USF Libraries
USF Digital Collections

The communicative two-way pre-writing task performed via asynchronous and synchronous computer-mediated communication an...

MISSING IMAGE

Material Information

Title:
The communicative two-way pre-writing task performed via asynchronous and synchronous computer-mediated communication and its influence on the writing expertise development of adult English language learners a mixed design study
Physical Description:
Book
Language:
English
Creator:
Sarieva, Iona
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
English as a second language
Writing processes
Pre-writing
Computer-mediated communication
Intertextuality
Dissertations, Academic -- Secondary Education -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This study addresses a gap in the second language writing research through examining processes occurring during the pre-writing and drafting stages of adult second language learners' writing when computer-mediated communication (CMC) pre-writing activities are involved. The theoretical framework adopted in the study is Writing-as-process approach with a focus being the pre-writing and drafting stages of the writing process. The design of the study is a parallel component mixed method design with an ongoing dominant qualitative stage and a nested less-dominant quantitative stage. In the qualitative stage of the project, two case studies were conducted: a group case study of the 60 intermediate level ESL learners who participated in the study and a more focused instrumental case study of eight learners selected based on their post-treatment writing gains.^ The research focus was on the social environment, including the learning task, peer interaction, mode of communication, and the intertextual connections between pre-writing discussions of the participants and their first drafts. The qualitative stage findings suggested that the CMC mode of communication (synchronous vs. asynchronous) affected differently the participants' patterns of interaction as well as the intertextual connections of their first drafts with the pre-writing discussions.^ In the quantitative stage, the researcher compared the first-draft writings of students who participated in asynchronous and synchronous pre-writing discussions (treatment) through the analysis of eight textual features of students' first drafts, namely: (1) syntactic complexity, (2) the amount of information present in a single focus, (3) the quantity of overall information present, (4) lexical information per clause, (5) vocabulary complexity, (6) rhetorical soundness, (7) presentation and development of main ideas, and (8) overall language use. The first five textual features, presented with continuous scores, were analyzed using five ANCOVA tests with significance level alpha being set at .05; the concomitant variables were the corresponding pre-treatment scores for each of the measures. Textual features 6-8, presented with ordinal scores, were analyzed through two-tailed Mann-Whitney U tests.^ While no differences were found for any of the eight proposed features when the writings of the participants in the asynchronous CMC and the synchronous CMC groups were compared, the consideration of the qualitative findings suggested that further analysis of an additional textual aspect of students' first drafts, more specifically - distinct lexical items, could be informative. The quantitative analysis of distinct lexical items of students' writings completed after synchronous and asynchronous pre-writing discussions was performed through the application of a two-tailed t-test.^ The results of this analysis led to the conclusion that at significance level alpha = .05, the CMC mode in which the pre-writing discussion was completed influenced differently students' first drafts on a lexical level: the intertextual connections between the pre-writing interactions and the first drafts of the participants from the asynchronous group at a lexical level were significantly stronger than those of their counterparts who participated in synchronous pre-writing discussions.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Iona Sarieva.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 625 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001969461
oclc - 272307058
usfldc doi - E14-SFE0002296
usfldc handle - e14.2296
System ID:
SFS0026614:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001969461
003 fts
005 20081117093040.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 081117s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002296
035
(OCoLC)272307058
040
FHM
c FHM
049
FHMM
090
LB1601 (ONLINE)
1 100
Sarieva, Iona.
4 245
The communicative two-way pre-writing task performed via asynchronous and synchronous computer-mediated communication and its influence on the writing expertise development of adult English language learners :
b a mixed design study
h [electronic resource] /
by Iona Sarieva.
260
[Tampa, Fla.] :
University of South Florida,
2007.
3 520
ABSTRACT: This study addresses a gap in the second language writing research through examining processes occurring during the pre-writing and drafting stages of adult second language learners' writing when computer-mediated communication (CMC) pre-writing activities are involved. The theoretical framework adopted in the study is Writing-as-process approach with a focus being the pre-writing and drafting stages of the writing process. The design of the study is a parallel component mixed method design with an ongoing dominant qualitative stage and a nested less-dominant quantitative stage. In the qualitative stage of the project, two case studies were conducted: a group case study of the 60 intermediate level ESL learners who participated in the study and a more focused instrumental case study of eight learners selected based on their post-treatment writing gains.^ The research focus was on the social environment, including the learning task, peer interaction, mode of communication, and the intertextual connections between pre-writing discussions of the participants and their first drafts. The qualitative stage findings suggested that the CMC mode of communication (synchronous vs. asynchronous) affected differently the participants' patterns of interaction as well as the intertextual connections of their first drafts with the pre-writing discussions.^ In the quantitative stage, the researcher compared the first-draft writings of students who participated in asynchronous and synchronous pre-writing discussions (treatment) through the analysis of eight textual features of students' first drafts, namely: (1) syntactic complexity, (2) the amount of information present in a single focus, (3) the quantity of overall information present, (4) lexical information per clause, (5) vocabulary complexity, (6) rhetorical soundness, (7) presentation and development of main ideas, and (8) overall language use. The first five textual features, presented with continuous scores, were analyzed using five ANCOVA tests with significance level alpha being set at .05; the concomitant variables were the corresponding pre-treatment scores for each of the measures. Textual features 6-8, presented with ordinal scores, were analyzed through two-tailed Mann-Whitney U tests.^ While no differences were found for any of the eight proposed features when the writings of the participants in the asynchronous CMC and the synchronous CMC groups were compared, the consideration of the qualitative findings suggested that further analysis of an additional textual aspect of students' first drafts, more specifically distinct lexical items, could be informative. The quantitative analysis of distinct lexical items of students' writings completed after synchronous and asynchronous pre-writing discussions was performed through the application of a two-tailed t-test.^ The results of this analysis led to the conclusion that at significance level alpha = .05, the CMC mode in which the pre-writing discussion was completed influenced differently students' first drafts on a lexical level: the intertextual connections between the pre-writing interactions and the first drafts of the participants from the asynchronous group at a lexical level were significantly stronger than those of their counterparts who participated in synchronous pre-writing discussions.
502
Dissertation (Ph.D.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 625 pages.
Includes vita.
590
Co-adviser: Wei Zhu, Ph.D.
Co-adviser: Jeffra Flaitz, Ph.D.
653
English as a second language.
Writing processes.
Pre-writing.
Computer-mediated communication.
Intertextuality.
0 690
Dissertations, Academic
z USF
x Secondary Education
Doctoral.
773
t USF Electronic Theses and Dissertations.
856
u http://digital.lib.usf.edu/?e14.2296



PAGE 1

The Communicative Two Way Pre Writing Task Performed v ia Asynchronous a nd Synchronous Computer Mediated Communication a nd i ts Influence o n t he Writing Expertise Development o f Adult English Language Learners: A Mixed Design Study by Iona Sarieva A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of World Languages College of Arts and Sciences and Department of Secondary Education College of Education Unive rsity of South Florida Co Major Professor: Wei Zhu Ph.D. Co Major Professor: Jeffra Flaitz Ph.D. Anthony Onwuegbuzie, Ph.D. Tony K. Erben, Ph.D. Date of Approval: November 9, 2007 Keywords: English as a second language, writing processes, pre writing, english as a second language, computer mediated communication, intertextuality 2007 Iona Sarieva

PAGE 2

DEDICATIONS To my h usband Andr i a n my daughter Iana, and my parents, Margarita and Ivan Sarievi, with my deepest love and appreciation for all they have done for me.

PAGE 3

ACKNOWLEDGMENTS allow sincere thankfulness to all the people who have contributed to the resear ch project presented in this dissertation. First of all, I would like to express my deepest gratitude to my committee members for supporting and advising me during the process of designing and implementing the research project presented in this dissertati on F oremost, I would like to express my great appreciation to my co major professors, Dr. Wei Zhu and Dr. Jeffra Flaitz who guided me through out the years of my graduate studies and spent many hours discuss ing with me my research plans and outcomes and p rovided me with their thoughtful feedback. I also owe my appreciation to Dr. Tony Erben who helped me focusing the theoretical framework of my dissertation study and to Dr. Antony Onwuegbuzie who guided me throughout the study design and data analysis proc ess. I would also like to thank Bethany Bell for her help, encouragement, and support. Here I would like express my deepest thankfulness to my husband Andrian Pervazov, who was with me at every step of my graduate studies and the dissertation writing. He not only supported me morally and emotionally throughout the process but

PAGE 4

also participated in the study designing the custom software which I used for my data analysis. I would also like to thank my daughter Iana for her understanding and emotional suppor t with which she was able to provide me despite of her young age and to ask her for forgiveness for the many hours I did not spend with her while writing this dissertation. I am very grateful to the ELI at USF staff, faculty, and students for supporting me during the process of data collection. I particularly would like to thank Ms. Barbara Smith Palinkas, the Assistant Director for Curriculum and Instruction at the ELI at the time the data for this study were collected, for her collaboration and experti se. I am also sincerely thankful to Ms. Jenny Schultz, who was the leading writing instructor during the two semesters of the data collection, for her patience and thoughtful insights she shared with me during our numerous debriefings. I thank the student s from Level III Writing classes as well for their participation in this study. Last but not least, my sincere appreciation is to my friends and fellow Ph. D. students, Aline Harrison, Dr. Annmarie Gorenc Zoran, Irshat Madyaro v, Darunee Dujsik and Dr. Ru i Cheng for their help and encouragement.

PAGE 5

TABLE OF CONTENTS LIST OF TABLES ... XIX LIST OF FIGURES XXI LIST OF CMC LOGS ... XXII LIST OF MATRICES XXIII ABSTRACT XXIV CHAPTER I: INTRODUCT ION 1 S TATEMENT OF THE P ROBLEM ................................ ................................ ....................... 1 T HEORETICAL B ACKGROUND ................................ ................................ ........................ 4 R ATIONALE OF THE S TUDY ................................ ................................ ........................... 7 R ESEARCH Q UEST IONS ................................ ................................ ................................ 9 Quantitative Stage Research Questions ................................ ................................ .. 10 Qualitative Stage Research Questions ................................ ................................ .... 12 S IGNIFICANCE OF THE S TUDY ................................ ................................ ...................... 13 I NFERENCE Q UALITY AND I NFERENCE T RANSFERABILITY C ONSIDERATIONS ................ 14 D EFINITIO N OF T ERMS ................................ ................................ ................................ 17 O RGANIZATION OF R EMAINING C HAPTERS ................................ ................................ .. 21 CHAPTER II: LITERATU RE REVIEW 23 O VERVIEW ................................ ................................ ................................ ................. 23 O N THE N ATURE OF W RITING ................................ ................................ ..................... 24

PAGE 6

vi Towards Understanding Text Construction ................................ ............................ 34 Examining Textual Features of Writing ................................ .............................. 35 T units ................................ ................................ ................................ ........... 36 Idea units ................................ ................................ ................................ ....... 38 Lexical density ................................ ................................ ............................... 41 Vocabulary complexity ................................ ................................ .................. 44 Examining Deep Textual Levels ................................ ................................ ......... 47 Coherence ................................ ................................ ................................ ...... 47 Functional dimensions of text: stance ................................ ............................. 48 Scoring Rubrics: Accounting for Deep Textual Levels ................................ ....... 49 Beyond Linguistic Boundary ................................ ................................ .................. 51 Pre writing: From Solitude to Collaboration ................................ ........................... 54 Collaborative Writing ................................ ................................ ............................ 56 Peer Feedback ................................ ................................ ................................ .... 57 Dynamics and Patte rns of Peer Feedback ................................ ........................... 58 Incorporating peer feedback ................................ ................................ ........... 60 N EW L ITERACIES AND SLA ................................ ................................ ......................... 62 CMC and its Role in the Writing Process ................................ ............................... 65 CMC in the Second Language Classroom ................................ .............................. 66 Dynamics and Productivity ................................ ................................ ................ 67 CMC: Language Use ................................ ................................ .......................... 69 CMC: Synchronous vs. Asynchronous ................................ ............................... 70 S UMMA RY ................................ ................................ ................................ .................. 71

PAGE 7

vii CHAPTER III: METHOD 73 O VERVIEW OF THE S TUDY D ESIGN ................................ ................................ .............. 73 Quantitative Stage Design Overview ................................ ................................ ...... 76 Qualitative Stage Design Overview ................................ ................................ ........ 79 Qualitative Sub stage I: The Gro up Case Study ................................ .................. 80 Qualitative Sub stage II: The Instrumental Case studies ................................ ..... 81 Pilot Study ................................ ................................ ................................ ............. 85 S ETTINGS P ARTICIPANTS AND S AMPLING P ROCEDURES ................................ .............. 87 Study Context ................................ ................................ ................................ ........ 87 The ESL Program ................................ ................................ .............................. 87 The Writing Class ................................ ................................ .............................. 91 Participants ................................ ................................ ................................ ............ 93 Sampling Techniques ................................ ................................ ............................. 95 Quantitative Stage Sampling Techniques ................................ ........................... 95 The Group Case Study Sampling Techniques ................................ ..................... 95 The Instrumental Case study Sampling Techniques ................................ ............ 96 The Non student Participants ................................ ................................ ............. 96 cy ................................ ............................ 97 R ESEARCHER S R OLE IN THE S TUDY ................................ ................................ ............ 98 R ESEARCH P ROCEDURES ................................ ................................ ........................... 100 Qualitative Stage Procedures ................................ ................................ ................ 100 Quantitative Stage Procedures ................................ ................................ .............. 105 The Treatment ................................ ................................ ................................ .. 109

PAGE 8

viii I NSTRUMENTATION ................................ ................................ ................................ ... 110 Quantitative Stage Instrumentation for Text Analysis ................................ ........... 110 T unit Analysis ................................ ................................ ................................ 116 Idea Unit Analysis ................................ ................................ ............................ 116 Lexical Density Analysis ................................ ................................ .................. 121 Vocabulary Complexity Analysis ................................ ................................ ..... 124 Multiple trait Scoring Rubrics ................................ ................................ .......... 125 Text Analysis Software ................................ ................................ .................... 126 Internet Based Instruments for Data Collection ................................ ................ 127 T REATMENT D ESIGN C ONSIDERATIONS ................................ ................................ ..... 129 Prompt Design Considerations ................................ ................................ ............. 129 Treatment Task Design Considerations ................................ ................................ 130 D ATA A NALYSIS P ROCEDURES AND R ELIABILITY AND V ALIDITY C ONSIDERATIONS ... 133 Text Normalization and Coding ................................ ................................ ........... 133 Multiple trait Rubric: Reliability and Validity Considerations .............................. 137 Q UALITATIVE S TAGE I NSTRUMENTATION ................................ ................................ .. 138 The Researcher ................................ ................................ ................................ .... 138 IEP Documentation ................................ ................................ .............................. 139 Observation Field Notes ................................ ................................ ....................... 139 Writing Samples and CMC Logs ................................ ................................ .......... 140 The CMC Interviews ................................ ................................ ............................ 141 T RUSTWORTHINESS C REDIBILITY AND T RANSFERABILITY C ONSIDERATIONS ............ 145 Q UANTITATIVE S TAGE : D ATA A NALYSIS ................................ ................................ ... 146

PAGE 9

ix Comparing the ACMC and the SCMC Groups ................................ ..................... 147 Group Comparison: Age ................................ ................................ .................. 147 Group Comparison: E ducational Background ................................ ................... 147 Group Comparison: Gender ................................ ................................ ............. 148 Group Comparison: Pre treatment Writing ................................ ....................... 149 Analysis of the Post treatment Paragraphs ................................ ............................ 149 Analysis of Covariance ................................ ................................ .................... 1 50 Mann Whitney U Test ................................ ................................ ...................... 152 Pulled Score Analysis ................................ ................................ ...................... 152 Q UALITATIVE D ATA A NALYSIS P ROCEDURES ................................ ............................ 153 Sub stage I: The Group Case Study ................................ ................................ ...... 154 The IEP Documentation Analysis ................................ ................................ ..... 155 Field Observations Analysis ................................ ................................ ............. 156 The Treatment CMC Interactions Analysis ................................ ....................... 160 CMC Student Interviews ................................ ................................ .............. 165 Sub stage II: The Instrumental Case Studies ................................ ........................ 166 Analysis of observation narratives and interview records .............................. 168 The CMC Interactions Analysi s ................................ ................................ ....... 169 The Paragraph Intertextuality Analysis ................................ ............................. 170 Comparing the Instrumental Cases ................................ ................................ ... 172 C OMBINING THE Q UANTITATIVE AND Q UALITATIVE A NALYSIS ................................ .. 172 C HAPTER S UMMARY ................................ ................................ ................................ 173

PAGE 10

x CHAPTER IV: THE GROU P CASE STU DY STAGE AND THE QUAN TITATIVE STAGE ANALYSES AND RESULTS 174 I NTRODUCTION ................................ ................................ ................................ ......... 174 Q UALITATIVE D ATA A NALYSIS : S UB STAGE I ................................ ............................ 176 Computer Laboratory Sessions: Behavioral Patterns and Class Dynamics ............ 176 Class Dynamics ................................ ................................ ............................... 176 Class Management ................................ ................................ ........................... 179 Working with Software ................................ ................................ .................... 182 ICA ................................ ................................ ................................ .............. 183 L.E.C.S. ................................ ................................ ................................ ....... 184 Microsoft Word ................................ ................................ ............................ 185 The Treatment CMC Interactions ................................ ................................ ......... 187 CMC Discussions: Specifics ................................ ................................ ............ 187 Descriptive characteristics of CMC interactions ................................ ........... 187 Patterns of Dyadic Interac tions ................................ ................................ ......... 190 The collaborative dyads ................................ ................................ ................ 193 The dominant/dominant dyads ................................ ................................ ...... 198 The expert novice dyad ................................ ................................ ................ 201 The dominant/passive dyads ................................ ................................ ......... 205 The disengagement pattern ................................ ................................ ........... 209 Dyads with mutuality and equality mismatches ................................ ............ 210 Collaborative/dominant dyads ................................ ................................ ...... 210 Class Dynamics a nd Patterns of Interaction ................................ ...................... 221

PAGE 11

xi Computer Mediated Interview Outcomes ................................ ............................. 223 Attitudes Towards Writing ................................ ................................ ............... 228 Computer and CMC Related Experiences ................................ ......................... 231 CMC Interaction and Task Perception ................................ .............................. 232 CMC Inter action Tools and Typing ................................ ............................... 235 The Low Computer Skills Proficiency Students ................................ ............... 235 Qualitative Stage Summary of Results ................................ ................................ 237 ................................ ........................ 237 Patterns of Dyadic Interaction ................................ ................................ .......... 238 Factors Influencing the CMC Pre writing Interaction ................................ ....... 239 The Use of Computer as an Interaction Medium ................................ ............... 239 The Mode of CMC Interaction ................................ ................................ ......... 241 Partner Interaction Style ................................ ................................ ................... 243 Opinion on the Importance of Writing ................................ .............................. 243 Q UANTITATIVE S TAGE : A NALYSIS AND R ESULTS ................................ ....................... 244 Comparing the ACMC and the SCMC Groups ................................ ..................... 245 Group Comparison: Age ................................ ................................ .................. 245 Group Comparison: Educational Background ................................ ................... 246 Group Comparison: Gender ................................ ................................ ............. 247 Group Comparison: Pre Treatment Writing Score ................................ ............ 249 Post treatment Data Analysis and Results ................................ ............................ 257 ANCOVA Analysi s of Post treatment Writings ................................ ................ 258 ANCOVA results: syntactic complexity ................................ ....................... 262

PAGE 12

xii ANCOVA results: information present in a single focus ............................... 263 ANCOVA results: quantity of the overall information ................................ .. 264 ANCOVA results: lexical information ................................ .......................... 265 ANCOVA results: vocabulary complexity ................................ .................... 266 Mann Whitney U Test ................................ ................................ ...................... 268 Rhetorical soundness ................................ ................................ .................... 268 Presentation and development ................................ ................................ ...... 269 Overall language use ................................ ................................ .................... 269 Pulled Scores Analysis ................................ ................................ ..................... 269 Quantitative Results Summary ................................ ................................ ............. 272 Additional Quantitative Findings ................................ ................................ ......... 273 Design of the Additional Quantitative Research ................................ ............... 273 Results of the Additional Quantitative Research ................................ ............... 278 Summary of the Additional Quantitative Findings ................................ ................ 280 C HAPTER S UMMARY ................................ ................................ ................................ 281 CHAPTER V: INSTRUMEN TAL CASE STUDIES 282 I NTRODUCTION ................................ ................................ ................................ ......... 282 T HE C ASE S TUDIES ................................ ................................ ................................ ... 289 ACMC Actors, Low Gains Case Studies ................................ .............................. 289 Case Study One: Azad ................................ ................................ ..................... 289 ................................ ................................ ....................... 289 The treatment int eraction ................................ ................................ .............. 290 The paragraph: intertextual connections ................................ ....................... 294

PAGE 13

xiii Intertextual connections: conclusions ................................ ........................... 304 Answering the Research Questions ................................ ................................ .. 305 Case Study Two: Shin ................................ ................................ ...................... 307 ................................ ................................ ....................... 307 The treatment interaction ................................ ................................ .............. 309 The paragraph: intertextual connections ................................ ....................... 311 Intertextual connections: conclusions ................................ ........................... 320 Answering the Research Questions ................................ ................................ .. 323 ACMC Actors, Low Gains: Comparison ................................ .............................. 323 ................................ ................................ ................ 323 The Pre writing Interaction ................................ ................................ .............. 324 Post treatm ent Paragraph Gains and Intertextual Connections .......................... 325 Answering the Research Questions ................................ ................................ .. 326 ACMC Actors, High Gains Case Studies ................................ ............................. 326 Case Study Three: Felipa ................................ ................................ ................. 326 ................................ ................................ ..................... 326 The trea tment interaction ................................ ................................ .............. 328 The paragraph: intertextual connections ................................ ....................... 333 Intertextual connections: conclusions ................................ ........................... 346 Answering the Research Questions ................................ ................................ .. 347 Case Study Four: Isabella ................................ ................................ ................. 348 ound ................................ ................................ ................... 348 The treatment interaction ................................ ................................ .............. 349

PAGE 14

xiv The paragraph: intertextual connections ................................ ....................... 353 Intertextual connections: conclusions ................................ ........................... 366 Answering the Research Questions ................................ ................................ .. 367 ACMC Actors, High Gains: Compa rison ................................ ............................. 368 ................................ ................................ ................ 368 Pre writing Interaction ................................ ................................ ..................... 368 Post treatment Gains and Intertextual Connections ................................ ........... 369 Answering the Research Questions ................................ ................................ .. 371 ACMC Actors: Comparison ................................ ................................ ................. 372 ................................ .................. 373 The Pre writing Interaction ................................ ................................ .............. 373 Post treatment Paragraph Gains and Intertextual Connections .......................... 374 Answering the Research Questions ................................ ................................ .. 381 SCMC Actors, Low Gains Case Studies ................................ ............................... 383 Case Study Five: Ajwad ................................ ................................ ................... 383 ................................ ................................ .................... 383 The treatment interaction ................................ ................................ .............. 386 The paragraph: intertextual connections ................................ ....................... 392 Intertextual connections: conclusions ................................ ........................... 409 Answering the Research Questions ................................ ................................ .. 410 Case Study Six: Kamil ................................ ................................ ..................... 411 Kam ................................ ................................ ..................... 411 The treatment interaction ................................ ................................ .............. 412

PAGE 15

xv The paragraph: inertextual connections ................................ ........................ 414 Intertextual connections: conclusions ................................ ........................... 422 Answering the Research Questions ................................ ................................ .. 422 SCMC Actors, Low Gai ns: Comparison ................................ ............................... 423 ................................ ................................ ................ 423 Pre writing Interaction ................................ ................................ ..................... 423 Post treatment Gains and Intertextual Connections ................................ ........... 424 Answering the Research Questions ................................ ................................ .. 426 Case Study Seven: Kang ................................ ................................ .................. 427 ................................ ................................ ...................... 427 The treatment interaction ................................ ................................ .............. 429 The treat ment paragraph: intertextual connections ................................ ........ 433 Intertextual connections: conclusions ................................ ........................... 445 Answering the Research Questions ................................ ................................ .. 446 Case Study Eight: Sun ................................ ................................ ...................... 447 ................................ ................................ ......................... 447 The treatment interac tion ................................ ................................ .............. 448 The paragraph: intertextual connections ................................ ....................... 454 Itertextual connections: conclusions ................................ ............................. 474 Answering the Research Questions ................................ ................................ .. 475 SCMC Actors, High Gains: Comparison ................................ .............................. 476 Comparison of the Par ................................ .................. 476 The Pre writing Interaction ................................ ................................ .............. 477

PAGE 16

xv i Post treatment Paragraph Gains and Intertextual Connections .......................... 477 Answering the Research Questions ................................ ................................ .. 480 SCMC Actors: Comparison ................................ ................................ ................. 481 Comp ................................ .................. 482 The Pre writing Interaction ................................ ................................ .............. 483 Post treatment Paragraph Gains and Intertextual Conn ections .......................... 485 Answering the Research Questions ................................ ................................ .. 493 ACMC and SCMC Actors: Comparison ................................ ............................... 495 ................................ ................................ ................ 496 The Pre writing Interaction ................................ ................................ .............. 498 Post treatment Paragraph Gains and Intertext ual Connections .......................... 499 Answering the Research Questions ................................ ................................ .. 504 Summary of Findings ................................ ................................ ........................... 508 C HAPTER S UMMARY ................................ ................................ ................................ 510 CHAPTER VI: SUMMARY, DISCUSSION, RECOMMEN DATIONS, AND LIMITATIONS 511 I NTRODUCTION ................................ ................................ ................................ ......... 511 S UMMARY OF Q UALITATIVE AND Q UANTITATIVE S TAGE F INDINGS ............................ 511 T HE Q UALITATIVE S TAGE F INDINGS : D ISCUSSION ................................ ..................... 5 14 writing Task ............................. 514 CMC Mode Influences on Pre writing Interactions ................................ ........... 516 Language use ................................ ................................ ............................... 516 Patterns of dyadic interaction ................................ ................................ ....... 518

PAGE 17

xvii CMC Task oriented Exper iences ................................ ................................ ...... 520 Limited Computer Skills Participants ................................ ............................... 521 T HE Q UANTITATIVE S TAGE F INDINGS : D ISCUSSION ................................ ................... 522 ACMC and SCMC Groups: Comparison of Textual Features ............................... 522 Comparison of Lexical Intertextual Connections ................................ ................. 523 I NSTRUMENTAL C ASE S TUDY F INDINGS : S UMMARY AND D ISCUSSION ........................ 525 Participation in Pre writing Discussion ................................ ................................ 525 The In fluence of the Pre writing Discussion on the First Drafts: .......................... 528 Differences and Similarities ................................ ................................ ................. 528 I MPLICATIONS FOR F URTHER R ESEARC H ................................ ................................ ... 532 P EDAGOGICAL I MPLICATIONS ................................ ................................ ................... 534 L IMITATIONS OF THE S TUDY ................................ ................................ ..................... 536 Limitations of the Quantitative Stage ................................ ................................ ... 536 Limitations of the Qualitative Stage ................................ ................................ ..... 538 Limitations of the Instrumental Case Studies ................................ ........................ 539 C ONCLUSIONS ................................ ................................ ................................ .......... 540 REFERENCES 542 APPENDICES 565 A PPENDIX 1 A: I NTERMEDIATE A CADEMIC W RITING C OURSE : G OALS AND O BJECTIVES ................................ ................................ ........................... 566 A PPENDIX 1 B: C OMPREHENSIVE E NGLISH L ANGUAGE T EST (CELT) ......................... 568 A PPENDIX 1 C: T HE M ICHIGAN T EST OF E NGLISH L ANGUAGE P ROFICIENCY (MTELP) ................................ ................................ ............................ 570

PAGE 18

xviii A PPENDIX 2: ESL/CALL C URRICULUM O BJECTIVES AND A CTIVITIE S ........................ 571 A PPENDIX 3A: M ULTIPLE T RAIT R UBRICS P R E T R E A T M E N T P A R A G R A P H .............. 573 A PPENDIX 3B: M ULTIPLE T RAIT R UBRICS P OST T REATMENT P ARAGRAPH ............. 574 A PPENDIX 4: T REATMENT CMC T ASK ................................ ................................ ...... 576 A PPENDIX 5: A DDITIONAL A SSIGNMENT ................................ ................................ ... 578 A PPENDIX 6: S AMPLE NVIVO R EPORT ................................ ................................ ..... 579 A PPENDIX 7: P ARTICIPANTS Z SCORES BY M EASURE ................................ ................ 583 A PPENDIX 8 : A ZAD ACMC I NTERVIEW R ................................ ............................... 585 A PPENDIX 9: A ZAD T RAINING T ASK : ACMC P OSTING ................................ ............ 588 A PPENDIX 10: S HIN ACMC I NTE RVIEW WITH THE R ESEARCHER ............................. 590 A PPENDIX 11: O RDERED C OMPARISON M ATRIX ................................ ........................ 593 A PPENDIX 12: F ELIPA ACMC I NTERVIEW WITH THE R ESEARCHE R .......................... 597 A PPENDIX 13: F ELIPA ACMC P OST TO T EACHER S W ELCOME M ESSAGE ................. 600 A PPENDIX 14: I SABELLA ACMC I NTERVIEW WITH THE R ESEARCHER ...................... 601 A PPENDIX 15: A JWAD ACMC I NTERVIEW ................................ ............................. 605 A PPENDIX 16: K AMIL ACMC I NTERVIEW ................................ .............................. 609 A PPENDIX 17: K ANG ACMC I NTERVIEW ................................ ................................ 612 A PPENDIX 18: S UN ACMC I NTERV IEW ................................ ................................ .. 615 A PPENDIX 19: S UMMARY OF Q UANTITATIVE S TAGE F INDINGS ................................ ... 618 A PPENDIX 20: S UMMARY OF Q UALITATIVE R ESEARCH F INDINGS ............................... 619 A PPENDIX 21: S UMMARY OF C ASE S TUDY R ESEARCH F INDINGS ................................ 622 A PPENDIX 22: A DDITIONAL Q UANTITATIVE F INDINGS ................................ ............... 624

PAGE 19

xix LIST OF TABLES Table 3.1: Quantitative Stage Instrumentation ................................ ................................ 111 Table 3.2: Interview Questions Types and Trends ................................ .......................... 143 Table 3.3: Stances of Interaction So rted by Dyadic Pattern ................................ ............ 162 Table 3.4: Partially Ordered Matrix Fields: Group Case Study ................................ ....... 164 Ta ble 4.1: Pre writing Discussion per Dyadic Interaction: Descriptive Statistics ............ 190 Table 4.2: Dyadic Patterns of Interaction by CMC Group ................................ .............. 192 Table 4.3: Interview Questions Outline ................................ ................................ .......... 225 Table 4.4: ................................ ....... 246 Table 4.5: Dyad Assignment by Group ................................ ................................ .......... 248 Table 4.6: Pre treatment scores: Descriptive Statistics ................................ ................... 250 Table 4.7: Pre treatment t test: Statistics and Out comes ................................ ................. 255 Table 4.8: Descriptive Statistics: ANCOVA Datasets ................................ .................... 260 Table 4.9: ANCOVA Results ................................ ................................ ......................... 268 Table 4.10: Pulled z scores: Descriptive Statistics ................................ ......................... 271 Tab le 4.11: Independent t test of the Pulled z scores: Statistics and Outcomes ............... 272 Table 4.12: Independent t test of the Distin ct Lexical Items (DLIs): Descriptive Statistics ................................ ................................ ................................ ..... 280 Table 4.13: Independent t test of the Distinct Lexical Items (DLI): Statistics and Outcomes ................................ ................................ ................................ 280

PAGE 20

xx Table 5.1: Instrumental Case Study Participants: Average Gains ................................ ... 284 Table 5.2: Paragraph idea units: Azad ................................ ................................ ............ 303 Table 5.3: Paragraph Idea Units: Shin ................................ ................................ ............ 322 Table 5.4: Paragraph idea units: Felipa ................................ ................................ .......... 344 Table 5.5: Paragraph Idea Units: Isabella ................................ ................................ ....... 363 Table 5.6: ACMC Actors: textual aspects improved ................................ ....................... 375 Table 5.7: AC MC Actors: Paragraph Idea Units and the Level of Intertextual Connections ................................ ................................ ................................ 377 Table 5.8: Paragraph idea units: Ajwad ................................ ................................ .......... 408 Table 5.9: Paragraph idea units: Kamil ................................ ................................ .......... 421 Table 5.10: Paragraph idea un its: Kang ................................ ................................ .......... 443 Table 5.11: Paragraph idea units: Sun ................................ ................................ ............ 471 Table 5.12: SCMC Actors: textual aspects improved ................................ ..................... 486 Table 5.13: SCMC Actors: Paragraph Idea Units and the Level of Intertextual Connectio ns ................................ ................................ ................................ 488

PAGE 21

xxi LIST OF FIGURES Figure 1.1: Stages of the Study ................................ ................................ .......................... 9 Figure 3.1: Research Flow ................................ ................................ ............................... 75 Figure 3.2: Research Stages and Steps ................................ ................................ ............. 84 Figure 3.3: ICA2 ................................ ................................ ................................ .............. 86 Figure 3.4: Language Education Chat System (L.E.C.S.) ................................ ................. 86 Table 3.4: Partially Ordered Matrix Fields: Group Case Study ................................ ....... 164 Figure 3.6: Data Analysis Flo wchart: Instrumental Case studies ................................ .... 168 Figure 4.1: Dyadic Interact ion and CMC Patterns Indicators ................................ .......... 191 Figure 4.2: DLI Data Analysis Flow ................................ ................................ .............. 277 Figure 5.1: Instrumental Case Study Findings ................................ ................................ 50 7

PAGE 22

xxii LIST OF CMC LOGS Log 4.1: ACMC Collaborative Discussion ................................ ................................ ..... 194 Log 4.2: SCMC Collaborative Discussion ................................ ................................ ...... 196 Log 4.3: ACMC Dominant/Dominant Interaction ................................ .......................... 199 Log 4.4: SCMC Expert Novice Discussion ................................ ................................ .... 202 Log 4.5: ACMC Dominant/Passive Interaction ................................ .............................. 206 Log 4.6: SCMC Dominant/Passive Interaction ................................ ............................... 208 Log 4.7: ACMC Collaborative/Dominant Intera ction ................................ ..................... 211 Log 4.8: SCMC Collaborative/Dominant Interaction ................................ ..................... 215 Log 4.9: SCMC Dominant/Expert Interaction ................................ ................................ 220 Log 5.1: ACMC Pre writing Discussion: Azad and Partner ................................ ............ 291 Log 5.2: ACMC Pre writing Discussion: Shin and Partner ................................ ............. 309 Log 5.3: ACMC Pre writing Discussion: Felipa and Pa rtner ................................ .......... 329 Log 5.4: ACMC Pre writing Discussion: Isabella and Partner ................................ ........ 350 Log 5.5: SCMC Pre writing Discussion: Ajwad and Partner ................................ .......... 387 Log 5.6: SCMC Pre writing Discussion: Kamil and Partner ................................ ........... 413 Log 5.7: SCMC Pre writing Discussion: Kang and Partner ................................ ............ 430 Log 5.8: S CMC Pre writing Discussion: Sun and Partner ................................ .............. 449

PAGE 23

xxiii LIST OF MATRICES ACMC Matrix 5.1: Paragraph Comparison: Azad ................................ ......................... 296 ACMC Matrix 5. 2: Paragraph Comparison: Shin ................................ ........................... 313 ACMC Matrix 5. 3: Paragraph Comparison: Felipa ................................ ........................ 336 ACMC Matrix 5. 4: Paragraph Comparison: Isabella ................................ ...................... 355 SCMC Matrix 5.5 : Paragraph Comparison: Ajwad ................................ ........................ 395 SCMC Matrix 5.6 : Paragraph Comparison: Kamil ................................ ......................... 416 SCMC Matrix 5.7 : Paragraph Comparison: Kang ................................ .......................... 435 SCMC Matrix 5.8 : Paragraph Compar ison: Sun ................................ ............................. 456

PAGE 24

xxiv The Communicative Two Way Pre Writing Task Performed v ia Asynchronous a nd Synchronous Computer Mediated Communication a nd i ts Influence o n t he Writing Expertise Development o f Adult English Language Learners: A Mixed Design Study Iona Sarieva ABSTRACT This study addresses a gap in the second language writing research through examining processes occurring during the pre writing and drafting stages of adult second mediated communication (CMC) pre writing activities are involved. The theoretical fra mework adopted in the study is Writing as process approach with a focus being the pre writing and drafting stages of the writing process. The design of the study is a parallel component mixed method design with an ongoing dominant qualitative stage and a n ested less dominant quantitative stage. In the qualitative stage of the project, two case studies were conducted: a group case study of the 60 intermediate level ESL learners who participated in the study and a more focused instrumental case study of eight learners selected based on their post treatment writing gains. The research focus was on the social environment, including the learning task, peer interaction, mode of communication, and the intertextual connections between pre writing discussions of the participants and their first drafts. The qualitative stage findings suggested that the CMC mode of communication (synchronous vs.

PAGE 25

xxv intertextual connections of their first drafts with the pre writing discussions. In the quantitative stage, the researcher compared the first draft writings of students who participated in asynchronous and synchronous pre writing discussions (treatment) through the analysis of eight textu (1) syntactic complexity, (2) the amount of information present in a single focus, (3) the quantity of overall information present, (4) lexical information per clause, (5) vocabulary complexity, (6) rhetorical soundness, (7) presentation and development of main ideas, and (8) overall language use. The first five textual features, presented with continuous scores, were analyzed using five ANCOVA tests with significance level alpha being set at .05; the concomita nt variables were the corresponding pre treatment scores for each of the measures. Textual features 6 8, presented with ordinal scores, were analyzed through two tailed Mann Whitney U tests. While no differences were found for any of the eight proposed fea tures when the writings of the participants in the asynchronous CMC and the synchronous CMC groups were compared, the consideration of the qualitative findings more specifically distinct lexical items, could be informative. after synchronous and asynchronous pre writing discussions was performed through the application of a two tail ed t test. The results of this analysis led to the conclusion that at significance level alpha = .05, the CMC mode in which the pre writing discussion was

PAGE 26

xxvi connecti ons between the pre writing interactions and the first drafts of the participants from the asynchronous group at a lexical level were significantly stronger than those of their counterparts who participated in synchronous pre writing discussions.

PAGE 27

1 CHAPTER I: INTRODUCTION Statement of the Problem Computers have been used in second language teaching for several decades now. However, the fast development of computer technol ogies changed their role in our everyday life in general and in the language classroom in particular The explosive growth of the computer mediated communication (e mail, chat, on line conferencing, blogging, etc.) is a phenomenon that greatly influences t he way people interact with each other in the 21st century. This dissertation addresses the question of how the development of writing skills and writing expertise of second language learners could be influenced by these new trends in human communication i n the context of the contemporary writing research and pedagogy in which writing is perceived as a process of text creation through interaction with peers, other texts, and communities. The emerging globalization and the technological revolution of the 2 0th and the 21st century raised new requirements and expectations for educators. The processes of democratization of the Western society and education influenced the way how language teaching and, more specifically, writing in educational settings are appr oached: the diversity of second language learners and the necessity to meet their needs were recognized (Grabe & Kaplan, 1996). The wide implementation of computer technologies and means of communication further created substantial changes in the second l anguage (L2) learning environments. These factors call for research on the specifics of the L2

PAGE 28

2 means of communication. Second language writing research and practice are co nsistent with the trends of the general composition theory development. In the past decades, the research community reassessed academic writing, stressing the importance of perceiving it as a recursive and complex process of text creation rather than simpl y as a final product. This view on writing guided the creation of the basic principles of writing as process theoretical and pedagogical approach (Ferris & Hedgcock, 1998; Grabe & Kaplan, 1996; Reither, 1994). The new understanding of writing was so promin ent that Hairston (1982) qualified it as a paradigm shift in writing theory. The globalization and democratization processes also influenced the development of foreign and second language education theory and practice. The necessity of an approach that wo uld enhance communication among different cultures was recognized, which directed foreign language and second language educators to aim their instruction at establishing teaching practices that would lead to strong communication skills development (Salaber ry, 2001; Warschauer, 1999). In the second half of the 20th century, the Proficiency Based Language Learning model of foreign language education was described and accepted (Ellis, 1994); this model met the needs of both language learners and society. The P roficiency Based Language Learning model is consistent with the view on writing as an evolving process because both recognize the importance of teaching language with a focus on purposeful communication (Cooper, 1989; Salaberry, 2001). Furthermore, the de velopment of computer technologies and local and global networks are constantly reshaping society and education. This implementation of

PAGE 29

3 computer technologies into everyday life and academic environments is transforming basic concepts such as literacy and c ommunication (Aronowitz, 1992; Leu & Donald, 2000; Ware & Warschauer, 2006; Warschauer, 1999). This major change in the way people communicate and the new communicational tools they use influenced language education: the interaction is not limited anymore t o the boundary of the classroom; the text nature of the computer mediated communication (CMC) allows learners to have an easy access to the language generated during the communication, often within the same communication act (Warschauer, 1999). These maj or shifts in communication, writing theory, and second language education posed a new question to researchers: How can the new views on language, communication, and technology be applied in L2 educational settings in order to improve language acquisition a nd support learners in gaining and mastering language skills relevant to the needs of the contemporary society? (Liu, Moore, Graham, & Lee, 2002; Sotillo, 2000; Warschauer, 2004a). The higher accessibility of computers nowadays encourages researchers and p ractitioners to emphasize the application of technology in the L2 classroom in order to enhance language teaching and learning processes (Liu et al. 2002). This emphasis identified a new area in language teaching practice and research Computer Assisted Language Learning (CALL). ge Levy, 1997 p.1 ). could be a potential waste of resources if pedagogy does not take ad vantage of new p. 39, Salaberry, 2001 ). Further, Chapelle (2000) states that the development of computer networks, including loc al and global networks, extends the

PAGE 30

4 possibilities for interaction beyond the human computer dyad, providi ng opportunities for human human interaction via computers; thus, CALL becomes an important venue for language use and learning that offers new perspectives and poses new questions. Today the question is not if CALL contributes to language learning but ra ther what tasks should be applied in the second language classroom in order to support language acquisition in instructional settings (Chapelle, 2000; Oxford, Rivera Castillo, Feyten, & Nutta, 1998). The process of adoption and implementation of new ped agogical approaches influenced by the above stated societal and educational shifts has already started in the past two decades, language teachers and researchers have been concerned with issues related to creating a cooperative classroom environment that would allow for effective student teacher interaction and peer collaboration (Ferris, 2003; McDonell, 1992; Ohta, 2000). The nature of collaborative learning environments and collaborative learning processes is further shaped by the increasing presence of technology in the classroom. This presence entails new channels for communication and collaboration as well as new shifts of the dynamics of communication processes, thereby influencing the way language learning and teaching, in general, and, more specifi cally, the writing process of language learners are approached (Liu & Sadler, 2003; Salaberry, 2001; Ware & Warschauer, 2006; Warschauer, 1999). Theoretical Background The current study is guided by Writing as process theory (Grabe & Kaplan, 1996) ; partic ularly, the broadened theoretical framework within which the current study is designed embraces the social context approach to the writing process. Within this framework, wringing is understood as a cultural activity, which incorporates the notions

PAGE 31

5 of self expressive, cognitive, and discourse processes (Matsuda, 2003). This trend in writing theory and research is influenced by the Vygotskian perspective on language, thought, and higher mental processes as sociocultural processes ( Faigley, 1994; Grabe & Kapl an, 1996; Matsuda, Canagarajah, Harklau, Hyland, & Warschauer, 2003; Warschauer, 1999). The researcher views the Writing a s process theory from the stand point that human language in general and writing in particular are not simply a product of an individu al consciousness; rather they are imbedded into societal processes and contexts and can be understood only if the specific community of writers is taken into account (Faigley, 1994; Grabe & Kaplan, 1996; Reither, 1994). Thus, writing is viewed not simply as a creation of a final product but rather as a process of text creation in which the writing ability and expertise develop. This understanding of writing is further informed by Post process theory that assesses critically the expressivist and cognitivis t approaches to writing. Researchers who adopt the Post process approach to writing point out that p. 53 ), writing should not be viewed as a process with rigid predetermined stages but rather as a socially embedded phenomenon that includes writers, genres, audiences, cultural practices, and social interactions (Kent, 1999; Petraglia, 1999; Reither, 1994). Thus, the mai n focal points of writing as a continuous social phenomenon rather than a creation of a product have been emphasized. In the process of data collection and interpretation in the current study, writing was approached as defined by Kent (1999). According to him writing is: (a) a public act;

PAGE 32

6 in other words, in the writing act other language users apart from the writer are relation of understanding wi th other language users Kent, 1999 p. 2 ); and (c) a situated act within which writers compose from a certain position. When defining these core assumptions in the context of social understanding of writing, the writing act, the writer, and the writing product are not separate will and action, internal agenc Couture, 1999 p. 31 ). In the view of the researcher, this theoretical framework supports the understanding of the complexity of the L2 writing proc ess employing the reference to social structures and processes that take place prior to and during text construction. In addition, approaching the writing act, the writer, and the writing product as integrated parts of one process allows for attending to t on both macro and micro levels (Ferris & Hedgcock, 1998; Grabe & Kaplan, 1996). Therefore this theoretical framework, as applied in the context of a study that aims to viewed to empower the recognition and interpretation of the specifics of the L2 learning processes and writing skills development. This theoretical framework influenced the design of the study, focus specification, type of language learning tasks employed and learning environments under investigation. The understanding of writing as a complex phenomenon encouraged the researcher to consider both qualitative and quantitative aspects of data analysis. The view on writing as a public and interpretative act g uided the focus to the pre writing stage design of the communicative pre writing task. Further, accepti ng writing to be a situated act guided the

PAGE 33

7 focus on the social context and interaction within which the specific writing processes evolve including macro and micro textual levels. Rationale of the Study The goal of the researcher in the current study was to answer questions related to the current processes in language education; the researcher investigated how the process of L2 writing could be supporte d by synchronous and asynchronous CMC used in the pre writing stages. The social context approach to writing guided the current research project towards examining the process of L2 literacy skills development in a social context that incorporates the inter actions among various actors as well as consideration of their backgrounds and goals. More precisely, the presented research study investigated how a specific communication task performed in asynchronous and synchronous CMC environments influenced the writ ing skills development of adult language learners their past and current writing experiences the researcher aimed to build a canvas within which the data analysis evolved an d conclusions were made. The study is needed because a broader conceptualization of writing processes viewed as they occur in socially and historically dynamic contexts would contribute to writing theory and pedagogy (Faigley, 1994; Matsuda, 2003); in ad dition, the learning processes supported by CALL and CMC and their influence over the language classroom dynamics need to be further studied, described, and presented ( Sotillo, 2000; Warschauer, 2004b) More research is needed about how language learning a nd L2 writing development are supported by tasks performed in asynchronous and synchronous computer mediated communication modes, how learners perceive and apply

PAGE 34

8 communicative tasks in the process of CMC interaction, and how this affects their language per formance (Blake, 2000; Warschauer, 2004b ). Such research would provide a basis for better understanding of the changing philosophical values of teachers and learners as well as of learning as a social activity (Chapelle, 2000; Wertsch, 1985). Further, thi s would allow for expansion and fine tuning of methodological techniques concerning design of CALL acti vities, and it will shade light on the ways language learners can be supported in the contemporary language classroom while accounting for the social pro cesses (Chapelle, 2000). The current study was guided by this need to better understand the above stated issues related to CALL. It focused specifically on communication process performed via computer and its influence on writing skills formation. This s tudy approached the research questions from a pragmatist paradigm applying a mixed method design that was perceived to be applicable and desirable in applied research (Teddlie & Tashakkori, 2003). The pragmatist paradigm approach provided better understand ing of CMC contribution to ESL stude nt writing. The researcher viewed this approach to sustain the process of identification and interpretation task in relation to their writing achievements, as well as how the CMC task performance supports their writing. The problem was approached through a study of a parallel component mixed method design. The ongoing dominant qualitative stage wa s extended with a multiple case study (Yi n, 2003) while the nested less dominant quantitative stage involved an experimental design (Caracelli & Greene, 1993; Tashakkori & Teddlie, 2003). The researcher perceives that applying more comprehensive analytical techniques

PAGE 35

9 allowed providing better and more thorough interpretations of the qualitative findings (Creswell, Plano Clark, Gutmann, & Hanson, 2003) which could reveal additional dimensions in the study and educational practices related to writing. Figure 1.1 presents graphically the stages of th is parallel component mixed design study. Figure 1.1 Stages of the Study Research Questions The study aimed to answer the following overarching research question: How can computer mediated communication influence the writing skills development of Engli sh as second language adult learners? This question was addressed in both the qualitative and quantitative stages of the study focusing on series of sub questions.

PAGE 36

10 Quantitative Stage Research Questions In the quantitative stage of the study, the followin g general research question was posed: How does the CMC mode (synchronous vs. asynchronous) used for the completion of a communicative two way pre writing task influence the quality of proached through following specific questions: Research Question 1: What is the difference in the syntactic complexity present in the post treatment paragraphs of the stud ents who participated in the synchronous versus asynchronous CMC pre writing discussion? Research Question 2: What is the difference in the amount of information present within a single focus of the post treatment paragraphs of students who participated i n synchronous versus asynchronous CMC pre writing discussion? Research Question 3: What is the difference in the quantity of the overall information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronou s CMC pre writing discussion? Research Question 4: What is the difference in the lexical information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Question 5: What is the difference in the vocabulary complexity present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion?

PAGE 37

11 Research Question 6: What is the difference in the rhetorical so undness present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Question 7: What is the difference in the presentation and development of the main point and its suppo rt present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Question 8: What is the difference in the overall language use present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Following the current views on interaction being beneficial for language learning and establishing of writing skills, it was expected that the synchronous and asyn chronous (Beauvois, 1997; Warschauer, 1999, 2004b). Further, it was expected that the discussion in the two CMC modes could influence the composition process in a different way considering the discourse differences of asynchronous CMC and synchronous CMC interactions (Sotillo, 2000). Thus, t he overall hypothesis was that the textual aspects to be studied, namely syntactic complexity, lexical density, idea units, vocabulary compl exity, rhetorical soundness, presentation and development of the main point and its support, and overall quality of language use, would be influenced differently by synchronous and asynchronous CMC pre writing interactions because of the specifics of the t wo CMC modes (Sotillo, 2000). It was difficult to formulate more specific hypotheses because of the limited research that focuses on the way CMC modes influence writing production of second language learners.

PAGE 38

12 Qualitative Stage Research Questions The lea ding research question addressed in the qualitative stage is the following: How do dyadic interaction, performed via asynchronous computer mediated communication (ACMC) and synchronous computer mediated communication (SCMC) pre iews on CMC and writing influence writing development of adult intermediate level ESL learners? This research question was approached by focusing on specific issues when analyzing the data collected during the qualitative stage. The following specific rese arch questions helped to guide this analysis: Research Question 1: process of establishing their writing skills? Research Question 2: What patterns of dyadic interaction do participants manifest during the asynchronous and synchronous CMC interaction process? Research Question 3: What are the factors that influence the CMC pre writing interaction process? How do these factors influence the interaction process? Further, the qualitative stage wa s extended through multiple case studies. They aimed to reveal the specifics of the processes that took place during the pre writing interactions and their relations to the following writing through closer examination of the pre writing interactions and fo llowing writings of eight purposefully selected participants. The researcher questions addressed through these case studies were: Research Question 1: How do peers participate in asynchronous and synchronous pre writing CMC interactions?

PAGE 39

13 Research Questio n 2: (a) How do they use the specific ideas and language generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas? Significance of the Study Today, th e field of L2 writing is affected by various changes these include changes in the society, demographics, technology, and the discipline of writing itself. These changes have brought about a research focus shift from teacher to learner: a substantial portion of the current writi ng research is focused on student collaboration during the revision and editing stages of the writing process. Recent studies have also examined student teacher interaction in the process of feedback provision during the writing process (Ferris, 2003; Hyla nd & Hyland, 2006a ). The current study aimed to investigate the interactions that occurred in pre writing stages and were completed via two different modes of CMC and how these interactions influenced the development of L2 writing skills. Thus, the study i s viewed by the researcher as a contribution to the body of writing research through shedding light on communication and composition that occur in the initial planning and drafting stages of the writing process. Such an understanding would allow teachers t o support learners through early intervention accomplished via more precise design and application of pre writing tasks and creation of appropriate collaborative opportunities for planning, rather than postponing these until later stages of the writing pro cess. In other words, it is viewed that understanding the specifics of the initial stages of the writing process would allow for providing learners with opportunities to construct actively and effectively their writing starting at an early

PAGE 40

14 stage which should positively influence the overall writing process (Ferris & Hedgcock, 1998). The current study also sought to provide a better basis for understanding how learners collaborate via CMC. It was thought that findings from this study would extend our kn owledge base of collaborative processes in the context of writing instruction and would provide a better foundation for developing effective teaching approaches and practices consistent with changes in the society, technology development, and writing theor y and practice. In order to answer modern societal demands for effective incorporation of technology into the learning process (Warschauer, 1999), the current study aimed to reveal how adult ESL learners used the documented CMC interactions when composing their first drafts. Thus, the researcher believed that studying factors that influence CMC interaction, CMC patterns, and how they influence pre writing processes addresses the current needs of learners and society and would add to the body of L2 writing r esearch. Inference Quality and Inference Transferability Considerations In the qualitative and mixed design research paradigm, different terminological approaches are used when the quality of a study is addressed. Maxwell and Loomis (2003) address the val idity of qualitative and quantitative stages of a mixed design studies subdividing it to internal validity and generalizability. Merriam (1998) discusses case study quality in terms of validity and reliability. Furthermore, Patton (2002) and Lincoln and Gu ba (1985) in the context of qualitative research, propose the terms credibility a term that corresponds to internal validity in the qualitative research,

PAGE 41

15 transferability corresponding to external validity, dependability corresponding to reliability, an d confirmability corresponding to objectivity. In this study, when issues related to the Qualitative and Quantitative stages were internal, external validity, reliability, objectivity for the Quantitati ve part were used; while in the Qualitative part the terms proposed by Lincoln and Guba (1984) and Patton (2002) presented earlier were used. When addressing issues related to the study as a whole, i.e. embracing the Qualitative and Quantitative stages, th e terms proposed by Teddlie and Tashakkori (2003) were used. Specifically, the authors suggest that the terms inference quality and inference transferability are more appropriate when summarizing issues related to mixed design studies. They define inferen ce quality as a construct that addresses issues related to the Quantitative stage internal validity and the credibility of the Qualitative stage, while inference transferability addresses issues related to the Quantitative stage external validity and Quali tative stage transferability (Teddlie & Tashakkori, 2003). Further in this section, the study design considerations related to inference quality and transferability will be shortly discussed; they will be addressed in mo re detail in Chapter III The data w ere collected at the intensive English language program (IEP) of a large urban state university in the Southeast United States. An important delimitation of the study was the fact that the researcher chose to collect data from a single educational setting; this decision was made in order to minimize the history threat to internal validity (Onwuegbuzie, 2003). To obtain higher inference quality of the findings and to guarantee better control of the experimental settings for the Quantitative stage of the stud y, the

PAGE 42

16 students were randomly assigned by the researcher to two treatment groups synchronous and asynchronous computer mediation groups, thereby diminishing the statistical regression and selection threats. However, the history threat may not have been e ntirely eliminated because the two groups had different instructors. This threat was further diminished by the fact that the groups followed the same curriculum and used the same textbook and they wrote the two paragraphs based on the same tasks and using the same referential material. The Qua ntitative stage of the study had a considerable threat to the external validity, using the framework of Onwuegbuzie (2003). The sample was unlikely to include representatives of the vast variety of the ESL student popu lation in the United States, which posed a threat to population validity. Taking into account the fact that the internal and external validity of the findings pertaining to the quantitative stage of the study may have be en affected by researcher bias (Onwu egbuzie, 2003), the use of T unit, idea unit mean length, idea unit number, lexical density, vocabulary complexity, and multiple trait scores of the writing samples was calculated This calculation was performed following strict unified procedures describe d in Chapter III In order to minimize the research threat further, the researcher trained the ESL instructors how to use the writing assessment instruments selected for the study. Thus, 25% of the writing outcomes were scored independently by the research er and the teachers, the outcomes of the scoring were compared, the achieved inter rater reliability was above 90%. The credibility, transferability, dependability, and confirmability of the qualitative findings were assured through the techniques of pro longed engagement, persistent observation, triangulation, and non student participant member checking (Lincoln &

PAGE 43

17 Guba, 1985). Prolonged engagement and persistent observation techniques are expected to allow the collection of thick data relevant to the rese arch questions and to support the credibility of the findings and interpretations produced in the study. Further, multiple sources of data were considered to confirm the emerging themes and t o formulate the findings, which wa s expected to strengthen the in ference quality and transferability of the study (Merriam, 1998). The researcher collaborated with the actors of the case studies in the process of specification and conceptualization of the emergent themes; these actors presented the cases that maximized the difference in the phenomena of interest strengthening the inference transferability of the study (Merriam, 1998). Assuring the quality of data collection and interpretation in each stage of the study is perceived to allow for high inference quality an d inference transferability of the findings of the study as a whole. In order to assure the quality of the inference, the study was carefully designed which provided a basis for rigorous interpretation of the data collected. Further, inference transferabi lity was determined by providing strict procedures for data collection and careful description of data analysis. This was viewed to be necessary criteria that assured sufficient information about the context in which the study took place and clear presenta tion of the reasoning process that led to drawing of the final conclusions. Definition of Terms Achievement l evel. In this study, achievement level was writing gain scores as measured by the pre treatment and post treatment multiple trai t rubric scores.

PAGE 44

18 Asynchronous Computer mediated Communication This is communication facilitated by computers that does not require a common clock for the interlocutors and can be performed with little or significant delay (Warschauer, 1999). This type of communication can be completed via e mail or electronic bulletin/discussion boards. In this study, the abbreviation ACMC will be used. The study focus ed on ACMC performed on a discussion board hosted on an online courseware Internet Classroom Assistant v ersion 2 (ICA2) (Nicenet, 1998). Computer Assisted Language Learning. Levy (1997) defines computer assisted language learning (C ALL ) Levy, 1997, p. 1). In this study, the abb reviation CALL will be used. Computer mediated communication. Computer mediated communication refers to communication of humans facilitated by computers. There are two modes of computer mediated communication: synchronous, in which the communication proces s develops in real time (usually performed in a chat format), and asynchronous, in which the communication is performed with delay via email or electronic bulletin boards (Warschauer, 1999). Communication exchange in both modes was investigated in the curr ent study. In this study, when referring to computer mediated communication, the abbreviation CMC will be used. Communication Task In this study, a task wa s viewed as communicative activity focused on meaning that involves language comprehension and manip ulation; this activity can stand alone but is also designed to support the development of specific genre writing skills through a range of communicative procedures in a specific sociorhetorical

PAGE 45

19 situation. When constructing this definition, the researcher w as guided by two definitions of tasks that complement each other, that of Nunan (1989) and Swales (1990). Communication Tasks Typology Tasks assigned to learners are classified depending on the way the information is available to them (Pica, Kanagy, & Fal odun, 1993). The Communication Tasks Typology includes three basic types of tasks. When performing the first type of task, one way exchange learners communicate based on information that is known only to one of the learners. The goal of the other members of the group is to find the missing information during the communication process. The second type, two way exchange requires pieces of information that are available to the learners to be used when constructing the whole. Finally, the third task type, rea ching agreement is performed when all learners have the same information and, based on this information, they need to reach an agreement. The two way exchange task will be used in the current study. Distinct Lexical Item The same lexical items that appea red one or more times in either the CMC interaction or the post treatment paragraph were considered to present one distinct lexical item (DLI). Idea Unit. In this study, idea unit is perceived to be a clause that contains one verb phrase and the noun and p repositional phrases, adverbs, and so forth that belong t o it is defined as an idea unit (Chafe, 1985). Interaction Patterns. This term addresses the d ynamics of small group or pair interaction behaviors (Stroch, 2002) Intertextuality. Intertextuality pre sents the understanding that a text is related to other texts and is a part of a network of textual relations (Allen, 2000). In the current

PAGE 46

20 study three levels of int ertextuality were considered: (a ) lexical level with unit of analysis lexical item; ( b ) ide a unit level with un it of analysis idea unit; and (c ) organizational level with unit of analysis the pre writing discussion organizational suggestions such as topic sentence, supporting ideas, and conclusions, and their relation to the draft organization. L1. This is an abbreviation for first language or native language used in literature related to language acquisition (Ellis, 1994). L2. This is an abbreviation for second language that refers to the language that is not the native language of the learne r and is learned in a country where it is either the official or the dominant language (Ellis, 1994). Lexical Density The ratio of the lexical items count to the total number of clauses is defined as lexical density. Lexical density reveals the average a mount of lexical vocabulary as demonstrated by the number of unique words used (Halliday, 1989). Lexical Item. In the current study, it was considered that for al differ, differed, different, difference, differing, differently 1989, p. 65). Multiple trait Rubrics Rubrics that st ate specific criteria for assessing the quality of writing accounting for the given topic, the genre, and specific textual aspects are defined as multiple trait rubrics (Ferris & Hedgcock, 1998). Post process writing theory A theoretical and pedagogical a pproach that supports the view that there is no one generalizable writing process. Rather, writing acts are communicative public interactions occurring in specific contexts.

PAGE 47

21 Synchronous Computer mediated Communication Type of communication facilitated b y a computer in which no time delay occurs. In this study, the abbreviation SCMC was used to refer to a text based synchronous communication performed via Language Education Chat System (L.E.C.S): a free web based synchronous conferencing software designed especially for language teachers and learners. T units This abbreviation stands for minimal terminal units. The T unit consists of the main clause of the sentence including the subordinate clau ses and is used for syntactical level text analysis (Hunt, 1 965). Vocabulary complexity Vocabulary complexity is the richness of the active vocabulary of the participants on a given topic. Unique Lexical Item. Each occurrence of a lexical item in the studied texts (CMC interactions and post treatment writings) wa s identified as a unique lexical item. Writing as process theory This theoretical and pedagogical approach presents writing as a non linear continuum of decision making, planning, and audience consideration (Grabe & Kaplan, 1996). Organization of Remaini ng Chapters The following chapter, Chapter II, provides a discussion of the literature related to the present study. The main trends in development of literacy theory, writing process theory and approaches, CMC in the second language classroom, and task ty pes are presented. In Chapter III, a comprehensive description of the study design, settings, and participants is provided. The data analysis and results of the qualitative and the quantitative stages of the study are presented in Chapter IV. Further, the dissertation continues with Chapter V, in which the eight instrumental case studies are represented.

PAGE 48

22 Finally, Chapter VI provides summary and discussion of the findings as well as recommendations and implications.

PAGE 49

23 CHAPTER II: LITERATURE REVIEW Overvie w This review of literature presents the current trends in research and education that influence the field of writing in general and, more specifically, the field of L2 writing and CALL. The significant influence of Writing as process theory on L2 writing research focus and practice is discussed based on the works of researchers such as Cooper (1989), Emig (1971), Faigley (1994), Grabe and Kaplan (1996), Hairston (1982), Johnson (2004), Matsuda (2003), etc. that contributed to the development of the field. Writing as process theory building is reviewed in conjunction with other trends in writing research that focus on the textual features of writing and is further examined in the context of the social trends in the writing literature. Another important deve lopment in L2 writing which is examined in the current chapter is the role of the new technologies in education and more specifically in L2 education. Researchers agree that in contemporary research of education, the question is not whether technology can contribute to language learning but rather which forms of technology and what particular tasks are most ef fective (Blake, 2000; Oxford et al.,1998; Warschauer, 1999). Seminal research publications such as and Swaffar (1998), Sotillo (2000), and Warschauer (1999) that provide the basis for such agreement are presented in the literature review as well.

PAGE 50

24 On the Nature of Writing In the current writing research literature, writing is viewed as a process rather than as a final product. However, the interpretati ons of the nature of this process vary significantly within the process writing theory. Currently the expressivist and cognitivist movements within the paradigm of the writing process are approached critically by theorists who call for understanding of wri perspective that is deeply concerned with the social nature of the writing phenomenon (Kent, 1999; Matsuda, 2003; McComiskey, 2000). Within the Post process, there are claims that Writing as process strives to pres ent a generalized process that would fit any writing; thus, it should be taken to the next level of theory building post process (Kent, 1999) on writing: within the para digm of Writing as of activities out of ea Bizzell, 1986 p. 49 ), and writing itself is strongly imbedded into the social context (Grabe & Kaplan, 1996; McComiskey, 2000). In this study, fol Post process theory is not viewed as a complete rejection of process pedagogy or process might be more productively defined as the r ejection of the dominance of process at the expense of other p. 78 79 ). Thus, the Post process theory is perceived to further develop the Writing as process theory adding a 1994 p. 109 ) to writing research and pedagogy and recognizing the uniqueness of each writing act. McComiskey (2000) states that the nature of writing as a social rather than a solely an expressive or cognitive phenomenon does not

PAGE 51

25 rejection of the process movement, but rather [presents] its p. 53 ). Thus, the multiplicity of writing theoretical positions could be better understood in the context of the history of writi ng process development. The view on writing as a process emerged from the social and educational changes of the 1960s and 1970s in the United States. These changes strongly influenced the current trends in writing research and practice (Hairston, 1982). The need to incorporate and support non traditional students in the higher education settings triggered a new view on writing and writing instruction among practitioners and researchers (Grabe & Kaplan, 1996). The work of researchers such as Emig (1971) an d Hairston (1982) analyzing the final writing product analysis. Emig (1971) outlined several basic notions about the writing process such as its non linearity influenced by audience, unevenness and recursiveness influenced by learning progression, and the effect of collaboration on the quality of the process and final product. Another reason for researchers to focus on studying the writing process was t he fact that the instructional practices based on this approach needed a theoretical foundation that would support them (Grabe & Kaplan, 1996). Thus, approaching writing as a process called for new understanding of text construction and the composition pro cess itself. However, researchers point out that there is no unity in the writing theory in regards with the nature of writing processes (Faigley, 1994; Grabe & Kaplan, 1996; Lockhart & Ng, 1995; Matsuda, 2003); moreover, in the field of L2 writing researc h and pedagogy, the rigid formulation of process approach has been opposed. In her 1991

PAGE 52

26 overview of teaching of writing traditions, Ann Raimes (1991) presents several stages of writing instruction and research development. In the late 1960s and early 1970s writing was approached with strong focus on form. In this period, writing instruction aimed to reinforce and test the accurate application of grammatical rules and sentence combining. The main limitation of this writing approach, as outlined by practition ers and researchers (Raimes 1991; Zamel, 1980), is that students were limited to syntax manipulation tasks, and the complexity of writing was ignored. Another influential trend in writing theory and instruction started in this period with that of contrastive rhetoric. raised the attention on cultural differences. Kaplan stressed that these differences may shape writing and provided the foundation for the development of the field of contrastive rhetoric. In the second half of 1970s the focus in writing research and instruction shifted to the writer. This shift was influenced by L1 writing research in the Uni ted States initiated by Emig (1971) and Zamel (1976). However, the overemphasis on personal reflection in writing allowed scholars involved in the study of academic writing to conclude that this approach cannot be a leading one in postsecondary academic se ttings. The late 1980s shifted the focus of writing towards the audience outside the classroom focus on the reader. This approach focused on academic demands and discourse communities (Raimes, 1991). A different classification of the development of writ ing approaches aligned with the L1 writing research development is provided by Grabe and Kaplan (1996). They

PAGE 53

27 contrasted two periods in writing research: the first period addressed was writing as a product approach (the 1960s) and writing as a process appro ach (a period that started at the beginning of the 1980s). According to these authors, the complexity of developing writing skills in general had been underestimated up until the end of the 1960s. Writing was taught following a prototypical model of instru ction, and little time had been devoted to the actual act of writing (Ferris & Hedgcock, 1998; Kroll, 1991). The product approach to writing, prevalent in this period, had been limited by concentrating on the final text produced by the students; thus, the procedures and strategies involved in the writing of coherent and readable texts had not been addressed. In the early 1980s, Writing as process was widely accepted in the effort to answer at least some of the questions related to writing instruction. It f reed the instruction from simplistic assumptions about writing, moved writing instruction from one draft writing assignments to multiple drafting, stressed the recursive nature of writing that involves multiple stages including planning, feedback, focusing on content, revision, and editing, and pointed out the importance of audience consideration (Ferris & Hedgcock, 1998). In fact, researchers viewed Writing as a process as a paradigm shift in the field of writing (Grabe & Kaplan, 1996; Hairston, 1982). It is important to note that Writing as process followed the development of the education philosophy. Four stages of development of Writing as a process approach are described in the literature (Faigley, 1994; Grabe & Kaplan, 1996): (1) The expressive stage w as the first stage of the development of Writing as a process approach. The main characteristic of this stage was the encouragement that writers received to look for their authentic voices. The main problem of this stage was

PAGE 54

28 that it was assumed that the wr iters already have all the intellectual resources and simply need guidance to an appropriate outlet (Grabe & Kaplan, 1996). (2) The recognition of this weakness of the expressive stage led the researchers towards the next stage, the cognitive (Grabe & Kapl an, 1996). In the early 1970s, the Writing as a process approach was influenced by the findings in the field of cognitive aloud methodology to investigate what writers are actually doing (at l east on the surface) while writing. (3) The beginning of the 1980s brought a new view on communication and writing in particular writing as a social event (Grabe & Kaplan, 1996). In this social stage views on writing were strongly influenced by the dev elopment of sociolinguistics and the Hallidayan functional linguistics. The social context that defines the particular writing purpose was viewed by a number of researchers such as Cooper (1989), Faigley (1994), and Ferris and Hedgcock (1998), and as cruci al for the understanding of writing. The sociolinguistic view on writing required new methodology that would allow revealing the specificity of both the learner and the social setting in which the learner is functioning. Ethnographic methodology allowed fo r collecting naturally occurring data in meaningful social contexts. The main critique addressing the issues in this stage of Writing as a process was methodological the qualitative data collected in case study research do not allow for generalization an d hypothesis testing (Grabe & Kaplan, 1996). However, the social approach brought new perspectives to language learning and writing research and situated it within a meaningful context, raising new research questions and providing foundation for new educat ional approaches such as whole language approach (Goodman, 1989).

PAGE 55

29 (4) Research and educational practices in discourse community stage address writing at post secondary levels as it is occurring in social rhetorical contexts (Cooper, 1986). The researchers who approached writing from the perspective of the social rhetorical context, addressed the discourse communities, groups that include readers, writers, texts, and social contexts in which information exchange is performed (Swales, 1993). Swales defined a discourse communities approach as one that shares common public goals, is a forum of discussion, provides feedback and information to its members, develops discourse expectations and genres, uses a specific set of terminology, and has enough members to di scuss matters of importance in a wider group. This approach is viewed as combining the whole language approach but on a tertiary level. There are a number of issues related to the discourse community approach, such as the danger of empowering the elite mem bers of the community, which could be an obstacle for knowledge exchange (Grabe & Kaplan, 1996). It is also important to state that the discourse community approach had a positive effect on writing curricula development in post secondary settings through p romoting the development of a community within a writing class. This level writing class. In the beginning of the 1980s, serious criticism of the expressive stage of the Writing as pr ocess emerged, such as lack of a strong theoretical foundation. In the late 1970s to mid 1980s, Flower and Hayes (1980) developed their cognitive model of writing. It was based on research involving protocol analysis. In their model, they interpreted the c omposing processes as consisting of three major components: (a) the term memory.

PAGE 56

30 According to the authors, within the composing processor, the following three processes are generated: (a) planning (generating ideas, organizing information, setting goals); (b) translating; and (c) reviewing. This model has met the needs for a theoretical foundation of Writing as process and influenced the writing research community; however, it also ra ised questions and criticisms. North (1987) states that the model is too vague to meet the requirements for theory model building and raises a concern about the methodology that protocol analysis (i.e. extraction of knowledge from verbal report) might no t be a primary valid methodology for building a theoretical model. Finally, it was pointed out that the model does not allow addressing the differences between writing processes of novice and experienced writers (Grabe & Kaplan, 1996). In the 1990s, Flower s and Hayes (1980) continued their research addressing a range of difficulties which skilled and less skilled writers might experience during the writing process, further investigating the task environment and writing processes and extending their approach to writing as a cognitively defined and contextually constrained activity. The main implications of their model are that it discovers that better writers have a richer sense of what they want to do when they write and the writing is a teachable process th at involves a socially contextualized approach to writing. The main critique of this model is that it presents the skilled and less skilled writers in the same continuum and seeks to describe features similar to all writers (Grabe & Kaplan, 1996). Bereiter and Scardamalia (1987) proposed a model that identifies the differences between novice and skilled writers. This model addresses the genre and audience differences, and provides specifics of revision process and tasks. They outline two

PAGE 57

31 models of writing: knowledge telling and knowledge transformation Knowledge telling is generally being employed by less skilled writers who use fewer strategies that are generally useful for personal experiences writing. Skilled writers may also use the knowledge telling mo del when the task has been internalized. The second model, the knowledge transforming model, explains a more complex writing process. In this model, knowledge telling is just one of the components. The starting points of the whole writing process are the m ental representation of the assignment and following problem analysis and goal setting. This larger model approaches writing as a problem solving activity in which the problem solving space and the rhetorical problem space are in constant interaction throu gh problem translation and knowledge telling processes. These spaces also are in constant connection with the content knowledge of the writer and his/her discourse knowledge. From a cognitive perspective, these two models represent the two different ways i n which writers compose (Bereiter & Scardamalia, 1987). The advantages of these models are that writing processes of skilled and novice writers are not viewed as one dimensional; in addition, the knowledge transforming model accounts for the heuristic natu re of writing, the content, rhetoric, and the discourse; it stresses the differences in the writing processes of writers with various levels of writing skills development (Grabe & Kaplan, 1996). Scholars in the field of composition research have also recog nized certain limitations of the model, such as the fact that it does not account for context influence on the writing process, as well as how and when the writer makes the cognitive transition between the two models (Grabe & Kaplan, 1996).

PAGE 58

32 In the 1980s, researchers also raised the question of the importance of the formal aspects of writing and problem solving strategies. Genre based approaches to writing addressed these questions identifying the importance of genre knowledge which supports ng development beyond the knowledge telling stage (Grabe & Kaplan, 1996). This approach views language as a functional construct in which form gives structure for meaningful communication and is an integrated system (Painter, 1989). In his study of element question of the importance of practicing writing in a range of genres, thereby, language for real pu rposes and integrating language and content. Researchers who adopted the genre approach to writing also concluded that genre is an important element of meaning making. Kroll (1991) points out that teaching how to work with language effectively in order to communicate content would make students more successful learners and writers. Recognizing the fact that genres are culturally embedded, Kroll (1991) points out The gen re approach also allows the beginning writers to address the text on content and genre levels. The genre approach to writing was extended to higher levels of education. Swales (1990) states that genres comprise a system for accomplishing social purposes by verbal means, and the ability to use genre would support students in their learning and communication in academic contexts. In the 1980s, in the field of second language acquisition (SLA), writing was studied based on the assumptions that first and second language writing processes share

PAGE 59

33 basic features (Ferris & Hedgcock, 1998). For example, it was shown that, in the case of experienced second language writers there is a noticeable transfer of their first language writing abilities into L2 writing processe s. In addition, it was observed that the text construction strategies of these writers were independent of and often more advanced than their L2 skills. Thus, it was confirmed that L2 writers have instructional needs comparable to first language writers (C umming, 1989 ). However, the results of English as a Second Language (ESL) writing research also showed that the L2 learners have their unique writing instructional needs. Less experienced writers in their first languages, learners with lower L2 proficiency as well as the sociocultural context of writing require special attention and additional research in the context of Writing as process (Ferris & Hedcock, 1998). These unique needs of L2 learners require close attention when evaluating L2 writing. While textual approach provides important information about language skills development such as syntax complexity and vocabulary, the process of writing also should be studied from the social perspective taking into account meaning constructing, planning, audie nce considerations, and revisions based on feedback received from peers and the instructor (Ferris & Hedgcock, 1998). Furthermore, social interaction is viewed special a ttention to teacher feedback and student collaboration when studying the writings of L2 learners could contribute to L2 writing practices (Ferris, 2003). These trends are consistent with the fact that the field of SLA emphasizes the dialogic nature of verb al communication in both oral and written discourse (Voloshinov, 2001). In the following sections of this chapter, writing research related to the text construction

PAGE 60

34 elements that influence the process of writing as well as the role the social interactions that take place during the writing process will be presented. Towards Understanding Text Construction Writing as a process approach focuses on text creation; however, the importance of accounting for more formal features such as syntax structure and vocabu lary is also recognized. This recognition is based on the fact that the nature of text involves both structure and appropriateness of text in the written discourse; that is, language has both pragmatic function and form (Widdowson, 1978). In a SLA context, vocabulary and syntax mastery, along with appropriate written discourse functions of the text, are important features that lead to a performance that is closer to the native like goal in the target language. The current research study attempts to address the process of text creation in the context of both surface and deep structures as well as to reveal the sociocultural factors that affect this process. The following sections of this review of literature address these different aspects of text constructio n. In the traditional paradigm of text analysis, the focus is on the final product, and procedures that result in composing coherent and readable texts are not taken into account (Ferris & Hedgock, 1998). Grabe and Kaplan (1996) propose an overall text mod el that considers several approaches to text, which could reveal the specifics of text construction such as coherence, cohesion, and text functional use dimensions. They (p. 79) and the model is rather descriptive. The authors attempted to lay out the components of text construction theory in order to provide guidelines for future exploration of how these components interact in a text. This model guided the

PAGE 61

35 decision regarding which textual features would be appropriate to examine un its, idea units, lexical density, and vocabulary will be presented in the context of model and the multiple trait grading rubrics that were used to analyze and assess level that examines syntax and cohesion components; the text is approached through analysis of its deep structure taking into account text semantics and coherence; thus, four basic components are outlined: on the surface level these are syntax and cohesion and on the deep level, these are semantics and coherence. These four basic components of the text can also be interpreted in terms of sent ential and whole text level and as well as in terms of writer reader relation (text interpersonal features). Further, the model presents the Lexicon as a textual feature of dual nature that affects and is being affected by the syntax, semantics, cohesion, and coherence; thus, lexis is related to the surface and deep structure of text. Last but not least, text construction is influenced by factors placed outside of the linguistic boundaries such as reference and world background knowledge. Examining Textual Features of Writing Grabe and Kaplan (1996) stress the importance of clear understanding of three theoretical aspects for writing instruction: (a) what is the array of knowledge that one needs to possess in order to be a successful writer; (b) what are th e skills that presuppose successful learning how to write; and (c) what are the social contexts that shape the

PAGE 62

36 successful writing instruction. They also point out that more research is needed in order to move towards clearer consensus regarding these aspec ts of writing instruction. The current research presented in this dissertation aims to focus on three main aspects of text that are outlined in the model proposed by Grabe and Kaplan (1996): (a) on surface (sentence) level analyzing the smaller portions of text, (b) on the overall text discourse embracing deep structures of text, motivations for form factors as well as factors outside the linguistic boundaries, and (c) how different texts relate to each other (Lemke, 1985). This third aspect is not explic itly presented in the model but is logically related to it through the features of the model that lay beyond linguistic boundary, namely T units In the research tradition analyzing the text surface l evel, the concept of the T unit has been viewed as one of the central constructs to be examined in the T unit as the main clause of the sentence including the subordinat e clauses, the following are examples of five separate T The sailor finally came on deck. He was tall. He was rather ugly. He had a limp. He had offered them the prize Grammatical Structures at Three Grade Levels Hunt found out increases its complexity across the grades, as measured by mean length of T units. He When discussing the application of T unit measure in analyzing writings of adult second language learners, Gaies (1980) pointed out that although L2 acquisition by adults does not follow the developmental sequence of first language development, the syntactic maturity of adult second language learners undergoes similar sequences and its growth

PAGE 63

37 could be measured through T unit analysis. Further, based on data collected in L2 context, Larsen Freeman (1983) concluded that T units could be a reliable indicator of nts involving T units in Larsen Freeman study were the average number of words per T unit and the average number of words per error free T unit, where each T unit was ranked depending on the error type in terms of syntax, vocabulary, and orthography. There fore, Larsen Freeman (1983) along with other researchers (Gaies, 1980), raised the issue of taking into consideration only error free T While the errors in the L1 learning process are less characterist ic, the occurrence of errors in the L2 learning process is frequent, and in order to account for the developmental errors in the T unit analysis, only error free T units should be considered. However, the criteria for error free T unit is interpreted diffe rently in the research (Gaies, 1980). In the current study, error free T units along with T units with minor morphosyntactical errors will be considered. Applied to instructional settings, the T unit approach offers a number of recommendations to instructo rs that includes raising conscious attention to syntactic form, providing tactics for revision, and increasing familiarity with syntactic patterns confirm that mean T un it length increases with the language and writing maturity of the writer. However, the limitations of this measure should be taken into consideration. Although T unit analysis presents a practical way to analyze student writing, it addresses the issue of w riting only on a syntactic level (Grabe & Kaplan, 1996). According to

PAGE 64

38 improves overall writing, and this type of analysis should not be used as a single predictor of th e overall quality of writing. In addition, keeping in mind that sentences are part of a larger text, the T unit analysis does not reveal accurately the way the learners acquire the sentence structure in a rhetorical context (Bardovi Harlig, 1992). Neverthe less, the T unit methodology provides the best way to assess writing development when comparing groups that are different in terms of age or proficiency level. Unfortunately, this methodology does not provide an indicator that is always sensitive enough fo r comparisons of learners of similar language proficiency levels (Gaies, 1980; Grabe & Kaplan, 1996). Taking into account these limitations of the T unit measure, in the current study, additional measures aiming to reveal the writing development of the par ticipants was used; these methods are presented and discussed in the following sections. Idea units Another measure that aims to reveal the specifics of the texts produced by the participants in the current study is idea unit. This measure, proposed by Ha lliday (1967, cited in Chafe 1980) and further developed by researchers such as Kroll (1977) and Chafe (1980, 1985), first emerged in the analysis of spoken language production. Researchers noticed that the human mind has the ability to repair the unevenne ss of spontaneous speech, thereby creating the impression of smoothness. If the speech is possible to notice that what seems to be a smooth speech is actually interm ittent -the term used in the expressions of focuses

PAGE 65

39 subject, and objects plus When discussing the idea unit concept in a written discourse, Chafe (1985) also limits the single focus to a clause; he points out that previous research shows th at difficulties in text production appear to take place when writers attempt to compose more than one clause at a time. Further, Chafe states that the fact that written language is produced at a slower pace in comparison with spoken language, that writers are usually construct their thoughts allow the written language to have generally longer idea units. However, Chafe (1985) notes that readable texts are organized in a f ormat that punctuation and other markers that would signal the boundaries of the idea unit; in addition, the length of an idea unit in a readable text is coordinated with the amount of information that could be assimilated by the reader to whom the text is addressed. Generally, the idea units in readable texts, as reported by Chafe (1985), are approximately 11 words long compared to 7 words in spoken language. In addition, idea units in a readable written text tend to be more complex and interdependent in comparison with spoken language: devices such as nominalization, attributive adjectives,

PAGE 66

40 pre and post posed present and past participles, prepositional phrases, and compl ement phrases (for a full list of devices see Chafe, 1985, pp. 109 110) are used to achieve this interdependency and complexity. In language research, idea units are most often used in studies related to recall as a measure of how information received eit her through listening or reading is remembered and reproduced by language learners (Carrell, 1987; Hildyard & Hidi, 1985; Wong, analysis. This decision was made base qualities, including its length, influence text readability. In addition, the idea unit presents a way to identify the quantity of information that the learners can handle in a single focus (Chafe, 1985). It is perceived that this measure would allow discussing quality and quantity of the texts written by the participants in the context of the research questions posed in the study. More specifically, the idea unit would help identify how much information the stu dents attempt to handle in one attention focus. This was measured through the mean length of idea units, while the quantity of information the students attempt to handle in a paragraph was measured by the total number of idea units used in the paragraph wr iting. Because using idea units allowed the researcher to focus on the quantity of information that the learners were able to communicate in their writings, and the morphosyntactical features of their writing were measured in T units (as discussed in this chapter and Chapter 3 of the current paper), the researcher decided to use morphosyntactically correct idea units as well as idea units that contained certain errors

PAGE 67

41 Lexical density Grabe and Kapl semantics, and coherence. It assists the s yntactic construction of text and influences text cohesion and coherence. Thus, the lexicon analysis offers a way to understand the text that goes beyond the boundaries of the context in which the clause is viewed when T units are considered. Halliday ( 1989) points out that for such a fundamental category as clause, it is Halliday (1989) extends the interpretation of clause, defining it as a functional unit that has a tr representation of the phenomena of experience, as these are interpreted by the members of the culture; (2) as the expression of speech function, through the categories of mood 3) as the bearer of the message, which is organized in the form of theme and addressing the difference between spoken and written language: written language is denser and this density could be measured when approached on a clause level. The shell colour s are yellow in green areas, pink of leaf litter, and reds and browns in beach Halliday, 1989, p. 69).

PAGE 68

42 cha racteristics that could reveal the specifics of written and spoken language. Further, in his discussion of lexical density, Halliday (1989) makes a distinction between two main vocabulary classes: lexical and grammatical. Lexical items are the content word s that participate in an open class membership, their semantic links could be extended indefinitely. He provides the following example to illustrate an open set of lexical items (the lexical items are italicized): door is in contrast with gat e and screen ; also with window, wall, floor and ceiling ; with knob, handle, panel and sill ; with room, house, hall ; with entrance, opening, portal there is no way of closing off the sets of items that it is related to, and new items can always come into original) While lexical items participate in an open set, grammatical items are part of a closed system, for example personal pronouns contrast with other pronouns only. Halliday also points out that many language ite ms could be easily classified as grammatical or lexical; however, with certain items, such as modal adverbs (e.g. always, perhaps), this distinction is more difficult and the lexical density analysis needs to be consistent when classifying such items. Wha t lexical density allows measuring via calculating the ratio of lexical items to the number of clauses is the information presented in a given text (Halliday, 1989; MacDonald, 2002; Read, 2000). In his analysis of written and spoken language produced by na tive speakers, Halliday outlines that language density is measurable, and when the

PAGE 69

43 language is produced in comparable oral and written discourses, the density of the written one is much higher; in other words, the information is organized in a more intrica te way. Moreover, according to MacDonald (2002), lexical density allows one to measure the degree of information packed into sentences. In her review of research involving lexical density, MacDonald reports that oral discourse contains about two lexical it ems per clause, whereas in scientific written discourse the lexical density could reach 11 to 13 items. In SLA research, lexical density has been used as one of the measures that would of spoken language production allowed Mehnert to conclude that lexical density is affected by planning time the more time learners have, the higher was the density of their language. The results of the same study also showed that lexic al density is closely related to fluency students who are more fluent tend to use a larger percentage of content words. Further, research shows that when considering lexical density as a measure, the prompts and the tasks that aim to elicit language prod uction should be designed in such a way that they call for production of comparable texts in terms of their complexity (Read, 2000). Reid (1990) compared two essays written during the Test of English as a Foreign Language (TOEFL), this comparison revealed that the two types of essays compared were significantly different in terms of the three lexical measures he applied: the average length of the words, the lexical density as percentage of content words, and the percentage of pronouns. When applying lexical density measure, it should be taken into consideration that lexical density is particularly higher in the noun and subject phrases (MacDonald, 2002).

PAGE 70

44 This fact would make it more difficult to imply that the information is equally presented in the structur es of the whole text produced by the student. Similar unevenness of lexical density was measured by Shih (2000) in a L2 production context. In her analysis of the largest Taiwanese Learner Corpus of English, she reports noun phrases to be four times higher in density than verb phrases. Therefore, lexical density is a useful measure that could be applied in text analysis including written language production of L2 learners (Cook, 2001). However, it may not be a representative indicator of overall text qualit y and information packing if used as single indicator (MacDonald, 2002). Moreover, this concern should be even higher when analyzing texts written by language learners. Although the texts written by language learners demonstrate the same relationship of de proficiency is reflected in language density (Lauren, 2002; Mehnert, 1998), it is important to consider other measures that would help to make conclusions about the co mplexity of a text written by a language learner. Vocabulary complexity In the field of SLA teaching, lexis was somewhat neglected in the period of 1945 1970s. In the SLA literature, the heavy emphasis on structure is pointed out to be a possible reason f or the fact that lexis was accounted for only as far as grading and selection of texts (Carter & McCarthy, 1988). In the early seventies, this lack of attention to vocabulary was recognized as a major gap in language teaching and consequently in SLA resear ch. Wilkins (1972, cited in Carter & McCarthy, the original).

PAGE 71

45 In recent SLA research, vocabulary acquisition and teaching is a frequent topic. Researchers and theorists such as Chomsky (1995) and Cook (1998) state that the lexicon has a central role for language acquisition; it needs to be acquired in meaningful chunks that involve not only the word with its basic meaning but also the projection that a word has on the syntax context in which it is used. The interpretation of lexis implies that the vocabulary is linked to the context as a whole and more specifically to the syntax (Cook, 1998) Further, Cook (1998) suggests that, in the process of acquisition, vocabulary is the they progress to phrases with grammatical inflections. Thus, vocabulary use is an important element of language acquisition and examining the development of L2 The current research on vocabulary acquisition is guided by questions that would disclose the nature of word and lexical competence, and how they could be approached in the L2 classroom. Laufer (1997) summarizes SLA research that reveals lexical problems to be in the core of reading/comprehension difficulties of language learners. Haynes and Baker (1993) outlined that while factors such as syntactic complexity have impact on based on empirical research results, researchers concluded that vocabulary accounts for the Nation, 2001). learners choose vocabulary while writing is perceived by both university professors (San tos, 1988) and language learners themselves (Leki & Carson, 1994) as one of the

PAGE 72

46 important factors that affect the quality of writing. A number of studies on writing use scoring methods that rank vocabulary as one of the main features of L2 writing (Nation, 2001). Furthermore, several different measures have been suggested in the research for examining the vocabulary development of language learners. The measure used in the nd described by him as the most complete profile that avoids the drawbacks of measures such as lexical originality, lexical sophistication, and lexical quality. Using a software package developed especially for this vocabulary analysis procedure, text lexi cal characteristics such as number of total words, number of unique words, that is word families, frequency level of the vocabulary used in the text can be obtained. The frequency level is determined based on three word lists: General Service List of Engli sh Words compiled by West in 1953 and Academic Word List compiled by Coxhead (1998, cited in Nation, 2001). Accounting for the frequency with which the vocabulary items used by language learners appear in the target language adds to the precision of L2 lea terms of quantity but also quality (Nation, 2001). In the previous sections of this chapter, textual features of writing that focus on text surface level and link it to the deep level were discussed. Although lexis is related to deep textual features as well as surface, it is impossible to see this relation to deep features such as coherence, semantics, and stance without reaching for different ways of student text analy sis. Further, literature related to functional analysis of texts will be presented.

PAGE 73

47 Examining Deep Textual Levels Coherence While connected closely with the surface features of text, the deep features have their own specifics; the analysis of these specif ics could reveal important issues of the writing process and development of language learners. An important concept that has been traditionally part of writing research and instruction is writing coherence (Bamberg, 1983; de Beaugrande & Dressler, 1981; Gr abe & Kaplan, 1996). The notion of c oherence specifies how the information is organized in a text, and how this organization affects the discourse topic (Lee, 2000). Text analysis that takes into account this textual characteristic could get into deeper le vel of understanding the processes of both writing and reading (Grabe & Kaplan, 1996). Brown and Yule (1983) argue that text coherence is to a greater extent a creation of the reader rather than of the author. Although Writing as a process approach recogn izes the importance of reader consideration in the composition process, as well as the fact that the readers may impose their own interpretations of texts, accepting that the reader is the constructor of text coherence would undermine important characteris tics that the text possesses. The text itself and consequently the writer are the main factors conveying coherence; moreover, research results provide evidence that coherence is contained in the text rather than imposed by the readers (Grabe & Kaplan, 1996 ). Grabe and Kaplan (1996), based on the analysis of several models (Mann & Thompson, 1988, 1992; Martin, 1992; Meyer, 1975, 1985; Sperber & Wilson 1986; van Dijk & Kintsch, 1983, cited in Grabe & Kaplan, 1996), outlined the following three key charact eristics of coherence: (a) it has a discourse theme; (b) it embraces the text from the level of sentence to the top level structuring; and (c) it provides information structure in

PAGE 74

48 order to guide the reader. These criteria closely relate to the functional dimensions of texts; in other words, they reveal how the writer considers the reader, how the subject matter about which the writer is composing affects the process of composition and the text itself, and how the background knowledge of the writer shapes t he text (Halliday, 1985). Functional dimensions of text: stance Coherence is also strongly connected with stance the functional dimensions of text that reach to text creation on interpersonal levels (Halliday, 1985). The examination of the functional o rganization of text could writing situation thereby disclosing the social functions of writing ( Cumming, Kantor, Powers, Santos, & Taylor, 2000; Grabe & Kaplan, 199 6; Halliday, 1985). Further, the analysis of text performed on the functional level allows one to evaluate the structural appropriateness of the text; in other words this type of analysis allows assessing if the text elements such as use of direct speech s tructures correspond with the text style (Grabe & Kaplan, 1996). The analysis of text functional dimensions is concerned with the way the text elements are combined rather than with what elements are combined. Different authors approach this notion of sta nce from different perspectives. In the early approaches to functional text analysis, Chafe (1982) develops two dimensions of text construction: text involvement/detachment and text integration/fragmentation. Enright, Grabe, Koda, Mosenthal, Mulcahy Ernt, and Schedl (2000) along with Grabe & Kaplan (1996 ) emphasize the relation of the writer to the reader and text. Further, Enright et al. (2000) and Grabe and Kaplan (1996) state that patterns of lexico syntactic elements construct the

PAGE 75

49 text and proposes text dimensions that are empirically identifiable in the texts rather than determined a priori. These dimensions are: rhetorical intention, interactivity, referentiality, immediacy of context, persuasion, abstractness, elaboration, evidentiality, and genre (Gr abe & Kaplan, 1996). Scoring Rubrics: Accounting for Deep Textual Levels researchers agree that there is no single measure that could be applied in order to reveal all or at l east most aspects of text that constitute its quality. To find an approach to text analysis that would allow maximizing the precision and detail of text construction analysis and assessment is still one of the main objectives of writing research (Cumming e t al., 2000). A possible way to approach writing while accounting for its complexity is the application of scoring rubrics that would be designed with both surface and deep features of writing in mind. There are four summative methods for approaching stud writing: holistic, analytic, primary trait, and multiple trait. These methods could be used both in large scale testing as well as in regular classrooms (Ferris & Hedgcock, 1998). Ferris and Hedgcock (1998) present holistic scoring as a way to asses s the overall text quality. Based on this assessment, the grader could make conclusions about the as one entity instead of looking at specific dimensions of writing ; this method emphasizes the strengths of the writer while involving examination of a wide range of text features (Cohen 1994; Ferris & Hedgcock, 1998; White, 1994). Another approach to text written by language learners is analytic scoring. This type of s coring would assist the text reader in the process of weighting distinct textual

PAGE 76

50 components. These components could include surface and deep textual features, for example the ESL Composition Profile created by Jacobs, Zingraf, Wormuth, Hratfiel, and Hughey (1981), contains five criteria that focus on content, organization, vocabulary, writings is provided by Ferris and Hedgcock (1998), this is the Essay Rating Profile an adap tation of Jacobs et al. (1981) model (Ferris & Hedgcock, 1998). Although this type of scoring is perceived to be easy for teachers to implement, it weakens the recognition of the complex connection in the written discourse (White, 1994). The third and the fourth types of summative writing assessment methods described by Ferris and Hedgcock (1998) are primary and multiple trait scoring. The strength of primary and multiple trait scoring is that the text is approached based on the context of the specific wri ting under analysis (Ferris & Hedgock, 1998; White, 1994). The goal of this approach is to develop salient criteria and traits for successful writing based on genre and topic in the context of the writing task. The fact that trait scoring does not address writing as generic texts but rather focuses on the specifics related to the writer (e.g. writing task, genre, topic, and audience), makes the method highly flexible and it allows guiding students throughout the writing process. The trait scoring technique could support different stages of the writing process teaching, drafting, and feedback (Ferris & Hedgock, 1998). Trait rubric also support the text analysis that seeks writing skills diagnostics. Cohen (1994) and Ferris and Hedgcock (1998) pointed out th at comparing to holistic and analytic scoring, trait based scoring has a number of advantages. First, this type of scoring has higher face validity because it is related directly to the task. Second, it provides higher content related and construct related validity

PAGE 77

51 because when the specific scoring instruments are developed, the expectations, textual based instruments, it is also easier to reach a consensus among raters applying trait based descriptions that are tied much closer to the task than the items of holistic and analytic instruments. However, the major drawback of this approach to L2 writing analysis is that it could be time consuming for classroom application. For the current study, the multiple trait method was applied using three traits that allow to focus on the deep and surface features of the text: (a) rhetorical soundness; (b) presentation and development of the main point and its support; and (c) overall language use (Fe rris & Hedgcock,1998). Beyond Linguistic Boundary The model of text construction proposed by Grabe and Kaplan (1996) and used to linguistic boundary incorporating featur es such as reference, world background knowledge, memory, emotion, perception, intention, logical arrangement (rhetorical patterns), and situation. Grabe and Kaplan (1996) point out these elements to be of a great importance for understanding text construc tion By including the features that lie beyond the linguistic boundaries in their model, Grabe and Kaplan (1996) suggest that staying within the boundaries of text surface and deep levels, would not allow one to understand how writers construct their texts. The authors of the model indicate that text construction is influenced by the previous experiences of the writers. Further, staying in the same stance, it could be implied that these exp eriences include other texts either read or created earlier by the writer in a

PAGE 78

52 different discourse, that a reader of a particular text brings his/her own frames for organizing content and context and transfers it to his/her own writing creating intertextua l connections. Considering the above from the perspective of the current study, whose goal is to analyze how pre writing, it was of interest to the researcher to find how the texts created in a CMC that intertextual analysis approaches the text from the standpoint of language use in communities (Lemke, 1985), and more specifically in language learning communi ties; this point of view on text does not contradict but rather extends the process view on writing (Cooper, 1986). The modern theorists engaged in text analysis view text as an entity that does not have an independent meaning (Allen, 2000). Thus, texts, literary or non literary, cannot exist as self sufficient entities because they do not function in a closed system (Still & Worton, 1990). Intertextuality in text analysis is an interdisciplinary approach to text (Worton & Still, 1990), not a transparent term that could be defined in a simple manner text as a part of a wider textual network. The concept of intertextuality is widely applied in the literary analysis and cultural studies (Allen, 2000; Lemke, 1985). In the 1960s, Julia Kristeva first coined the term in her early attempt to combine Saussurean and Bakhtinian theories of language and literature (Allen, 2000); it was further utilized in poststructuralilst theo Further, Lemke (1985) points out that the meaning of every text and discourse is

PAGE 79

53 influenced in an explicit or implicit way by other texts or discour ses; thus, the discourse practices of a community affect the understanding of a particular text by creating systems of texts related through community practices. In language learning context, intertextuality is perceived as preparing learners to deal with 1999). This preparedness could support in various ways the learning process: (a) through incorporating reading practices that promote comprehension and writing (Ferris & Hedgcoc k, 1998), (b) through engaging in writing activities that expand beyond the telling to writing as knowledge transformation (Bereiter & Scardamalia, 1987), (c) through encouraging learners to engage in self reflection (Blanton, 1999). Intertextuality is directly connected to the literacy practices of the community. Although it is traceable in the written product (the text), it presents actually part of the writing process through re vealing the engagement of the writer with other texts before, during, and after text creation (Blanton, 1999). Lemke (1985) proposes three specific features of texts that could help interpreting if two or more texts form a strong intertextual set: the Mode the Tenor, and the Field. Considering these dimensions, relationships and interactional ploys (Tenor), and to some degree across activity type and overt topic (Field ) (Lemke, 1985, p. 279). This approach to the text would allow one to follow the text construction process, taking into account the specifics of the participants: other, may make meanings through the relations they construe among them that penetrate

PAGE 80

54 Although in the above quote, Lemke does not refer specifically to language learners, th e quote is perceived by the researcher as applicable to them based on the fact that learners create and function in discourse communities that have their specific socioeducational dimensions (Ferris & Hedgcock, 1998; Grabe & Kaplan, 1996). Pre writing: Fr om Solitude to Collaboration In the writing process research literature, the interpretation of the pre writing stage is consistent with the development of the understanding of writing process. Rohman (1965) defines pre writing as a stage of discovery withi writing within a pre defined, linear process of pre writing, writing, and re writing. Further, Morris (1968) describes pre writing as an important period that f student probes his vast memory banks and attempts to find new, untried angles from 24). Building on the pre writing research of the 1960s, Kytle (1970) proposes a three stage pre writing strategy that involves analogy: (a) first stage exploration and discovery of the subject through outlining as many points as possible and classifyin g them; (b) reduction of the subject; (c) thesis formation. While these studies recognize the importance of the pre writing stage, they view it as a predetermined and linear sequence of steps that involves little or no interaction among the writers; thus, isolating the writer form the social environment within which writing process develops.

PAGE 81

55 The recognition of the social nature of writing and the importance of interaction for the writing process brings a different perspective on pre writing Rodrigues (198 3) proposes collaborative activities that aim to generate and organize ideas before students start the actual writing. B ased on her study of L1 remedial high school students, Abbott (1989) suggests that after being involved into brainstorming small group a ctivities during which ideas were talked out, students were more selective when choosing ideas to include in their essays and felt more positive towards their writings. In the L2 writing research of the 1980s, the process of schema building that aims to a ssure the readiness of learners to write and involves reading assignments as well as in class collaborative activities was recognized (Ferris & Hedgcock, 1998; Melvin & Stout, 1987; Smith, 1996). Although Spack (1984) in her study of one ESL college studen t does not discuss the collaboration as part of the pre writing process, she addresses pr e passively; it is the result of diligent effort on the part of each writer to self d efine a writing process of L2 learners was supported through subject knowledge building, idea sharing, and discussion of authentic materials in a third year French composition course. Further, F erris & Hedgcock (1998) provide pedagogical suggestions proposing several pre writing strategies that can be performed within a L2 collaborative process. They divide them into two categories: (a) unstructured pre writing freewriting, speed writing, brain storming writing, for example clustering and cubing activities that assist students in the process of topic exploration, strategies development, and relating new to already exiting knowledge.

PAGE 82

56 In t he contemporary writing research and pedagogy collaboration is recognized as an important vehicle of the learning process, it is an essential component of writing instruction that promotes the creation of writing communities within the class and further pr epares learners to participate in larger discourse communities (Grabe & Kaplan, 1996); limited when the pre writing stage is considered. The studies discussed in the co mprehensive writing research reviews by Ferris (2003), Ferris and Hedgcock (1998), Grabe and Kaplan (1996), Silva and Matsuda (2001) mainly address the collaboration in the writing process at stages that take place after the preparation for writing and pla nning are completed, focusing on collaborative writing and teacher and peer response. In the following section, a review of literature that presents interaction in the process of collaborative writing and peer feedback is presented. The researcher views th ese studies to be relevant to the present study; they depict specifics of the meaning negotiation process and its influence on text construction, and it is assumed that the discussion of their findings will support the current study. Collaborative Writing In their analysis of suggested teaching practices in writing at intermediate levels, Grabe and Kaplan (1996) present five main types of cooperative learning: (a) the Group Investigative method, developed by Sharan and Sharan in 1992, (b) the Learning Toge ther Method, proposed by Johnson and Johnson in 1991, (c) the Structural Approach, proposed by Kagan in 1992, (d) Jigsaw, discussed by Aaronson in 1978, and (e) the Student Team Learning approach, designed by Slavin in 1990. In the context of Writing as pr ocess approach, the collaboration models are mainly researched when

PAGE 83

57 students compose together or when they provide feedback to each other analyzing and discussing already produced drafts (Ferris, 2003; Ferris & Hedgcock, 1998; Liu & Sadler, 2003; Nelson & Carson, 2006; Storch, 2002, Ware & Warschauer, 2006). Peer Feedback Researchers agree that collaboration encourages students to approach their work adopting the role of interested readers, commentators, elaborators, and evaluators (Liu & Sadler, 2003; Sto rch, 2002, Ferris & Hedgcock, 1998). Furthermore, students receive and provide feedback to their peers, thus assuming multiple roles in the writing process. This unique character of collaborative composition process gives writers the opportunity to approac h the text from different perspectives and act on different levels of competency (De Graaf, Jauregi, & Nieuwenhuijsen, 2002; Dochy, Segers, & Sluijsmans, 1999). Based on their study of university level students in a second year Spanish as a second langua ge class, De Graaf et al. (2002) found that when writing collaboratively, students need to participate continuously in a negotiation process, reflecting on, analyzing, clarifying ideas and strategies. De Graaf et al. also reported that the students partici pating in their study felt that their collaboratively created texts were written better than were the ones they wrote individually. In the same study, the teachers shared the perceptions about the higher quality of the texts as well. However, the researche rs found In the context of collaborative writing, texts are composed in th e process of mutual efforts of two or more students and the contribution of the individual writer is difficult to assess; however, researchers found that individual writing supported by peer

PAGE 84

58 feedback promotes collaboration and has significant benefits (Fer ris, 2003). Peer feedback practices in L2 educational settings became widespread because of their consistency with the principles of Writing as a process Approach (Ferris, 2003). Both practitioners and researchers shared this interest in peer feedback; thu s, peer feedback in L2 writing was not only widely adopted in L2 writing classrooms but also became the center of numerous research studies (Ferris, 2003; Ferris & Hedgcock, 1998; Nelson & Carson, 2006; Silva & Matsuda, 2001). Dynamics and Patterns of Pee r Feedback Ferris (2003) points out that the dominant theoretical perspectives of peer response research approach writing as a socially constructed activity that requires audience consideration, self direction, and critical reflection. In their review of r esearch on second language peer response, Ferris and Hedgcock (1998) outline peer review and feedback as a complex process that incorporates a variety of interaction types. nd collaborative; these stances incorporate various categories of comments such as generic, critical evaluations, critical evaluations and suggestions, and critical evaluations and extended suggestions (Mangelsdorf & Schlumberger, 1992, cited in Ferris & H edgcock, 1998). Further, researchers (Lockhart & Ng, 1995) linked the stances used in the feedback, personalities, and the nature of the feedback task. Ferris (2003), in her summary of peer feedback research, outlined the following findings: (a) there were complex interaction processes that took place during providing/receiving feedback; (b)

PAGE 85

59 the major stance types of peer feedback identified were prescriptive, interpretive, a nd collaborative; with the collaborative stance being related to higher achievement; (c) the stances in the peer feedback process were further extended and described as authoritative, interpretative, probing, and collaborative; it was concluded that the au thoritative and interpretative stances contributed better to the peer review/writing process; (d) while participating in writing activities that required peer feedback, students stayed on task and sometimes could assume roles that are counterproductive to the functioning of the group. In addition, Storch (2002) addressed the issue of the dynamics of feedback interaction. She studied the patterns of interactions of 33 students in an advance university Storch (2002) accounted for the way learners approached the task they were given, the roles they adopted in the process of interaction, the involvement and contribution of each learner. The patterns of dyad interaction that were depicted by Storch were th e following: collaborative, dominant/dominant, dominant/passive, and expert/novice. Two indexes were used to describe the specifics of each pattern: mutuality and equality. Mutuality was described by Storch as the level of engagement of each participant, w hile equality referred to the degree of control over the interaction. The dyads that demonstrated alternative views on the discussed problems, looking for mutual agreemen ts and resolutions demonstrating interaction of high mutuality and equality. When adopting dominant/dominant patterns of interactions, students expressed an inability to reach consensus or employed a high division of labor, which resulted in low level of m utuality. In dominant/passive dyads, the dominant participant demonstrated an authoritarian stance

PAGE 86

60 appropriating the task while the other participant took more passive role; thus, the level of mutuality and equality were fairly low, which resulted in less negotiation. Finally, when expert/novice patterns were observed, the expert participant, although taking control over the task, encouraged the novice to contribute to the collaborative process, which resulted in higher mutuality with lower equality level. Storch (2002) suggested that there were more evidences of knowledge transfer when collaborative and expert/novice patterns were observed. Incorporating peer feedback One of the major questions related to the student feedback process outlined by Ferris (20 03) and addressed by multiple studies is how the that approximately 53% of peer suggestions were used in the following drafts; further, 766). The benefits of collaborative and interactive feedback process are also evident in a higher percentage of suggestions incorporated in the fol lowing drafts while non interactive and defensive in its nature feedback process resulted in feedback that was not utilized (Nelson & Murphy, 1993). The type of revisions incorporated by students as a response to feedback made by peers and teachers addres s mostly the surface features of the text (Berger, 1990). Investigating the same issue, Connor and Asenavage (1994) noted that feedback sessions that focus more on surface issues may lead students to emphasize these features when revising. Therefore, the q uestion of feedback training was raised in the L2 writing research community. In relation to this issue, Berg (1999) outlined the importance of

PAGE 87

61 students to address no t only the surface features of writing such as grammar and spelling but also to share considerations that influenced text meaning and organization. Based on their review of peer feedback research, Ferris and Hedgcock (1998) suggested principles of effectiv e peer response activities. These principles include incorporating peer response into the writing course and modeling the process of feedback provision in order to support the development of skills that would allow carrying on an effective peer response in teraction. The peer feedback process should sustain various peer response activities that would be designed with consideration of the individual student giving and inco rporating feedback. More recent research on peer response considers the channel through which this response is delivered. Liu and Sadler (2003) studied the process of peer response in nous FTF and delayed paper asynchronous feedback and multi user domains object oriented (MOO) synchronous feedback. The researchers found that the incorporated feedback was much highe r for the traditional groups, while the technology enhanced groups provide a greater amount of feedback. They also concluded that in the traditional mode, feedback was more effective when commenting synchronously, while the technology enhanced group worked more effectively when using the Word editing features. technology enhanced feedback process is the fact that there is little research that

PAGE 88

62 compares the effectiveness of tra ditional peer review with electronic peer review, as well as the fact that the rapid computerization of university classrooms calls for better understanding of the processes that take place in technology enhanced environments. The use of computers for prov iding peer response and teacher feedback is in the focus of current research (Milton, 2006; Ware & Warschauer, 2006). These research trends, along with the recognition that in the contemporary society, the new technologies transform writing and more genera lly literacy practices (Warschauer, 1999), guide the following section of this chapter which will address the impact that the new technologies have on education and L2 teaching and acquisition. New Literacies and SLA In this section, literature related to incorporating computers into the processes of teaching and acquisition of L2 literacy and writing is reviewed in order to situate the current study within the discourse of the new literacies and SLA. First, an outline of the contemporary understanding of literacy is presented. This presentation is followed by discussion of literature that focuses on computer mediated communication and its influence on literacy and writing in an L2 context. The contemporary views on literacy place it within the sociohisto rical context in which the literacy practices evolve, and are taught, acquired, and applied (Simpson, 2005; Warschauer, 1999); thus, literacy is viewed as a discourse that reaches beyond the ability to read and write; rather it is related, as other discour ses are, to ways people choose to communicate with each other, express their deeds, and social group associations, their styles, and interests (Tyner, 1998). Graff (1981) formulates some common conclusions about literacy based on research: (a) literacy is embedded within history; (b) literacy is a

PAGE 89

63 complex phenomenon both on theory and practice levels; (c) literacy incorporates social assertions and expectations as they vary in cultures and historical periods; (d) there are many if not infinite forms of lite racies; (e) there are continua of literacies; (f) mastering of alphabetic literacy is not simple; (g) the individuals may take multiple paths to promote, teach, and acquire literacy; (h) society may take multiple paths to promote, teach, and acquire litera cy; finally, (i) different forms of literacies coexist. These conclusions support the better understanding of fast emerging new literacy forms in the modern society. Western society is reshaped by fast emerging new technologies (Lindenau, 1984), these te chnologies in turn, reshape our understanding and application of literacy practices (Kasper, 2000; Tyner, 1998; Warschauer, 1999, 2006, Ware & Warschauer, 2006). Moreover, literacy plays the role of a gate keeper for accessing and using these practices (Wa rschauer, 2006): the digital, information oriented society requires literacy skills and competences that are closely related to the technology used for receiving information and communication (Kasper, 2000). This interrelation of technology and literacy is not new, technology and literacy have always been in constant interaction influencing and reshaping each other (Leu & Donald 2000). action of literacy and technology. While in the pre Gutenberg era writing mainly involved transcription of memorized orally composed speeches and sermons as well as copying manuscripts, the invention of the printing press in 1450 shifted the scholarly read ing and writing towards gathering and comprehending information from different sources. The availability of printed material also changed

PAGE 90

64 education, allowing for new ways of conceptualization in the learning process. The mass industrialization of the 20th century added new dimensions to the Western concept of Warschauer, 1999, p. 3). Further, L eu & Donald (2000) point out the importance of the change in the forms and functions of literacy. Today, changing technologies for information and communication and cha nging envisionments for their use rapidly and to the point where technolog y is being perceived as a new form of life (Aronowitz, 1992); the extent to which technology is altering modern life is determined by social and economic contexts (Warschauer, 1999). In Western society, technology decentralized the workspace changed the n ature of communication skills; it defined the fast and effective access to and processing of information as a key to success (Leu & Donald, 2000). Technology changes the classroom as well, researchers and practitioners point out that the modern pedagogical practices should aim for the development of literacies that are coherent with the societal transformations and would allow learners to fulfill their goals and dreams (Kasper, 2000; Warschauer, 1999). The Internet is one of the most important technological developments that is affected by modern literacy practices, and affects in its turn the literacy practices and the nature of literacy as a whole. In this context it is important to distinguish the computer mediated communication and the World Wide Web as literacy and information

PAGE 91

65 sharing environments (Warschauer, 1999). Following is a review of research that aims to reveal the specifics of the interaction and language learning processes that are supported by and situated in CMC environments. CMC and its Ro le in the Writing Process Warschauer (2004b) points out that constantly developing new technologies, along with the wide accessibility of personal computers, altered the reasons and the ways people write. Further, Waschauer states that these changes are so dynamic in their nature that it is difficult to document and analyze them. He views such an attempt for analysis as crucial for our understanding of how computer mediated communication can be incorporated into the language teaching process. Computer media ted communication (CMC) includes three main modes: synchronous, asynchronous, and hypermedia authoring (Warschauer, 1999, 2004b). Because the synchronous and asynchronous communication is in the focus of this study, the research on use of asynchronous and synchronous computer mediated communication in teaching writing is presented further. CMC has the unique nature of incorporating features of written and oral discourse. d that oral discourse requires a feedback loop that permits modification and correction synchronizing the meaning constructed during the act of communication. This provides basis for shaping the world in accordance to the language used. Therefore, as Kapla n (1987) states, even if the L2 speaker does not possess a complete inventory of sociolinguistic alternatives specific for the language in use, the feedback loop in oral communication will allow for modification and correction. Kaplan also points out that the speed of oral communication does not allow for extensive modification and planning,

PAGE 92

66 n. The above outline of oral written discourse characteristics allows scholars to conceptualize research directions and findings in the field of SLA: CMC, due to the unique combination of oral and written discourse characteristics, could provide greater support in SLA discussion (Blake, 2000; Swaffar, 1998; Warschauer, 1999). Moreover, Warschauer (1999) stresses the fact that electronic literacy becomes an important aspect in the contemporary society that is essential to be addressed in second language cl assrooms. Swaffar (1998) summarizes the benefits of CMC for second language learners concluding that in CMC, second language learners engage in class interaction more frequently and with greater enthusiasm, producing greater amounts of language. Blake (200 0) notes the importance of considering tasks that learners are assigned while communicating via the computer. He points out that the task typology developed by Pica et al. (1993) can be applied in computer mediated communication environments. In his 2000 s tudy, Blake validates task typology in a computer mediated communication environment with Spanish language learners. He also stresses the importance of further studies that would reveal how tasks are implemented and viewed by instructors and learners in co mputer mediated communication environments. CMC in the Second Language Classroom There are three main aspects of communication in the focus of researchers when approaching CMC in second language context: (a) how students participate in CMC environments i n terms of direction of their participation; (b) what is the target language output; and (c) what is the quality of the target language used in CMC interactions

PAGE 93

67 (Ortega, 1997). Warschauer (2004) outlines similar areas of interest in his review of research on computer assisted classroom discussion. In the next two sub sections, published findings in the areas related to asynchronous and synchronous CMC are presented: the Dynamics and Productivity subsection address the quantity and direction teraction, and in CMC: Language Use research findings about the quality of the target language output are presented. Dynamics and Productivity Researchers have found that the network communication environment creates a special communicative and linguistic community that differs in a number of ways from the face to face classroom environment (Beauvois, 1997; Kelm, 1992; Waschauer, 2004a). Networks provide a communication space that allows for noticing input and planning the output (Warschauer, 1997). The CM C environment allows for more opportunities for equal communication (Sullivan & Pratt, 1996; Warschauer, 1996). In this environment the communication process is more student centered, compared to the face to face classroom: students can communicate in thei r own pace without waiting for guidance rather than models, thereby empowering the students to control their own discussions and develop cooperative relationships (Sotillo, 2 000; Warschauer, 1996, 1997, 1999). Learners who have introvert communicative styles are more willing to communicate compared with the traditional face to face (FTF) classroom where more vocal students often overtake the discussion. Students report feelin g more comfortable when using CMC and have positive attitudes towards this mode of communication

PAGE 94

68 (Beauvois, & Eledge, 1996; Kelm, 1992; Kern 1995). This increases the quantity of student participation as compared to FTF discussion: several studies reported three reported the total amount of student participation to be 92%, in the two classes Kern (1995) reported that the student participation was 86% and 88%, Sullivan and Pratt (1996) reported 85% student participation. The increase of student participation reported in all three studies cited above is significant considering the fa ct that the face to face participation, as measured by these three researchers, varied between 35% and 60%. Although these results pertain to synchronous computer mediated communication (SCMC) environments, similar processes were observed by researchers w ho studied asynchronous computer mediated communication (ACMC) and, more specifically, e mail in language classroom settings. Hartman, Neuwirth, Kiesler, Sproull, Cochran, Palmquist, and Zubrow (1991) found that students communicated more when they had the opportunity to use e mail. Wang (1996) in a comparison of traditional paper pencil and e mail journal writing, found that students used longer sentences in their e mail journals and asked more questions than did their peers who wrote the traditional form of journal. These results suggest that CMC dynamics increase the opportunities for student collaboration, Warschauer (1996, 1999, 2004b) points out that although computer assisted classroom discussion is not the only way to involve students in collaborativ e activities, it is an effective way.

PAGE 95

69 CMC: Language Use Not only is the pace and dynamics of the interaction different in a CMC environment, the language that students produce differs as well. Considering that language learning as well as communication are complex processes, it is not surprising that the results of studies aiming to investigate the establishment of language skills when supported by ACMC and SCMC have been mixed. On the other hand, researchers have found that in CMC, students use simple sent ences more frequently than complex ones (Chun, 1994; Kern, 1995). On the other hand, Warschauer (1996) observed language of higher syntax complexity and lexical range. Despite these contradictory findings, the majority of researchers confirm the positive c ontributions of CMC to the process of language learning (Liu, Moor, Graham, & Lee, 2002). Based on their research, Beauvois (1994) and Kelm (1992) report less frequent code switching (altering the languages used in the communication process). Kelm (1992) a communication. Ortega (1997) suggests that in a computer mediated environment, students produce greater amount of language and communicate using longer sentences. In the ir study of adult German language learners, St. John and Cash (1995), who analyzed the learning process that took place in a six month e mail interaction of and use new vocabulary and structures. Another important characteristic of computer mediated communication is that it provides an advantage for language learners. More specifically, CMC gives the opportunity to participants of the communication act to go back and re view their

PAGE 96

70 communications (Beauvois, 1997; Warschauer, 1999, 2004b). It could be concluded that CMC creates an environment that allows learning the language, learning about the language, and learning through the language. CMC: Synchronous vs. Asynchronous In the contemporary L2 classroom, the positive role of CMC is recognized; however, in order to support L2 educators in the process of implementing CMC into their classroom, it is important to consider the specifics of the two different modes synchronous and asynchronous. Researchers found that both modes provide benefits to foreign and second language learners (Beauvois, 1994, 1997; Kern, 1995; Warschauer, 1999). Sotillo (2000) points out that qualitative and quantitative differences between the language outcomes of the two modes exist. These differences should be taken into account when planning to incorporate CMC into the curricula so the CMC activities correspond with the overall goals of the curriculum as well as with the objectives of the specific ins truction activities. SCMC interaction is highly dynamic and student controlled. This nature of synchronous interaction supports the development of student centered environment and active meaning negotiation (Sotillo, 2000). Further, CMC provides the lear ners with opportunities to participate in a communicative exchange in the target language that is at their level of language proficiency, in other words CMC supports the delivery 5). The intense social interaction, textual meaning construction and negotiation in synchronous discussion are crucial for learning and development of higher order cognitive functions and support the establishment of communities of learners.

PAGE 97

71 Sotillo (2000 ) found that asynchronous interaction offers similar advantages. In delay of the interacti on changes its nature, making this interaction more like the dynamics of face to face classroom interaction settings. However, one should be cautious making such generalizations about the nature of CMC since the types of communicational tasks performed can significantly alter the nature of communication (Blake, 2000). The mode of CMC also affects the functional use of language (Sotillo, 2000). The fast pace of the synchronous interaction makes it more similar to communication in informal oral acts, produc ing shorter electronic utterances. Sotillo (2000) reports evidences of error self correction and peer correction in this mode of interaction. Asynchronous interactions lead to more syntactically complex language output through the use of subordinate clause s and longer sentences that are an indicator of active cognitive processes involved in text construction. This language production could be the result of a number of factors such as having more time to construct the message, considering the audience expect ations, and the nature of performed tasks. Including challenging tasks, asking students to post carefully prepared responses would encourage them to think critically and focus on both meaning and form to a greater extent than happens during synchronous com puter mediated communication interactions. Summary design better instructional strategies, writing should be approached as a process situated into a social context (Grabe & Kaplan, 1996; Kent, 1999, Matsuda, 2003). In addition, it

PAGE 98

72 should be taken into account that textual level writing features contribute to this understanding (Grabe & Kaplan, 1996; Snellings, van Gelderen, & Glopper, 2002). Accounting for the social nature o f writing allows for the recognition of the communication events that occur among learners and between learners and the instructor and their impact on text construction. It is essential to note that the communication events differ in various discourses suc h as oral, written, and CMC (Swaffar, 1998; Warschauer, 1999). In the contemporary SLA research, it has been concluded that closer studies of CMC tasks and activities are important in order to assure effective inclusion of this unique communication environ ments in the L2 educational settings (Blake, 2000). Further, while interactions during the revising and editing stages of the writing process have been in the focus of writing research for more than a decade now (Ferris, 2003; Grabe & Kaplan, 1996; Hyland & Hyland, 2006a), research that would shed lights on the first two stages of the process, pre writing and drafting, is limited. In order to address the new trends in the writing research and the existing gap in the literature, the research project presente d in this dissertation aims to investigate the specifics and dynamics of these stages of the writing process within the context of computer mediated interaction.

PAGE 99

73 CHAPTER III: METHOD This chapter presents the design of the current research study justi fying the basis for making the overall and the specific design decisions. The methodological design is nature of the research questions, settings, and data addressed. T he chapter also describes the participants, outlines the instruments used for data collection, the process of data collection, as well as the data management procedures applied. Further, a description of the data analysis procedures employed is provided. Overview of the Study Design This study is guided by the following overarching research question: How can computer mediated communication influence the writing skills development of English as a second language adult learners? In order to answer this quest ion, a mixed method study was conducted (Onwuegbuzie, 2003; Teddlie & Tashakkori &, 1998, 2003). It aimed at investigating how a two way communicative pre writing task performed via asynchronous and synchronous computer mediated communication supported adu lt writing and first draft writing perceptions of CMC, CMC modes, and the two way communication task perform ed during synchronous and asynchronous computer mediated interactions. During the process of investigation in the current study, data were collected and analyzed in two stages: qualitative and quantitative. The qualitative stage consisted of

PAGE 100

74 two sub stage s, namely a case study of the entire group of the intermediate English language learners enrolled in the targeted IEP and a collection of eight instrumental case studies. The qualitative findings were complemented by the quantitative findings; thereby allo wing for comprehensive understanding of the phenomenon under investigation. This design, as presented graphically in Figure 1.1, can be defined as a parallel component mixed method design with an ongoing dominant exploratory qualitative stage and nested le ss dominant quantitative stage of an experimental design (Caracelli & Greene, 1993; Teddlie & Tashakkori, 2003). The specific educational setting, the research paradigm, and the data collection eoretical orientation, philosophical beliefs, and the research questions of the current study (Seliger & Shohamy, 1989). Thus, two focal issues were considered in the design. The first one was the research focus specified prior to conducting the study and the research questions formulated based on this theory. The second issue was how the research could be When considering the first issue, it is important to state that the dominant quali tative stage extended t hrough multiple case studies was guided by a research orientation, which allowed formulating the research questions, focusing the inquiry, determining the type of data collected, as well as determining the specific steps for data col lection and the applied data analysis strategies (Yin, 2003). The steps for data collection and analysis flow are presented in Figure 3.1. In this figure, the graphical representation of a multiple stage study flow used by Yin (2003, p. 50) has been adapt ed.

PAGE 101

75 Figure 3.1 Research Flow The overarching research question guided this multiple stage study developed within the framework of Writing as a Process approach discussed in detail in Chapter II. The researcher proposed that the quantitative analysis of first drafts and the group case qualitative analysis pertaining to the whole setting and extended with eight instrumental case studies would show how CMC supported the writing development of ESL adult learners at pre writing and first paragraph writing stages. More specifically, the researcher proposed that the quantitative analysis of first drafts would reveal how some of the key text features were influenced by the CMC pre writing exchange. Further, in the

PAGE 102

76 qualitative stage, a case study of the whole g roup of intermediate level ESL learners enrolled in the IEP was extended through the instrumental case studies. The goal of this stage was to show how the CMC pre class interaction, and iting development. In the context of the second issue considered in the process of study design, namely how the research could be conducted in a way that was consistent with the al views of the world that is constructed of multiple subjective realities (Dewey, 1964; Maxcy, 2003). The researcher also believes that studying interactions in a social context supports the understanding of the multiple subjective realities and reaching agreement on them. These beliefs directed the researcher towards applying a combination of design approaches that would reveal a kaleidoscopic picture of the studied pheno menon and elucidate its various aspects. Thus, the paradigm that supported the researcher in the process of addressing the pragmatist paradigm. This paradigm guides resear ch to the application of mixed method studies through a planned and justified combination of quantitative and qualitative methodologies allowing for a better understanding of the studied phenomenon (Maxcy, 2003). Quantitative Stage Design Overview The qua performance after completing a communicative two way pre writing task, as described and validated by Pica et al. (1993) via CMC (synchronous versus asynchronous modes).

PAGE 103

77 Specifically this s tage investigated the influences of pre writing task CMC mode on the quality of ESL writ ing as measured by assessing: (a ) the syntactic complexity (measured by calculating the mean length of T units), (b ) the amount of information present in a single focus (measured by mean length of id ea units), (c ) the quantity of overall information present (measured by the number of idea units) (d ) lexical information per clause (measured throu gh lexical density analysis), (e ) vocabulary complexity (measured by analyzi ng the frequency of the unique words used), (f ) and overall quality of writing using multiple trait rubric accounting for rhetorical soundness, presentation and development of main ideas, and overall language use. In this stage, the following research que stion was posed: How does the CMC mode (synchronous vs. asynchronous) used for the completion of a communicative two way pre This research question was answered through analyzi writings, assessing the criteria discussed above, and answering the following specific questions: Research Question 1: What is the difference in the syntactic complexity present in the post treatment paragraphs of the students who participated in the synchronous versus asynchronous CMC pre writing discussion? Research Question 2: What is the difference in the amount of information present within a single focus of the post treatment paragraphs of students who participa ted in synchronous versus asynchronous CMC pre writing discussion?

PAGE 104

78 Research Question 3: What is the difference in the quantity of the overall information present in the post treatment paragraphs of the students who participated in synchronous versus asynch ronous CMC pre writing discussion? Research Question 4: What is the difference in the lexical information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Ques tion 5: What is the difference in the vocabulary complexity present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Question 6: What is the difference in the rhetoric al soundness present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Question 7: What is the difference in the presentation and development of the main point and its support present in the post tre atment paragraphs of students who participated in synchronous versus asynchronous CMC pre writing discussion? Research Question 8: What is the difference in the overall language use present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? The study was conducted over the course of two semesters with a total of 60 participants. The quantitative stage took place during Weeks 5 and 6 of each the semes ter. In this stage, the students prepare d for writing a paragraph by completing pre writing activities, namely brainstorming and planning, preformed in randomly assigned dyads, these activities were performed in a face to face environment. Immediately afte r

PAGE 105

79 this pre writing session, they wrote their paragraphs (referred henceforth as pre treatment paragraph) using computers. Then, each dyad was randomly assigned to two treatment groups: asynchronous computer mediated communication (ACMC) group or synchrono us computer mediated communication (SCMC) group. The ACMC dyads used asynchronous computer mediated communication, while the SCMC dyads used synchronous computer mediated communication to perform a communication task (referred henceforth as a treatment tas k). This task aimed at preparing the participants for writing their post treatment paragraphs. After the task was performed, each student wrote individually the post treatment paragraph. The quality of their pre treatment and post treatment writings was as sessed based on the criteria described further in this chapter. Qualitative Stage Design Overview In the on going qualitative stage, the researcher aimed to explore the processes under investigation for the overall group of participants in the study and th en focus on extreme cases identified in this group. Thus, the design of this stage wa s exploratory (Yin, 2003). The qualita tive stage of the study consisted of two sub stages: the first one followed a group case study design (Miles & Huberman, 1994) that p resented the whole group of participants in the context of the learning and communication environment and then addressed specific variables. The goal of this stage was to identify and analyze processes and their outcomes within a group of language learners accounting for specific language learning conditions in which the language learning occurs; thus, a community of practice was in the focus of this group case study (Miles & Huberman, 1994). The qualitative stage was further extended through a second sub s tage: eight individual case studies were conducted that allowed focusing on extreme cases identified

PAGE 106

80 within the group. The instrumental case study methodology was used to build the eight individual case studies with unit of analysis being intermediate ESL adult learners who demonstrated the highest and the lowest gains in their writing after participating in a CMC pre writing task. After building each individual case, the cases were cross opted. He describes an instrumental case study as a study that examines a particular case aiming to provide an understanding of an issue or to refine a theory; t hus, the case itself facilitated the understanding of the studied phenomenon (Stake, 1995). Qu alitative Sub stage I: The Group Case Study The participants in the group case study were the 60 students who took part in the quantitative stage. This stage of the study aimed to investigate a group of purposefully selected participants (Hatch, 2002; Kemp er, Stringfield, & Teddlie, 2003) within the context of the studied phenomena. The data were analyzed in order to reveal the overall group participation in the writing class with focus on their performance duri ng the completion of a computer assisted pre writing activity. In this stage, the following variables related to the studi ed phenomena were considered: (a ) the descriptive characteristics of the CMC pre writing discussions p erformed by the participants; (b ) the patterns of dyadic interactions used by the participants in C MC environments; (c ) participa writing task as performed in a CMC environment. They were presented within the context of the learning environment. Multiple data sources were used in this stage. Observations and debriefing with the non student participants were conducted to collect data related to class organization,

PAGE 107

81 dynamics, management, and working with software. The CMC logs produced during the pre writing task were us ed in order to address the descriptive characteristics of the pre writing discussion. The patterns of dyadic interactions during the CMC discussions were identified applying the model of dyadic interaction proposed by Storch (2002). Finally, using a combin ation of the standardized open ended interview approach (Patton, 2002), writing task as performed in a CMC environment were elicited. The leading research question addressed was: How do dyadic interactions performed via asynchronous computer mediated communication (ACMC) and synchronous computer mediated communication (SCMC) pre views on CMC and writing influence writing development of adult intermediate level ESL learners? Thre e specific research questions helped to guide this analysis. Research Question 1: establishing their writing skills? Research question 2: What patterns of dyadic interaction do partici pants manifest during the asynchronous and synchronous CMC interaction process? Research Question 3: What are the factors that influence the CMC pre writing interaction process? How do these factors influence the interaction process? The goal of this stag e was to provide a better understanding of the studied group of learners and the learning environment within the context of the studied phenomena which facilitated the following instrumental case studies. Qualitative Sub stage II: The Instrumental Case S t udies In the instrumental case study stage, the researcher identified eight participants who demonstrated the highest and lowest gains in their writings after performing the

PAGE 108

82 CMC pre writing task. Each case was studied individually and then the cases were c ross examined. The analysis of the cases and their comparison aimed to disclose (a ) the factors that influenced the CMC pre writing interaction process for each partici pant and across participants; (b ) the patterns of dyadic interaction that the actor of e ach case manifested during the asynchronous or synchronous CMC interaction process and how these patterns were affected by the factors that in fluenced the CMC interaction; (c ) the f the writing skills in particu lar; (d ) the ways in which the actors used the specific language constructs and ideas produced during the asynchronous or synchronous CMC interactions in their writings. These comparisons were viewed as instrumental to expla in (1) how ESL students of intermediate level proficiency who had the highest and the lowest gains in their post treatment writings participated in the communication activity performed via CMC, and (2) how they applied the results of the peer CMC interacti on in their writing. In order to understand this complex process, data collected during on going in class observations, the pre treatment and post treatment writing samples, and CMC logs were analyzed. e communicative task elicited during interviews were considered. The research questions addressed through these case studies were: Research Question 1: How do peers participate in asynchronous and synchronous pre writing CMC interactions? Research Question 2: (a) How do they use the specific ideas and language generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas?

PAGE 109

83 The setting in which the study was conducted was perceived t o be naturalistic the ESL classroom. Although every effort was made to collect the quantitative data in a controlled manner, this control was not transparent to the participants, and the overall quantitative data collection process did not differ signifi cantly from the usual classroom and overall class participation and were linked to the results of the quantitative stage of the study in order to approach the process holistically Figure 3.2 represents the data collection schedule of qualitative and quantitative stages of the study.

PAGE 110

84 Figure 3.2 Research Stages and Steps Weeks 1 2 3 4 5 6 7 8 9 10 11 12 13 14 qant *Pre treatment writing; *Assigning ACMC and SCMC groups; *ACMC homework (due the beginning of week 6) *SCMC task; *Post treatment writing. QUAL Observations and document data collection Step 1 Non Student Participants Meetings 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Inte rview 1 2 3 4 5 6 7 8 9 10 11 12 13 14 CMC Interviews: Participants -all students Observations of the language learning setting CMC logs and writing: data collection Step 2 Step 3 Step 4 Post data collec tion period QUAL stage extension: Eight instrumental case studies

PAGE 111

85 Pilot Study The researcher conducted a pilot study in the Fall semester of 2004. The goal of the pilot study was to validate the scores obtained by the m ultiple trait rubric used in the current study and described further in this chapter, as well as to facilitate the design process of the interview questions and data collection procedures. Another important objective of the pilot study was to ensure that t he interface of the CMC applications used for the treatment dyad collaboration allowed for smooth navigation and flow of the CMC interactions between dyads of participants. The software applications that were used for collecting ACMC and SCMC data during the qualitative and quantitative stages of the study, namely Internet Classroom Assistant, version 2(ICA2) and Language Education Chat System (L.E.C.S) were tested during the pilot study. The Internet Classroom Assistant is an online application that allo ws asynchronous conferencing, document posting, and link sharing, Figure 3.3 presents the ICA2 class home page (Nicenet, 1998).

PAGE 112

86 Figure 3.3 ICA2 The application used for the SCMC interaction and data collection is Language Education Chat System (L.E.C.S ): a free web based synchronous conferencing software designed especially for language teachers and learners. Figure 3.4 presents a chat session entry screen of this application (Kanto Gakuin School Corporation, Japan n.d.). Figure 3.4 Language Education Chat System (L.E.C.S.)

PAGE 113

87 During the pilot study, it was interrupted by technical and navigational difficulties. Furthermore, it was concluded that in class and homework activities conducted prior the trea tment had been beneficial in terms of establishing the necessary computer skills that would assure the productive use of the applications. Both applications are discussed in more detail later in this chapter. Settings, Participants, and Sampling Procedure s Study Context The ESL Program The overall sampling technique used to select the IEP among the numerous programs existing in the United States wa s purposeful sampling (Hatch, 2002; Kemper, Stringfield, & Teddlie, 2003) which aimed to study language learn ing processes within contexts common for IEP settings in the United States. The researcher decided to collect data at this setting based on the fact that the targeted program passed the Commission on English Language Program Accreditation review. This accr editation assured the researcher that the language learning context of the program was common to other accredited IEPs in the country and that in the program similar enrollment and proficiency level assessment procedures were utilized. A careful study of the enrollment and level placing procedures employed at the targeted IEP as well as its teaching practices and the philosophy confirmed that the sample drawn from this program was representative to a significant extent to the overall IEP population in the USA. Thus, the participants in this study were chosen among the students of this IEP. They were intermediate adult ESL students who spoke various

PAGE 114

88 native languages, had different cultural backgrounds, and were enrolled in an IEP offered at a large urban are a university in the southeastern part of the United States. As outlined in the mission statement of the targeted IEP, the teaching and learning practices encouraged and implemented in this language school aspired to prepare its English language learners fo r successful participation in the U.S. academic environment and support them in the process of their professional development ( ELI @ USF, 2002). The continuous professional growth of the instructors was supported by the IEP administrators through organizin g professional discussion forums and workshops, providing a CALL support during the semester by a specially hired CALL consultant, as well as through providing connections and opportunities for active scholarly partnership with the university graduate prog rams in Second Language Acquisition and Instructional Design and Applied Linguistics. The curriculum designers and instructors at the IEP were well informed about the development of the current trends in SLA and applied modern teaching approaches including a carefully planned implementation of a computer assisted language learning curriculum for all proficiency levels. This curriculum aimed to develop computer literacy skills, as discussed in Chapter II, that involve effective use of computer accessible med ia for language learning, research, and communication. In addition, the goals of the writing course targeted in this study were formulated within the framework of Writing as a process approach, which was the guiding theory for the current study. For the fu ll text of the Intermediate Academic Writing course goals see Appendix 1 A. The program required the ESL students to be enrolled on a full time basis receiving 25 hours of weekly instruction during a 14 week semester or to be part time

PAGE 115

89 students with a mini mum of 10 hours of weekly instruction during a 14 week period. In addition, there was a second enrollment after mid terms (Week 7 of the semester). The students who took classes on a part time basis with less than 15 hours of weekly instruction or were adm itted during this second enrollment were not included in the study. Another requirement at the IEP was the strict placement testing administrated at the beginning of each semester. A combination of two placement tests was used: (a) Listening and Grammar se ctions of the Comprehensive English Language Test (CELT) (Canadian Test Centre Inc., 2003) and (b) Vocabulary and Reading Comprehension sections of the Michigan Test of English Language Proficiency (MTELP) (The University of Michigan, English Language Inst itute, Testing Certification Division, 2003). For more details on CELT and MELTP see Appendix 1 B and Appendix 1 C. At the beginning of each semester, after the proficiency levels of the students had been identified, they were assigned to the correspondin g proficiency level sections by the language and gender. The goal of applying these criteria was to assure native language and gender diversity in each section. In orde r to assure quality instruction the IEP administration limited the section size to a maximum of 18 students. Each intermediate level group participated daily in the following classes: Listening/Speaking and Pronunciation, Academic Writing, Grammar, Cultur al Contacts with a focus on reading, and Test Preparation Electives: the Test of English as a Foreign Language (TOEFL), Scholastic Aptitude Test (SAT), Graduate Record Examination (GRE), or Graduate Management Admission Test (GMAT) Furthermore, the classe s of a

PAGE 116

90 particular proficiency level followed the same syllabi and used the same textbooks, although different instructors would usually teach the level sections. The instructor meetings for each level were conducted on a regular basis during the semester t o assure frequent instructor communication. In addition, the instructors of the same class in each level coordinated their instruction throughout the semester to assure that the same or similar topics were taught. Last, but not least, the IEP was designed as a research laboratory for graduate students involved in language acquisition related programs, and the administrators were not only willing to support research studies but also had the required expertise. Initially, a second IEP was considered for the study; however, this consideration was abandoned because of the concern that the differences in the curricula, course syllabi, and textbooks may negatively affect the results. That is, it was assumed that the use of a second IEP would threaten the validity of the findings via differential selection of participants (Onwuegbuzie, 2003). T he administrative support and expertise, the identical and strict placement of students into proficiency levels and the instructional procedures applied increased the interna l validity of the quantitative findings as well as the credibility and transferability of the interpretation s pertaining to the collected qualitative data. Instead, in order to obtain a sample whose size would allow the application of inferential statistic al analysis, the researcher decided to collect data in two consecutive semesters at the same IEP. This data collection strategy was perceived to pose less threat to the validity of the findings by diminishing the threat of differential selection of partici pants.

PAGE 117

91 The Writing Class The writing class at the target IEP was titled Academic Writing. During both semesters in which the data collection was performed, the class was taught for a total of five hours a week, consisting of three single sessions and on e double session, four days a week. During both semesters, the students spent two of these class sessions in the computer laboratory working on assignments that were part of the writing class. The instructors were educators who held their Master of Arts de gree in the field of Applied Linguistics. In addition to the strong linguistic and language teaching background, the participation in in service workshops and conferen ce participation. The textbook used in the writing class during both semesters was Writing to Communicate written by Boardman and Frydenberg (2002). The authors of the book designed it as a guide through a recursive non linear writing process that uses t he principles of scaffolding. This choice of textbook revealed the intentions of the writing curriculum designer to promote a writing course that would carefully guide learners and encourage active collaboration throughout the writing process. The research er has been a part of this ESL community for several years, she also systematically observed the writing classes during the data collection period (the entire Fall 2005 and Spring 2006 semesters) ; her experiences showed that the above intentions, were shar ed and implemented by the instructors of the writing course. During the Fall 2005 semester, the intermediate level had two sections instructed by the same teacher. During the Spring 2006 semester, there were four sections of Level III with four different i nstructors. The instructors were required to use the same textbook but were given the freedom to create

PAGE 118

9 2 their own supplemental activities, assignments, and assessments throughout the semester; the only compulsory assessment was the final examination admini stered at the end of the semester. However, as it was mentioned earlier, the tradition at this IEP was for the instructors of the level sections to communicate with each other throughout the semester, sharing ideas, activities, and assessment tools. The we ekly level meetings held on a regular basis throughout the semester further supported this communication. Furthermore, in order to facilitate the data collection process, during the Spring 2006 semester, the four writing instructors made extra efforts to u nify the timing and sequence of topics introduced in the different writing sections of Level III writing class. Activities that incorporated CMC had been part of the learning process at this institution for several years. These activities took place during the computer laboratory sessions and were assigned as homework. During the Fall 2005 and Spring 2006 semester, all writing instructors used ICA2 for their writing classes to organize material and communicate with their students. During the pre treatment p eriod, the researcher L.E.C.S. was a less frequently used application, the writing instructor in the Fall 2005 semester was familiar with it and successfully imple mented chat activities using the application. Prior to the treatment, three out of the four Spring 2006 writing instructors were not familiar with L.E.C.S. and two had limited exposure to and use of ICA2 which required individual training sessions, conduct ed by the researcher at the beginning of the semester. During both semesters, the researcher actively supported and participated in the design and implementation of the pre treatment CMC activities that aimed to build mediat ed communication skills.

PAGE 119

93 While designing the current study and during the data collection process, the researcher carefully studied the practices and the materials used in the writing course offered in the targeted IEP and implemented by the instructors du ring both semester of data collection. The researcher concluded that the instructional practices as well as the texts used in the course were consistent with the Writing as a process approach theory informing the study. It was also concluded that these pra ctices were considerably unified across the two semesters in which data were collected as well as across the different Level III writing sections. The learning and teaching prac tices and environment described allowed for collection of data without destruct ion of the naturalistic educational setting; in other words, the data were collected without imposing practices such as collaboration, scaffolding, writing of multiple drafts, and CMC, because they were a natural part of the learning process. Participants The participants in the qualitative and the quantitative stage of the study were 60 adult students who were enrolled in an intensive ESL program affiliated with a southeastern American university during the Fall semester of 2005 and the Spring semester of 2006. Although the total number of Level III students in the Fall 2005 and Spring 2006 semesters combined was 103, only the students who participated in dyads with a full set of treatment activities completed were considered as participants in the study an d their writings and CMC discussion data were analyzed; in other words, both members of the dyad had to have the pre treatment discussion and writing, the treatment and the post treatment writing completed. The majority of the participants (59) were full t ime ESL students who had 25 hours of weekly instruction; there was one part time

PAGE 120

94 student who had 15 hours of weekly instruction. There were 24 female (40%) and 36 (60%) male students. The average age of the participants was 25.55. The youngest participant was 17 years old, whereas the oldest was 46. They came from 21 countries and spoke 13 native languages. All of the new coming students had an intermediate level of language proficiency, as measured by CELT and MTELP in the beginning of the semester. The st udents who were with the IEP during the previous semester and successf ully completed Level II were not tested for proficiency in the beginning of the semester. Considering the fact that the program was similar to other IEP programs in the US and was follo wing common registration procedures, this sample may be viewed to be representative of the university IEP population in the US. Nevertheless, it should be taken into account that these students may not encompass the whole variety of different nations and c ultural backgrounds represented in the IEP population in the United States and other countries where ESL instruction is offered which posed a threat to population and ecological validity of the quantitative findings of the study. All students participate d in the initial phases of the qualitative stage (prior to qualitative data collection and analysis). They were observed during their face to face classroom interactions and weekly computer laboratory classes, and were interviewed as well. After the quanti tative data were collected and analyzed, the researcher focused on eight participants chosen for the instrumental case studies. These participants were chosen based on their gains in the quality of their post treatment writings as compared to the pre treat ment writings. These gains were measured based on the difference of the pulled z scores of the eight measures used for the

PAGE 121

95 paragraph analysis of the pre treatment paragraph and the post treatment paragraph. These participants are presented in Chapter V. T he non student participants in the qualitative stage of the study included the researcher who was also the CALL consultant at the IEP for the Fall semester of 2005, the writing instructors, and curriculum coordinator at the IEP. All of them had a master de gree in the field of Applied Linguistics, foreign and second language teaching and had considerable (more than 5 years) experience in teaching ESL. All of them except for the researcher were native speakers of English. Sampling Techniques Quantitative S tage Sampling Techniques The sampling technique applied in the quantitative stage of the study were purposive homogeneous sampling (Hatch, 2002; Kemper et al., 2003). The goal was to choose ESL learners of a specific level of proficiency. During the Fall s emester of 2005 and the Spring semester of 2006, 103 students were enrolled at the target proficiency level; however, only 60, for whom a full data set was obtained, as specified earlier in this chapter, were selected as participants in the study. The Grou p Case Study Sampling Techniques Similarly to the quantitative stage of the study, for the group case stage, the participants were selected using purposive homogeneous sampling (Hatch, 2002; Kemper et al., 2003). The students who participated in the quant itative stage were the participants of the group case stage of the study. These were students who were enrolled in two consecutive semesters in the Intermediate level writing class at the targeted IEP.

PAGE 122

96 The Instrumental Case study Sampling Techniques The s ampling technique used in the instrumental case study stage was purposive sampling involving an Extreme/Deviant sampling technique. The researcher decided to apply this purposive technique because it allowed for identification of the most outstanding cases based on writing gains as a result of the treatment (Kemper et.al., 2003); these cases were used for the creation of multiple case studies that were then analyzing, and compar ing a collection of cases rather than one or two separate cases would reveal better the processes under investigation. The participants in the instrumental case study stage were eight students selected from the overall group of participants. Four students from each CMC mode group participated in the qualitative stage of the study. This selection was based on their writing gain scores and CMC group participation. The gain was identified based on the difference of the pulled z scores of the eight measures u sed for the paragraph analysis of the pre treatment paragraph, completed after a face to face pre writing discussion, and the post treatment paragraph, completed after the CMC pre writing discussion. Thus, the low gain participants were the two students fr om the ACMC and two students from the SCMC group who obtained the lowest pulled z score difference. Their high gain counterparts were the two students from the ACMC and two students from the SCMC group who obtained the highest pulled z score difference. Th e Non student Participants The non student participants in the study were chosen via application of purposive criterion sampling (Patton, 2002). The criteria were either for those

PAGE 123

97 participants to be involved in the instruction process of the student partic ipant and/or to be involved in the curricula design and development of the IEP. The researcher believes that these participants, chosen using the predetermined criteria stated above, contributed to the qualitative data collection process. Protecting Parti During the study, the required Institutional Review Board procedures were followed. I n order to participate, all students of the Intermediate (Level III) writing classes were informed about the nature of the study. They were ask ed to sign an informed consent form indicating their agreement or disagreement to participate in the study. They also were notified that they can discontinue their participation in the study at any point. It was explained to the students that if they decid e d to discontinue their participation or if they decide d not to participate at all in the study, this would not free them from completing the communicative and writing tasks because they were be part of their overall writing class activities. The students also were informed that their privacy would be protected: it was explained to them that their real names were to be kept confidential would be used. Alphanumerical codes were assigned to each of the quantitative stage participants after the data were collected. These codes replaced the names of the students and were used in all tables presenting the data. Further, during the analysis and the presentation of the qualitative dat a pseudonyms were assigned. The researcher used names that were typical for the native culture and country of the participants.

PAGE 124

98 In this study, the researcher played multiple roles; these roles were defined by various facto and educational background. During the Academic Writing curricula revision and course syllabus creation process, the researcher collaborated with the non student participants in order to ensure that the CMC training and treatment tasks as well as the topics of the pre and post treatment paragraphs were c onsistent with the intermediate level Academic Writing course and were compatible with the overall goals of the course. This al lowed for preservation of the naturalistic character of the setting. The researcher introduced the instructors to the writing measurement instruments, namely the multiple trait rubric, and trained them how to use the rubric. In addition, as mentioned earli er in this chapter, the instructors who had limited familiarity with ICA2 and L.E.C.S. were trained how to work with these applications. Further, prior to the quantitative stage, the researcher discussed with the instructors the design of the in class bra instorming sessions for the pre treatment paragraph writing. In order to ensure that the instructors follow similar procedures, the researcher observed these in class sessions and took notes that supported the qualitative data collection, analysis, and int erpretation. In the Spring semester of 2006, one of the pre treatment and one of the treatment sessions were observed by a trained research assistant because two of the four sections of Level III had their classes in two different computer laboratories sim ultaneously. Along with the IEP instructors, the researcher scored the pre treatment and post treatment first drafts ; however, only the instructors provided feedback to the students for further improvement of their post treatment drafts.

PAGE 125

99 The researcher was an active participant and observer during the whole semester. During the pre treatment period, she observed and assisted activities that aimed to train the students how to work with ICA2 and L.E.C.S., she also assured that the training activities were sim ilar and/or identical across the two semesters and the six sections of Level III Academic Writing class. beliefs play in data collection and interpretation of results. Further, t he complexity of the personally involved in the processes under investigation and consequently becomes the key measuring instrument (Gall, Borg, & Gall, 1996; Pa tton, 2002). Thus, it is important The researcher is a native speaker of Bulgarian and Russian. She was raised in Bulgaria and acquired her second native language Russian first in home setting s and later, starting at age of 10, in immersion educational settings. She learned English in her Russian Culture and Linguistics, as well as a graduate certificate in Teachi ng English as a Foreign Language, and currently is pursuing her doctoral degree in Second Language Acquisition and Instructional Technology. During the study design period and the first semester of data collection, she was the CALL consultant in the IEP wh ere the study was conducted and was involved in teacher training and instructional support related to TESOL, ESL, and CALL. Her personal beliefs as an educator and researcher encompass supporting classroom practices that encourage creating communities of l earners,

PAGE 126

100 providing authentic opportunities for language learning, and promoting the development of new literacy skills. Research Procedures This study utilized a parallel component mixed method design. There were two stages utilized, namely: an on going dominant qualitative stage and nested less dominant quantitative stage sequential mixed method design (Caracelli & Greene, 1993; Teddlie & Tashakkori, 2003). The research was conducted with students who were intermediate level ESL learners at the targeted program. Qualitative Stage Procedures In the on going qualitative stage of the study, four steps were undertaken. The steps were identical during both semesters of data collection. In these steps, the researcher aimed to collect data presenting the speci fics of the learning environment, writing and opinions about the treatment pre writing task. Following is a detailed description of the procedures that were applied in each step. For a graphical presentation of each step within the Qualitative stage schedule, refer to Figure 3.2. Qualitative Stage, Step 1 : Prior the beginning of the semester, the researcher had informal meetings with the non student participants of the s tudy, namely: the writing class instructors and the curriculum designer at the IEP. The topics discussed during these meetings concerned: (a) the overall instructional and assessment procedures planned for the semester; (b) the educational goals set for t he semester; and (c) the CALL approaches planned in connection with the integrated CALL curricula and how they were situated within the overall learning process. The outcomes of these informal interviews

PAGE 127

101 were documented. The data collected during these ses sions were used when describing the overall research setting. During these sessions, the researcher informed the non student participants about the study goals and procedures. The researcher confirmed that the procedures planned were consistent with the e ducational practices at the IEP, and that they could be applied in accordance with the research plan. The researcher and the non student participants compared the writing class syllabi and research project plans and schedules in order to confirm that the r esearch procedures were applicable and would not disturb the planned instructional and learning process. During this step, the familiarity of the writing instructors with the proposed CMC software was discussed and additional support was provided to the in structors who were less familiar with the CMC applications. Qualitative Stage, Step 2: This step started prior to the quantitative stage and continued after this stage was completed. Starting from the first week of classes, the researcher observed the int ermediate level writing classes. She took the role of a participant observer who assisted the instructor during class activities and participated in these activities. On a regular basis, the researcher held a short informal debriefing sessions with the ins tructors, in order to verify and clarify (if necessary) the observations. These debriefing sessions with the non student participants were viewed as an important element of the data collection that provided opportunities for verification and deeper underst anding of the processes observed in the classroom. The field notes were transcribed immediately after each session and were transformed into a narrative that described the class session. These narratives allowed for achieving a more complete description o

PAGE 128

102 based both on the raw field notes and on what the researcher remembered (Hatch, 2002). It is important to note that the researcher participated in and observed selected class sessions focusing on the compu ter laboratory sessions; however, when the schedule allowed face to face sessions were observed as well. Detailed information about the number and type of observed sessions is provided in Chapter IV. During Step 2, that is, prior to and during the quantit ative stage, the students participated in unified activities introducing the use of computers for writing and research purposes following the CALL curricula guidelines and objectives. The outline of the CALL curricula objectives is available in Appendix 2. The participants were introduced to the applications used for ACMC and SCMC, namely ICA2 and L.E.C.S., and performed various communicative CMC tasks including participation in the first round of iarity with the interface of the CMC applications (ICA2 and L.E.C.S.) and their exposure to computer mediated interactions in synchronous and asynchronous environments, thereby, eliminating the novelty effect that could limit student performance during the treatment. These activities were planned by the instructors of each class and coordinated when necessary with the researcher who was also the CALL consultant at the IEP during the first semester in which the data were collected. They are described in deta il in Chapter IV. Another reason for introducing the CMC applications in Step 2, prior to the quantitative stage, was to preserve the naturalistic nature of the educational setting; in other words, to go along with the path that students would follow duri ng their study of English language, as outlined in the CALL integrated curriculum. However, it is important to state that a communication task of the same design as the one that was

PAGE 129

103 performed for the actual data collection was not applied during the pre qu antitative period. This decision was made based on findings in SLA research that reveal that students participating in unfamiliar design tasks tend to demonstrate greater conversational commitment and involvement, whereas task familiar students demonstrate d signs of disinterest that affects the intensity of their communication (Plough & Gass, 1993). In addition, in Step 2, prior to and during the quantitative stage, the students, following the writing curriculum, were exposed to the specifics and mechanics of descriptive paragraph organization and performed practice exercises. The instructors of the writing classes also introduced the multiple trait rubrics (Appendix 3A and 3B) that were used for assessment and discussed them with the participants. These pr eliminary procedures were coordinated with all four instructors participating in the study in order to ensure uniformity of their application. The decision to start the observations i n the beginning of the semester was made based on the following consider ations: (a) the prolonged engagement and observation would allow the collection of thick and rich qualitative data (Lincoln & Guba, 1985); (b) the researcher was able to enter the setting in a more natural way at the beginning of the semester before the st udents knew each other well (Patton, 2002); (c) the drawbacks of the overt (full disclosure) data collection process was minimized by the fact that the students became used to the presence of the researcher in the classroom (Patton, 2002); and (d) the proc ess of building and maintaining trust relationships between the researcher and the participants allowed collecting credible data (Lincoln & Guba, 1985). In addition, the data collection process during the pilot study showed that asynchronous CMC

PAGE 130

104 interviews required more time for question/answer exchange; thus, starting the interviews at the beginning of the semester lowered th e burden on the students and allowed for more effective interview process. Qualitative Stage, Step 3: This step started in the begi nning of the semester and continued to its end. Through the whole semester, open ended CMC interviews were conducted. The interviews were conducted via ICA 2, the courseware used in all six sections of the intermediate level writing class. The total of 18 questions was asked during the interviews. Questions 1 background as well as current and past experiences related to writing and CMC. These questions were asked during the weeks preceding the nested quantit ative stage. Questions 11 18 addressed opinions, feelings, and experiences related to the studied CMC interactions and the following writing. The actual interview question s are presented in Chapter IV, T able 4.3. The interview design considerations are dis cussed later in this chapter. Qualitative Stage, Step 4: This step took place after the data collection was completed. After the quantitative and qualitative data were collected and analyzed, the specific partic ipants in the instrumental case studies were identified. From the overall body of qualitative and quantitative data, the researcher extracted the data related to the selected instrumental case study participants. Initial ly, the researcher planned for S tep 3 and 4 to coincide during the post quantita tive stage of the study. The researcher intended working closely with the students selected to participate in the instrumental case study stage. However, the textual data analysis based on which these participants were selected required more time than anti cipated; thus, the instrumental case study participants were

PAGE 131

105 selected after the end of the second semester of data collection. This late identification of the instrumental case study participants posed certain limitations: the researcher was unable to cond uct focused observation s of these participants and missed the opportunity to elicit more detailed answers related to their participation in the pre writing activity and writing task completion. However, considering that the focus of the instrumental case writing interactions and drafts composed based on these interactions, as well as that the researcher was involved in prolonged participation and observation of the targeted learning community, the researcher believe s that the available qualitative data extracted from the instrumental case study participants was sufficient for the planned analysis. Quantitative Stage Procedures In this stage of the study, quantitative data were collected and statistically analyzed in order to answer the quantitative research questions. The CMC pre writing task (treatment) was performed, and data were collected during Weeks 5 and 6 during both semesters. To assure an educational setting that was controlled, the participants underwent t he pre treatment and the post treatment writing tasks on the same day and in similar computer laboratory environment. During the Spring semester of 2006 an additional laboratory was used at the IEP due to the high enrollment; however, this additional compu ter laboratory was very similar in terms of organization, settings, and computer availability to the main one. There were three experimental computer laboratory sessions: (a ) pre treatment di scussion and writing session, (b ) treat ment SCMC discussion sess ion, (c ) post treatment writing session. The ACMC discussion session was assigned as homework due

PAGE 132

106 to the nature of the asynchronous computer mediated communication that requires an extended period of time for message exchange. During the three experimental sessions and the homework ACMC session, the students worked in dyads that were randomly assigned within each Level III section. These dyads were also randomly assigned to either the ACMC or SCMC group. The class schedule of the Fall 2005 writing class was changed, so that the two sections could work together in the computer lab oratory during all three experimental computer laboratory sessions, thus assuring that they completed the pre writing, treatment, and writing assignments under the same conditions. I n Spring 2006, no schedule changes were made during the W eeks 5 and 6 three of the classes experienced the three experimental sessions at the same time but one of the sections completed the task in the additional computer laboratory to which it was assig ned from the beginning of the semester. The fourth section completed the experimental computer laboratory sessions one hour later following their regular schedule s On Thursday of the fifth week of each of the two semesters, the participants gathered in th e computer laboratory where they wrote their first drafts of a paragraph. The same prompt was assigned to all participants during both semesters of data collection (Appendix 3A). This writing assignment was completed during a 50 minute class session. The f irst 20 minutes of the class were devoted to completing a pre writing task that involved oral brainstorming and planning of the paragraph to be written based on the given prompt. This pre writing task was performed as a dyad activity. The students were rem inded that they could take notes during the discussion but only a few of them actually followed this suggestion. In the remaining 30 minutes, each participant wrote his/her first

PAGE 133

107 draft of the paragraph. At the end of the class, they saved their work on the server and/or a floppy disk and submitted a printed and an electronic copy to their instructor. After the pre treatment paragraphs were collected, the students were informed to which group ACMC or SCMC their dyad belongs. The ACMC participants receiv ed detailed instructions and materials (pictures) needed to perform the treatment decision making task as a homework activity (the treatment task is presented in Appendix 4). Considering the fact that some of the students did not have computers at home and needed to use the two university computer laboratories to which they had access, the period in which they were allowed to work on the ACMC task was from Thursday afternoon (Week 5) to Tuesday morning (Week 6); this allowed ample time to complete the async hronous interaction. This interaction aimed to support the participants during the pre writing stage and helped them in the process of generating ideas and planning the post treatment paragraph. The SCMC students received a homework reading assignment: the y had to read a text and answer questions based on this text. The text and the questions instructed to submit their answers by Tuesday morning of W eek 6 via ICA2 (the fu ll text of this additional assignment is presented in Appendix 5). On Monday of Week 6, a regular writing class session took place. During the Monday session the treatment task was not discussed. On Tuesday (Week 6) all the participants had a double sessio n (100 minutes) in the computer laboratory. During the first 50 minute session, the SCMC dyads participated in a synchronous discussion following the same pre writing task assigned as homework to the ACMC group. During the same session, the ACMC group perf ormed the reading task via ICA2 assigned as

PAGE 134

108 homework to the SCMC group (Appendix 5). As it was expected, the reading task took approximately 35 minutes. In the time remaining until the end of the class, the ACMC participants who completed the reading task were asked to review their ACMC interactions; however, they were instructed not to post any additional messages. The post treatment writing was performed during the second computer laboratory session on Tuesday, W eek 6 Each participant wrote his/her firs t draft of a descriptive paragraph following the post treatment prompt within a 35 minute period. During their writing, the participants had access to their CMC logs and were encouraged to use them. It was expected that the CMC log reference would take mor e time compared to recalling of or referring to notes taken during a face to face discussion; therefore the researcher decided to allow an additional five minutes for the post treatment writing session. At the end of the writing session, each student save d his/her paragraph draft on a floppy disk or the laboratory server and submitted a hard copy to the instructor. In order to assure that the experimental procedures were followed consistently, the researcher overlooked and assisted all but one Level III w riting sections during the three experimental sessions. As stated earlier, in the Spring of 2006, one of the Level III writing sections had its computer laboratory classes in a separate laboratory. The researcher trained an assistant who was familiar with CALL, the computer laboratories, and the IEP to observe and assist the computer laboratory experimental sessions and assure that all of the required procedures and time lines were followed. During pre and post treatment writing sessions, the students were offered technical assistance in order to ensure that technical problems did not influence the writing outcomes. No frequent or serious problems were observed at this point of the

PAGE 135

109 and they felt confident when working with L.E.C.S. and ICA2. The Treatment The specific pre writing task that the participants in the current study performed via CMC was a decision making task as classified by Pica et al. (1993). Prior to performing the t ask, each participant was given a set of six pictures that aimed to elicit discussion on the following topic: The People of the United States At this stage, the participants also were instructed to brainstorm together the main idea, supporting ideas, and paragraph organization of a descriptive paragraph that each one of them would write individually following the CMC discussion. The full task including the directions and the pictures are presented in Appendix 4. The topic was chosen considering several fac tors such as age appropriateness, interest, expected knowledge about US culture and history based on the curricula material covered in the writing class as well as in the other classes prior to the treatment, and the rhetorical features of the expected pos t treatment writing outcome a descriptive paragraph. It was decided to use pictures rather than text in order to assure elicitation of student produced language and active meaning negotiation during the discussion process. When given the pictures, the p articipants were explicitly instructed not to share them with other participants in a regular face to face interaction and to discuss them with their partner only via the assigned CMC mode using the respective application. The researcher ensured this behav ior during the synchronous discussion session. However, it was impossible to control the participants in the ACMC group in this respect. However,

PAGE 136

110 the later analysis of the ACMC interaction logs did not raise a suspicion that the paragraph pre writing discu ssion took place outside of the CMC environment. Instrumentation Quantitative Stage Instrumentation for Text Analysis The instruments used in the quantitative stage of the study aimed to measure eight paragraph writings that reveal specific language competencies as demonstrated in a written text: (a) syntactic complexity, measured through T unit analysis; (b) the amount of information that the writer attempts to handle in a single focus, measured throu gh mean length of idea units; (c) quantity of the overall information that the writer presents, measured through the number of idea units; (d) lexical information presented by the writer, measured through lexical density; (e) vocabulary complexity, measure d through the ratio of unique word of frequency groups used in the writing; (f) rhetorical soundness; (g) descriptive presentation; and (h) overall language use. The last three competenc ies were assessed with multiple trait scoring rubrics. The scores of t he pre treatment and post treatment writing samples for both the synchronous and asynchronous groups were calculated and used for statistical analysis in order to answer the quantitative research questions. Table 3.1 presents a matrix summarizing the instr uments used in the quantitative stage of the study.

PAGE 137

111 Table 3.1 Quantitative Stage Instrumentation Type of Analysis Definition of measured item Writing competency measured Writing Production Considered Tools for Analysis T unit mean length (Hunt, 1965 ) Main clause including all the subordinate clauses Syntactic Both syntactically correct T units and T units that are unambiguous, in other words, idea units that have minor grammatical errors. The re searcher and a trained rater T unit identification M ean length calculation software

PAGE 138

112 Table 3.1 (C ontinued) Quantitative Stage Instrumentation Type of Analysis Definition of measured item Writing competency measured Writing Production Considered Tools for Analysis Idea Unit mean length (Chafe, 19 85) A single clause containing one verb phrase, along with whatever other phrases (e.g. noun, prepositional phrases), adverbs etc. How the generated ideas single attention focus. Both syntactically correct idea units and idea units tha t are unambiguous, in other words, idea units that have minor grammatical errors. The researcher and a trained rater idea unit identification Mean length calculation software

PAGE 139

113 Table 3.1 (C ontinued) Quantitative Stage Instrumentation Type of A nalysis Definition of measured item Writing competency measured Writing Production Considered Tools for Analysis Lexical Density (Halliday, 1989) Lexical items as a ratio of a total number of clauses Lexical information presented by the writer Only lexica l items and clauses that are unambiguous (i.e. grammatically correct or with minor grammatical and/or spelling errors). The researcher identification of lexical and grammatical items in the text Lexical density calculation software

PAGE 140

114 Table 3.1 (C o ntinued) Quantitative Stage Instrumentation Type of Analysis Definition of measured item Writing competency measured Writing Production Considered Tools for Analysis Vocabulary Complexity (Nation, 2001) The index of unique words frequency distribution Vocabulary used Only vocabulary that is unambiguous (i.e. with minor grammatical and/or spelling errors). Frequency distribution software calculations based on published word lists 1 999 and 1000 2000 (Nation, 2001) and compiled by the researcher

PAGE 141

115 Table 3.1 (C ontinued) Quantitative Stage Instrumentation Type of Analysis Definition of measured item Writing competency measured Writing Production Considered Tools for Analysis Multiple Trait Rubrics CARLA (n.d.); Ferris and Hed gcock (1998). Assessing the quality of writing accounting for the given topic and the genre The following traits were considered: (a) rhetorical soundness; (b) presentation and development of the main point and its support; and (c) overall language use. The whole paragraph The researcher and trained rater

PAGE 142

116 T unit Analysis The syntax complexity, as measured by minimally terminable unit (T unit) writing on a syntacti c level. A T unit, as defined by Hunt (1965), is the main clause of the sentence including the subordinated clauses. In this study, the mean length (number of words) of the T units in each writing was measured. The researcher included both syntactically co rrect T units as well as T units that were incorrect but unambiguous; in other words, T thoughts. Park (1988) summarizes research findings confirming that mean T unit length increases with the language and writing maturity of the writer. However, using only the mean length of the main and subordinate clauses is not perceived to be an adequate predictor of writing quality (Crowhurst, 1987; Ferris & Hedgcock, 1998). Therefore, analysis of Idea Unit Analysis s were analyzed further in order to identify how the participants unit was used: that is, it contains one verb phrase, along with 106). Chafe (1985) argues that although idea unit analysis has been initially developed for studying spoken language, it could be applied to written language as well. He proposes that ide a unit analysis would allow researchers to evaluate text readability, and information quantity.

PAGE 143

117 Chafe, along with others (Carrell, 1987; Hildyard & Hidi, 1985; Wong, 2003) 1985, p. 106). Thus, idea units were first identified and employed for spoken language anal ysis and criteria that constitute idea unit in spoken language were defined. These criteria are: (a) clause final intonation contour; (b) pausing; accompanying noun phrases final 1980, p. 14). In his 1985 publi cation on linguistic differences of written and spoken language, Chafe (1985) continues developing the idea unit construct presenting it in written discourse. He states the following about the idea units in written language production: most readable writi ng shows idea units rather clearly. Readable writers seem to organize their material intuitively into this format, using punctuation marks to show idea unit boundaries, or to show the same intonational and hesitational patterns that in spoken language w oul d signal idea unit boundary. (p. 107) Further, Chafe presents the specifics of the idea units identified in writing: compared with the idea units produced in spoken discourse, written idea units tend to be longer, the mean length of spoken language idea u nit is 6 7 words (Chafe 1980, 1985), whereas the mean number of words per idea unit in written language is approximately 11 (Chafe 1985).

PAGE 144

118 When considering the construct of idea unit in the context of this study, it was important to establish a definition t hat would provide some concrete criteria for identifying i dea units in written texts, thereby allowing the researcher to approach the writings of the participants in a consistent manner. In his 1985 publication, Chafe does not provide an explicit definitio n of idea unit in written language production. However, if we re discourse, it would become clear that the two criteria (intonation pattern and pausing) stated for idea uni t of spoken language could not be applied. Thus, the third criterion is adopted by Chafe (1985): the tendency of an idea unit to consist of a single clause should be considered as the leading formal criterion. However, there are four exceptions stated in C expansion in written language (for the complete list of criteria, refer to Chafe, 1985, pp. 108 110); while 10 of the expansion devices keep the idea unit within the bound aries of one clause, the following four expansions suggest the use of more than one clause per idea unit: 10) Complement clauses: Certain interesting aspects of the situation indicate that we are not witnessing obligatory synchronic rules at all in the yo unger speakers Her realism involves the refusal to recognize that literary language has no referent 11) Restrictive relative clauses:

PAGE 145

119 The rules developed here have environmental constraints that are important for some speakers but not applicable for oth ers 13) Indirect questions: It is important to ask whether a given theory deals with the kinds of cognitive skills that children acquire normally how he or she looks 14) Indirect quotations: Lakoff say s that a sentence like 6b presupposes that Easter Williams is not a fish Some speakers say they heard someone say once that this referred to reeds in the lake here (Chafe, 1985, p. 110, emphasis in the original) Further, Chafe suggests that two or more idea units can be integrated into one sentence, by using (a) dependent clauses conjoined by a different coordinating conjunctions such as after, although, as, as if, if, in order to and so forth, should be considered as separate idea units. In addition, (b ) appositives (The teacher opened the book, a green volume always kept on the desk ); and (c) participial clauses ( Considering the consequences of failing to the programmer decided to test the software packet one more time.), sugge sting that these three constructs are to be considered as separate idea units.

PAGE 146

120 S the current study, the following working definition of idea units was used in the process of iden p. 106 ); however, if a complement or restrictive relative clause is present or indirect question or indirect quotation is used, these belong to the idea unit presented by the main clause. Dependent clauses, appositives (including examples provided by the writer, i.e. for are separate idea units. In addition, when analyzing CMC discourse production, phrases indicating agreement, disagreement, understanding and so forth were be considered as (1980, 1985) idea unit presentation is based on language data produced by native speakers or speakers whose written and oral language production abilities are close to native like. In other words, there are no morphosyntactically incorrect idea units considered in the discussion. The language data analyzed in the current stud y were produced by English language learners of intermediate level of proficiency; morphosyntactical and semantic issues were evident in these data. In this study, idea units were applied as a measurement of effective written language use that shows abilit y to present information in a way that is complex and reflects the specifics of the written discourse being at the same time reader sensitive. In addition, the goal of this measurement was to identify how much information the students attempt to handle in one

PAGE 147

121 attention focus. This was measured through the mean length of idea units, while the quantity of information the students attempt to handle in a paragraph was measured by the total number of idea units used in the paragraph writing. Because using idea u nits allowed the researcher to focus on the quantity of information that the learners attempted to communicate in their writings, the researcher decided to use morphosyntactically correct idea units as well as idea units that contained certain errors but w ere unambiguous accounting for unreadable portions of writing. Lexical Density Analysis Besides the formal characteristics of a clause, it is also important to consid er that the clause is a functional unit that carries the construction of meaning in a triple function. Halliday (1989) outlines these three functions of a clause as follows: (1) as the presentation of the phenomena of experience, as these are interpreted by the members of the culture; (2) as the expression of speech function, through the categories of mood th e form of theme plus exposition. (Halliday, 1989, p. 67). The former three character istics of a clause lead to application of lexical density as a measure that would further disclose the characteristics of the texts written by the participants. While the vocabulary complexity (the measure discussed next) allowed identification of the rich number of unique wor ds used, lexical density revealed the average amount of lexical

PAGE 148

122 information per clause. In order to perform the lexical density analysis, the formula proposed by Halliday (1989 p. 65) for calculating lexical information per clause was applied. This involved counting the lexical items as a ratio of the total number of clauses: Total number of lexical items = Lexical density (1) Total number of clauses Lexical items are the content words which function in a system that is indefinitely extendable. Halliday provides the following example to illustrate the participation of door is in contrast with gate and scr een ; also with window, wall, floor and ceiling ; with knob, handle, panel, and sill (Halliday, 1989, p. 63). Thus, the lexical item is a word or a combination of words that enters an indefinitely extendable set of items. Halliday contrasts lexical items to grammatical items which function in a closed system and include determiners, pronouns, most prepositions, conjunctions, some classes of adverbs, and finite verbs. The lexical ite ms may consist of one or two words in the usual sense. For example, in the phrasal he researcher specified the follo wing grammatical items lis t used for the current study: (a) determiners (b) pronouns, (c ) primary (verbal) auxiliaries, (d) modal auxiliaries, (e ) subordinators, (f) coordinators, (g ) negations not and no (h ) the infinitive marker to (i ) ordina l and cardinal numerals, and (j ) inserts (injections) as defined in Biber, Johansson, Leech, Conrad, and Finegan (1999). The analysis was performed in as follows: all the unique the researcher and confirmed by a co rater. Only words that belong to the 10 grammatical

PAGE 149

123 items listed in this section of the chapter were considered for further analysis. The next step was to find and analyze these words within the specific context in whic h they were used. This analysis was necessary based on the fact that homonymy is common in English language and some English words can be members of both grammar and lexical groups (Biber be a preposition, an adverb, or a subordinator. The list of the grammar items was updated with a description of each item outlining it possible usages in different contexts. As a guideline for this description and for the provision of a comprehensive list of possible usages of the items, the Merriam Webster On line Dictionary (2006) was used. Then each usage of each item of the list was analyzed within the context and it was determined if it belongs to the lexical or grammatical items group. This analysis was performed by the researcher and a trained co rater assured the accuracy with more than 98% agreement achieved. During the analysis, all items that were identified as grammatical were marked with an asterisk This marking aimed to signal to the softwar e program used to perform the calculations of lexical density formula that these items should be subtracted from the total number of words in each writing which would provide the number of lexical items in the writing. Taking into account that clauses are an essential element for meaning construction, when applying lexical density analysis, only lexical items and clauses that were unambiguous (i.e., grammatically correct or with minor grammatical and/or spelling errors) were considered in the analysis. Thi s lack of ambiguity was determined by two raters (one of them the researcher and the other a trained co rater) who scored the rater reliability (>90%) was achieved.

PAGE 150

124 Vocabulary Complexity Analysis The vocabulary c omplexity score aims to allow one to make the assertions of the the frequency characteristics of this vocabulary. The vocabulary complexity score was calculated bas ed on the number of unique words (families) used in each paragraph. These counts were divided into three groups: (a) words that are in the list of the 1,000 most frequently used groups; (b) words that belong to the second group of most frequent 1,000 2,000 words; and ( c) words that do not belong to Group 1 and 2 but are recognized as existing words. The sources of these lists for the first two groups (the first 2,000 words) are A General Service List of English Words designed by Michael West, and The Academ ic Word List holding 570 word families designed by Coxhead (Nation, 2001). The third list, Above 2,000 words list, was compiled by the researcher; it included the words used in the texts composed by the students. The criteria considered when compiling this list were for the words to be existing words in English language and not to appear in the first two frequency lists; however, several exceptions were made for the names of foreign holidays and foods. The analysis was performed using customized software de signed especially for the current study. The functions performed by this software are discussed in detail in the current chapter. Two independent raters verified the accuracy of the Above 2,000 Words List. The inter rater reliability was higher than 99%. Th e few instances of discrepancies were discussed and corrected; these instances were based on human error rather than on disagreement about the list structure. For each text three raw vocabulary complexity scores were identified unique words (families) based on which the final vocabulary complexity score was calculated:

PAGE 151

125 (a) Group 1: 1,000 word frequency group, based on the 1 999 word frequency list (Nation, 2001); (b) Group 2: 1,000 to 2,000 word freq uency group, based on the 1,000 to 2,000 word frequenc y list (Nation 2001); (c) Group 3: words that do not belong to the first 2,000 most frequent words but were recognized as existing words, based on the above 2,000 word frequency list, compiled by the researcher. The final vocabulary complexity score was pr esented as a ratio of Groups 2 and Group 3 raw score to Group 1 raw scores. The formula used for the calculation of the final vocabulary complexity score was: Raw Score Group 2 + Raw Score Group 3 Raw Score Group 1 Multiple trait Sc oring Rubrics The last instrument used to assess the pre treatment and post treatment writings was multiple trait scoring rubric, with a maximum score of six. The rubric design decisions were made based on the rubric creation guidelines presented on the Vi rtual Assessment Center, a web resource published and maintained by the Center for Advanced Research on Language Acquisition (CARLA, n.d.), as well as on the writing rubric analysis presented by Ferris and Hedgcock (1998). The full text of the rubric is av ailable in Appendix 3A and 3B. An even number of levels for the multiple trait rubrics was chosen based on the CARLA Virtual Assessment (CARLA, n.d.) guidelines. According to these guidelines, the even number of levels allows raters to make a more precise judgment of writing performances. As stated by Ferris and Hedgcock (1998), multiple trait rubric allow for the development of sound criteria for assessing the quality of writing in accordance with the given topic and genre. This type of rubric allows both the instructors and the students X 100 = Vocabulary Complexity Score (2)

PAGE 152

126 to focus on specific textual aspects. Further, the scoring rubric designed for this study aimed to elicit the articulation and development of a narrative/descriptive paragraph focusing directly on the purposes of the writi ng task. The following traits were considered when evaluating the paragraphs: (a) rhetorical soundness; (b) presentation and development of the main point and its support; and (c) overall language use. In order to ensure that the multiple trait scoring rub ric instrument was sensitive to the requirements of the writing task performed by the participants, the rubric designed for this study were discussed with the writing instructors participating in the pilot study, and their feedback was considered in the pr ocess of the instrument design. Rubric design considerations During the rubrics design period and the pilot test, the validity of the multiple trait rubric scores was ensured using the guidelines outlined by Ferris and Hedgcock (1998). To ensure construc t related validity, the rubrics were written closely following the paragraph rating scale and multiple trait scoring guide presented by Ferris and Hedgcock (1998). Content related validity was guaranteed through careful consideration of the information abo ut narrative paragraph structure requirements, which was communicated to the students by the instructors and stated in Boardman and Frydenberg (2002) the writing textbook used by the students. Further, it was ensured that the participants, the instructor s, and the researcher had a common understanding of the rubrics items, thereby maintaining integrity of the rubrics. Text Analysis Software A custom software application was designed by an experienced programmer who followed the function specifications pr ovided by the researcher. These specifications requested the calculation of the foll owing text analysis features: (a ) the total number of

PAGE 153

127 words, (b) the mean length of T units, (c ) the mean length and number of id ea units, (d ) the number of lexical items, and (e ) word frequency distribution. The application was written in the PHP 5 open source programming language (PHP, 2006) using as a backend the open source MySQL database server (MySQL, 2006). The reason for employing a computer application rather than m anual counting and calculation is that, although all of the text analysis operations could be performed by hand, when working with a large data sample the possibility for human error increases However, selected portions of the data were analyzed manually and the outcomes of this analysis were compared with the computerized analysis in order to assure the reliability of the computer algorithms. Internet Based Instruments for Data Collection The application that was used for both the asynchronous CMC pre w riting tasks and the interview process is Internet Based Classroom Assistant 2 (ICA2). This is an online courseware developed by Nicenet (1998), which is a volunteer, non profit organization committed to providing free services to the educational Internet community. The access to the application is free of charge, and no advertising is displayed on any of the screens. In addition, ICA2 is attractive for educational use because no supplementary software is necessary to run the application. The only requireme nt is Internet connection; once accessed online, the application is opened through the web browser installed on the using it should not impose an additional burden on the students or teachers (Zoran & Sarieva, 2003). The ICA2 user friendly interface allowed to keep the training time of both teachers and students to a minimum. This courseware has been used by most of the instructors in the IEP for several years prior to the data collection period, and during

PAGE 154

128 professional meetings, the instructors reported it to be a useful class management tool (ELI CALL, 2003). Among the multiple features offered on ICA2, there is an asynchronous communication conference forum avai lable. In this forum, the instructor can post discussion questions and assign students to multiple discussion groups. Another important only users who have the unique class key generated by ICA2 a nd given only to the instructor can join a particular class. The application that was used for synchronous communication is Language Educational Chat S ystem (L.E.C.S.). This is a web based chat application created by Kanto Gakuin School Corporation, Japan (n.d.). The creators and researchers who have used the application report that this is a flexible and easy to use tool designed especially for l anguage teachers and learners, ( Kanto Gakuin School Corporati on, Japan, n.d.) ; Freiermuth, 2002). As with ICA2, the only requirement is Internet connection. Once data generated from thes e logs, such as counts of overall words used, target words, and consuming or complicated; the student simply needs the chat. Although simple, these procedures protect users from possible intruders. A possible drawback of the application might be the fact that along with the English explanations, explan ations in Japanese are appearing on the log on screen. However, after consulting with instructors who have used the application as well as with the students

PAGE 155

129 who participated in the pilot study, it was concluded that this is not a distracter that could impe de the SCMC process, in particular, and the language learning process, in general. Treatment Design Considerations Prompt Design Considerations The prompts designed for the pre and post treatment paragraph writing assignments were the following: (a) Pre Trea tment Writing Prompt: The University International Student Organization has invited you to write a note for their newsletter. The topic of your note is: My favorite U.S. holiday. Your note needs to be in the format of a descriptive paragraph. Write the f irst draft of this descriptive paragraph. You have 30 minutes. Title: My Favorite U.S. Holiday (b) Post treatment Writing Prompt: The University International Student Organization has invited you to write a note for their newsletter. (c) The topic of your note i s: The People of the United States Your note needs to be in the format of a descriptive paragraph. Based on the discussion you had with your partner using L.E.C.S. or ICA2, write a descriptive paragraph about the people of the United States. Make sure t hat it corresponds with the pictures you received for the discussion. Include in your description all the ideas presented by all the pictures. You can use the records of your L.E.C.S. or ICA2 discussion when composing the paragraph. You will have 35 minute s to write your first draft of this paragraph.

PAGE 156

130 The writing tasks were presented based on a situation frame stated in the prompts; this allowed the pre and post format (Kroll & Reid, 1994). Several var iables were considered in the process of prompt the prompts were designed in accordance with the Intermediate Academic Writing course goals which specified that students would master par agraph writing (including descriptive paragraph) during the first half of the semester; (b) content the prompts aimed to generate writing on a content familiar to the students through their personal experiences and previous discussions of U.S. culture in their Academic Writing and Cultural Contacts classes; (c) prompt linguistic features the language of the prompt was adapted to the proficiency levels of the students, and prompt clarity was confirmed by the writing instructors and the participants in th e pilot test; (d) task specificity the task, the writing time frame, as well as the required rhetorical features were explicitly presented; (e) rhetorical features the expected writing outcome (a descriptive paragraph) and the audience/purpose of the paragraph were explicitly stated in the prompt. When designing the prompts, the level, and the content of the instructional material used in the writing class. Treatmen t Task Design Considerations Before discussing the details of the treatment task design, it is important to state the definition of task adopted in the current research. After carefully studying a collection of language learning task definitions developed and adopted in language acquisition research and presented by Johnson (2003), the researcher was guided by two definitions

PAGE 157

131 work which involves learners in comprehending manipulating, producing or interacting in the target language while their attention is principally focused on meaning rather than form. The task should also have a sense of completeness, being able to stand alone as a (p. 10). The second definition that of Swales (1990) comprehension and communication in the target language involving meaning construction. Swales further defines task goal directed activities drawing upon a range of cognitive and communicative procedures relatable to the acquisition of pre genre and genre skills appropriate to a foreseen or emerging sociorhetorical situat communicative activity that focused on meaning and involving language comprehension and manipulation; this activity can stand alone, but is also designed to support the development of specific genre w riting skills through a range of communicative procedures in a specific sociorhetorical situation. Further, the specific features of the CMC treatment task were planned following the communication task design guidelines and communication task typology vali dation presented by Pica, Kanagy, and Falodun (1993). Pica et al.(1993) outlined specific task features as indicators for interaction opportunities in the target language. These features are: (a) comprehension during the communication process; (b) receivi ng feedback; and (c) generating modified production. Pica et al. (1993) concluded that, to insure productive interaction, the participants should work towards a common goal. In the treatment task, the following relationships, requirements, goals, and outc omes were considered. First, the task was designed as an interactional activity

PAGE 158

132 between two ESL learners of the same proficiency level (intermediate) where each dyad member held the same information in a picture format. This information needed to be verbal ized and interpreted by the participants during the interaction process. Second, this design set the interaction requirements, i.e. each dyad participant had to supply and request information in the process of meaning negotiation and task outcome construct ion. Third, during the interaction, both participants should be striving to achieve the same goal, which was to brainstorm paragraph content and organization. Fourth, for the CMC task itself, more than one possible outcome was acceptable, namely the interp retation of the specific pieces of information that were presented in the pictures given to the expected (these expectations were confirmed by the pilot study data analysis). However, the organization and interpretation of these single outcomes in the paragraph can be different. The CMC task is presented in Appendix 4. Taking into account the fact t hat communicative tasks discussed by Nunan (1989), Swales (1993), and Pica et al. (1993) were performed in a face to face environment, validation of communicative tasks performed via CMC was considered (Blake, 2000). Based on the findings of the study of i ntermediate level Spanish language learners who participated in synchronous CMC while solving different communicative tasks, Blake concluded that in a synchronous CMC environment, the same factors outlined by Pica et al. (1993) contribute to successful int eractions. No studies that compare asynchronous and synchronous student student interactions generated during completion of a communicative task as categorized by Pica et al. (1993) were found by

PAGE 159

133 the researcher. However, Sotillo (2000) in her study of adva nced ESL writing classes, compared reading discussions performed by two groups of students who participated in synchronous and asynchronous CMC. She found that in both groups the discussion shared similar discourse features; however, the SCMC discussion da ta demonstrated greater variety of discourse features and were more interactive, while the ACMC postings were more focused on the discussion topic and were lengthier and syntactically more complex. Therefore, based on the fact that there were similarities in the two CMC modes of communication that allow effective communication, and that the communicative tasks which were designed following the principles that Pica et al. (1993) were concluded to be successful in a n SCMC language learning environment, the re searcher decided to apply the same task for both ACMC and SCMC groups. Further, this decision was driven by enhanced interaction (Salaberry, 2001). He points out that the cost of dela y in communicational media without contemporality constraints such as e mail is minimal. Data Analysis Procedures and Reliability and Validity Considerations Text Normalization and Coding After the data were collected, the CMC and the pre and post treatm ent paragraph texts were normalized to assure reliable word counts. The nature of this normalization was the following: (a) the paragraphs were carefully read by the researcher and all the ambiguous phrases and sentences were identified by putting them in square brackets in order to flag them for exclusion during the computerized text analysis, (b) the CMC interactions and the paragraphs were read by the researcher and all misspelled words were replaced with their correctly spelled counterparts while the in their original spelling

PAGE 160

134 were bracketed. Following is the explanation of the reasons for these normalization procedures and the way they were implemented. The necessity to identify ambiguous text segments was based on the fact that the researcher made th e decision to use both incorrect and correct text units (T units, idea units, and clauses). This decision was determined by the specific nature of second text segment wit hout making grammar, vocabulary, and/or spelling deviations from standard English, they are able to construct texts that are readable and convey both facts and ideas. Thus, the researcher realized that excluding all the incorrect but unambiguous clauses an process and development in the context of the current study. The working definitio n of unambiguity/ambiguity constructed and adopted by the researcher was the following: unambiguous text segments (clauses and sentences) may have some grammatical inconsistencies; however they are easy to understand and do not impede the readability of t he text. If a text segment requires using extensive assumptions or sentence) should be considered as an ambiguous and excluded from the text analysis. In addition, words which were repeated or did not fit semantically into a phrase were considered as ambiguous text and thus excluded from the analysis. The ambiguity of the text segments was established following strict procedures: (a ) the researcher read carefully the paragraphs, identified and marked in the text the amb iguous phrases and sentences; (b ) to assure intra rater reliability ( Moskal & Leydens,

PAGE 161

135 2000), the researcher reviewed the paragraphs one more time one day later; (c ) All the excerpts identified as ambig uous were reviewed by two independent raters who were trained by the researcher. These raters used the working definition of unambiguity/ambiguity when identifying the ambiguous text segments in the data. The percentage of agreements about the ambiguous te xt segments was 89.29% with Rater 1 and 85.71% with Rater 2 The researcher was not satisfied by this percentage of agreement and reviewed carefully the portions of the data that was interpreted differently by the raters and herself, after the researcher a nd the raters discussied the problematic segments of the data, the agreement percentage achieved was above 95%, which is considered to be a high percentage of agreements ( Miles & Huberman, 1994); and (d ) each of the two independent raters reviewed the full text of 14 randomly chosen paragraphs which presents 10% of the paragraphs; thus the total of 20% of the paragraphs were reviewed. The goal of the raters was to identify excerpts of ambiguous text unrecognized by the researcher. High inter rater reliabili ty above 98 % was achieved with both raters ( Miles & Huberman, 1994) The second text normalization procedure was spelling normalization. Normalizing the spelling was necessary in order to be able to use computer software for the lexical density and vocabul ary complexity analysis. In this analysis a comparison of the vocabulary items used by the students with electronically stored grammatical items list and frequency word lists (Nation, 2001) was performed. The researcher identified all the misspelled words in the data, marked them by placing square brackets around these words and provided the correct spelling for them immediately after the brackets. For [nest] next

PAGE 162

136 In this procedure misspelling wa s defined a s an instance when (a ) a word is spelled so it appears to be wrong and this spelling does not correspond with any English word, for example [whant] want ; or (b ) a word is spelled as an English word that belongs to another word family: for example [nest] ne xt ; instances when the same family word was used in a inappropriate form, for example make made, student students etc., were not cons idered to be misspellings; or (c ) a word that contains two stems is spelled as two words: for example [some where] somewher e It is important to note that when identifying the misspelled words and making the decision about the correct spelling, the the reliability of these spelling int erpretations, the researcher used a second rater to check 100% of the writings and verify that all of the spelling mistakes as identified in the operational definition explained With this rater, 100% agreement was achieved. Next, two independent raters re viewed all of the identified spelling mistakes and their corrections accounting for the context in which they appeared. High inter rater reliability was achieved: 97.88% with Rater 1 and 98.94% with R ater 2 above described way, the number of words in each paragraph was calculated. These numbers were stored in a database that was built to assist the calculation process of the mean length of T units, idea units, lexical complexity, and lexical density of stude paragraphs were coded identifying the T units, the idea units, and the clauses for each paragraph. The researcher perf ormed the coding twice on two different days to assure higher intra rater reliability of the coding. Next, a trained rater reviewed 40% of the

PAGE 163

137 coded data and confirmed or disconfirmed the coding of the T units, idea units, and clauses. For all three crite ria high inter rater reliability was achieved: for T units the percentage of agreement was 98.17%, for idea units it was 97.95%, and for clauses 98.18%. Multiple trait Rubric: Reliability and Validity Considerations The instruments had been examined bas ed on the results of the pilot test conducted in the Fall semester of 2004 in the targeted IEP. The writing instructors had been trained to use the rubric. They and the researcher had assigned a score to each student paragraph produced during the pilot tes t using the multiple trait rubrics, and the reliability of each score had been assessed by comparing the scores assigned to the pilot test writing tasks performed by the participants. For the pilot pre treatment and post treatment paragraphs, the inter rat er reliability for the rubrics scores had been calculated (1994, p. 64) formula This formula presents the inter rater reliability as the following: Number of agreements = Inter rater Reliability (3 ) Number of possible agreements According to Miles and Huberman (1994) an inter rater reliability should be in the 90% range; therefore, the researcher was satisfied with the results and concluded that the designed rubric had generated consistent scores f rom different raters when applied by trained evaluators. Following the procedures applied in the pilot test, in the current study both the pre and post treatment paragraphs were scored by two independent raters, one of them the researcher and the other t he ESL writing instructors who participated in the study. They were trained in applying the research instrument discussed above. To ensure the

PAGE 164

138 reliability of scores, strict application of the multiple trait rubrics was required. High inter rater reliabilit y the scores (in the range of 90%) was sought. In order to guarantee high reliability of scores, if a score differed by more than 10%, raters discussed the specific writing sample and came to a consensus. Further, to obtain comparable scores for the statis tical analysis, all scores were converted to z scores (Glass & Hopkins, 1995). Qualitative Stage Instrumentation The Researcher In a qualitative study, the researcher is the main data collection and interpretation instrument (Lincoln & Guba, 1985; Patton 2002). The nature of qualitative inquiry ). In other words, in this specific study the rather to investigate certain issues, in this case, issues related to incorporating CMC into the language learning process, a nd, more specifically, in the process of establishing writing skills in a second language. However, the neutrality of the researcher is not equivalent to detachment; it rather involves empathy showing understanding of, interest in, and caring about the p articipants (Patton, 2002). During the qualitative stage of the current study, the researcher aimed to collect data that would disclose both the observable behaviors of the qualitative participants and their internal views, feelings, experiences, and opin ions, as well as questions that would reveal the intertextual relation of the CMC discussion text created by the participants in the process of pre writing collaboration and the texts that each participant produce

PAGE 165

139 afterwards. The qualitative stage was an o ngoing stage that involved data collection before, during, and well after the quantitative stage data collection and analysis was completed. The qualitative data collection process started before the beginning of the semester during the informal meetings w ith the non student participants in the study: the curriculum designer, the CALL consultant, and the writing class instructor. Qualitative task logs, writing samples, CMC inter views, and informal interviews with the instructors. Following is a description of additional instruments intended to be used in the study. IEP Documentation The researcher collected data through various documents related to the IEP where the study was c onducted and more specifically to the writing classes of the targeted level. These include d : (a) the Goals and Objectives document for Academic Writing III class (Appendix 1A); (b) Placement Test info rmation (Appendix 1B and 1 C); and (c) Integrated CALL C urriculum Objectives and Activities document (Appendix 2). These documents were provided by the non student participants and contributed to the collection of thick and rich qualitative data. The data collected using these instruments supported building the background of the case studies and the understanding of the observed processes. Observation Field Notes During S tep 2 of the Qualitative stage, the researcher focused the class observations on all stude nts. The following trends that e merge d from the obse rvation data were noted: (a) group interactions during the face to face and computer laboratory classes; (b) student/teacher interactions during the face to face and computer laboratory

PAGE 166

140 classes, (c) attitudes and opinions that students express and demonstr ate towards writing; (d) attitudes and opinions that students express and demonstrate towards technology, and more specifically towards computers, and their implementation in the language learning process. During each class session, raw field notes were t aken. The raw field notes were used to create extended field notes in a narrative format. This aimed to organize and reduce the amount of observation data by creating condensed notes of interactions, events, behaviors, attitudes and opinions. The informal debriefing sessions with each of the four teachers participating in the study informed the researcher in the process of field observations. In order to assure the credibility and reliability of the data, in addition to the informal debriefing session held on a regular basis, the researcher discussed these field notes with the teachers in order to assure that the dynamics of the observed classes were depicted and interpreted correctly. While it is difficult to report the agreement percentage as an outcome o f these discussions, based on these outcomes, the researcher concluded that the qualitative data depicted in the field notes reflected the targeted processes. In addition, the final narrative of the analysis of the computer laboratory sessions presented in Chapter IV was shared and discussed with the teacher who was a participant in the study during both semester in which the data were collected. She confirmed the trends depicted, described and interpreted by the researcher. Writing Samples and CMC Logs Th e paragraphs written by the participants following writing task described earlier in this chapter and the pre treatment CMC discussion logs produced by the students during their pre writing discussion represented another set of qualitative instruments that

PAGE 167

141 provide d insights about the processes under investigation. The ACMC logs constituted of via ICA2; while the SCMC logs presented the synchronous discussion exchange of the participants of the SCMC group. These exchanges were performed in order to complete the pre writing treatment CMC task described earlier in this chapter. The students wrote their first drafts of the descriptive paragraph of the assigned topic. The CMC log s also provided insights about how each student approached the computer mediated communication act. In addition, the aim was to collect data that would allow the investigation of the intertextual relations that exist ed between the paragraphs created by stu dents and text generated during the CMC discussion. The CMC Interviews The computer mediated interviews were conducted via ICA2. The Conferencing section of the courseware was used for the interview. The overall goal of the interview was to reveal the pas t and current experiences of the participants and how they perceive these experiences in the context of the pre writing CMC task and their ESL writing development. More specifically through the interviews the researcher aimed to find out what were the par the role of computers in education and for acquiring writing skills. In addition, the which they participated, in particular, and how this supported their writing performance level and writing skills development. The interview questions are presented in Table 4.3, the table denotes type of questions and provides the question text. These interviews se rved as a tool for the

PAGE 168

142 collection of self reported data from the participants. Further, these data were treatment task, first draft paragraph writings, and field notes taken during the overall data collection process. In t he beginning of the interview (Q uestions 1 10), the researcher collected demographic data about participants including gender, age, country of origin, native language, and educational background such as educa tional level, country(s), and educational settings where education was received. Issues related to education, writing, and CMC were elicited from the participants. Further, Q uestions 11 18 were related to the treatment and its effect on writing. The resear cher aimed to elicit information students participated in the interviews. A combination of interview approaches was planned to be applied: standardized open ended interv iew and interview guide (Patton, 2002). The researcher decided to follow these approaches because the standardized interview format allows each interviewee to be asked the same set of initial questions, hereby permitting the identification of common themes and issues. On the other hand, the follow up questions qualitative research sub questions used as an interview guide, were expected allow the researcher to explore in more depth issues related to the basic lines of inquiry in the types: (a) background questions, (b) experience and behavior questions, (c) opinion and value questions, and (d) feeling questions (Patton, 2002). The goal of these questions was

PAGE 169

143 to elicit data related to four trends: (a) attitudes towards writing, (b ) computer and CMC related experiences (c ) CMC interact ion and task perceptions, and (d ) CMC interactions: t ools and typing. Table 3.2 presents the question types and trends. Table 3.2 Interview Questions Types and Trends Question Type Question Trend 1 4 Background (a) age; (b) country of origin; (c) native langu age; (d) educational background 5 Opinion att itudes towards writing 6 Current Experience attitudes towards writing 7 Current Experience attitudes towards writing 8 Current Experience computer and CMC related experiences 9 Past Experience computer and CMC related experiences 10 Past Experience co mputer and CMC related experiences 11 Opinion CMC interaction and task perceptions 12 Feelings CMC interaction and task perceptions 13 Current Experience CMC interaction and task perceptions 14 Opinion CMC interaction and task perceptions 15 Current E xperience CMC interaction and task perceptions 16 Current Experience CMC interaction and task perceptions 17 Current Experience CMC interactions: tools and typing 18 Current Experience CMC interactions: tools and typing

PAGE 170

144 To assure the credibility and t he transferability of the interview data, three steps were taken. (a committee members who are experts in the field of educational research, ES L education, and ESL writing, (b) t he questions also were discussed with the ESL instructors teaching the pilot study participants in the Academic Writing classes to assure that the language of evel of language proficiency, and (c) d uring the pilot study, in terviews with sample questions were conducted with all participants. The and to determine if the information elicited during the interview would allow the qualitati ve research questions to be answered. It is important to attend in more detail to the mode of the interviews because the CMC mo de chosen for this study differed from the traditional face to face way of conducting interviews. The CMC mode of interviewing was chosen based on the following reasons outlined in the research on CMC and second language learning (Beauvois, 1997; Kern, 1995; Sengupta, 2001; Warschauer, 1999): (a) CMC provides a discussion space in a less threatening environment compared with face to face interactions; (b) the students have more time to construct their responses and tend to produce longer sentences expressing their ideas in more depth; and (c) learners express themselves at their own pace and can review, revise, and clarify their po stings before submitting them. Researchers from fields not related to ESL report computer supported interviews to be less threatening: during this type of interview, people tend to provide more responses than in situations in which they feel insecure or i ntimidated (Nathan, 2001).

PAGE 171

145 Using asynchronous CMC interview mode was expected to eliminate some of the disadvantages of face to face interviews, such as unintentional interviewer influence due to gestures, mannerisms, or face expressions that indicate subt le signs of agreement or disagreement (Tashakkori & Teddlie, 1998). This mode of interviewing was expected also to allow for easier member checking following data analysis through using ICA2 for posting questions to the participants without requiring a spe cial meeting. However, it is important to note that these expectations were not met in the current study: some of the open ended questions and the majority of the follow up interview guide questions based m the participants which was attributed to the asynchronous nature of the interview many of the students ignored these questions. This aspect of the interviews is further discussed in the Limitations of the Qualitative Stage section of Chapter IV. Trustwo rthiness, Credibility, and Transferability Considerations In order to establish the trustworthiness of the qualitative stage of the study, specific techniques suggested by Lincoln and Guba (1985) were considered and applied. The credibility of t he data col lected in the study wa s assured through the prolonged engagement of the researcher. Together with the non student participants in the study, the researcher took part in the writing curricula discussion and adjustment in order to confirm that the study woul d not have a negative impact on the naturalistic nature of the learning environment. In other words, the tasks applied in the study fitted within the philosophy of the IEP where the data were collected, and they were consistent with the goals and procedure s of the writing course. Further, the researcher closely observed the intermediate level Academic Writing classes during the whole semester. During the

PAGE 172

146 whole process of data collection, the researcher was taking field notes that account for the face to fac e and computer laboratory sessions and developed research protocols that reflected the processes observed. This provided the opportunity to further collect thick description assuring the transferability of the study as well as information that would reveal the interpretive nature of the research process. During data collection and analysis, the researcher collaborated with a peer and the members of the dissertation committee to ensure that the data were collected and interpreted truthfully and without error s and the conclusions were justifiable. The researcher sought high inter rater reliability during this process (Miles & Huberman, 1994). Quantitative Stage: Data Analysis The statistical procedures applied at this stage allowed the researcher to answ er the quantitative research questions of the study as outlined in this chapter. The researcher first compared the ACMC and the SCMC groups in order to identify the degree to which that the groups differed in terms of: (a) age, (b) educational background, (c) gender, (i.e. ) and (d) writing abilities. Further, the post treatment writing scores of the two groups were compared for each of the eight proposed textual features in order to find out if th e post treatment writings were significantly different for the two groups. Finally an additional quantitative analysis was performed comparing the intertextual connections on a lexical level that the participants in the two groups established between thei r treatment discussions and post treatment writing. The data analysis tests and procedures that were employed are presented further.

PAGE 173

147 Comparing the ACMC and the SCMC Groups Group Com parison: Age The ACMC and the SCMC groups were compared in terms of the age of the students assigned to them. The goal of this comparison was to find out if there were significant differences in terms of age composition between the two groups. The ages of the students were self reported in the interviews conducted during the qualitative on going stage. In order to perform this comparison, a two tailed t test was conducted. The Null Hypothesis for the t test was that the means of the two groups did not diffe r significantly. Group Comparison: Educational Background The ACMC and the SCMC groups were compared in terms of the educational background of the students assigned to them. The goal of this comparison was to find out if there were significant differences in terms of educational background composition between the two groups. Three levels of educational backgrounds were identified bas ed on the interview outcomes: (a) high school; (b ) at least one semes ter of college education, and (c ) college degree either (or its equivalent) and/or employed prior to their enrollment at the IEP. In order to perform the statistical comparison of the groups, the researcher collapsed educational level one (high schoo l) and two (college) into one Pre College/College. This decision was made due to the small sample size. The decision to collapse these two levels was based on the fact that the majority of the students who already received their college degree were profe ssionals with starting or already established careers; thus, it was viewed that the college level group would be more

PAGE 174

148 similar to the high school level group in terms of their educational background. Consequently, two levels were considered for further comp arison analysis of the ACMC and the SCMC groups: (a) Pre College/College and (b ) Post College level. The groups were compared in terms of the odds of a participant of a particular educational level to be assigned to a particular group (ACMC or SCMC). The o dds ratio analysis allowed finding out if there was a statistically significant difference between the two groups in in order to find out if there was an association of student education level and CMC group assignment. Group Comparison: Gender This comparison was performed in order to find out if the random assignment of the participants to either an ACMC or SCMC treatment group resulted in groups of similar gender compo sition. The groups were compared in terms of the odds of a male or a female participant to be assigned to a particular group (ACMC or SCMC). The odds ratio analysis allowed finding out if there was a statistically significant difference between the two gro ups in terms of gender composition. Further, the groups were compared in terms of dyad gender composition: mixed or the same gender dyads. The odds ratio analysis allowed finding out if there was a statistically significant difference between the two group in order to find out if there was an association of dyad gender composition and CMC group assignment.

PAGE 175

149 Group Comparison: Pre treatment Writing In order to compare the outcomes of the pre tr eatment writings, the eight proposed textual measures were compared, namely: (a ) the syntactic complexity (measured by calculating the mean length of t units), (b ) the amount of information present in a single focus (measured b y mean length of idea units), (c ) the quantity of overall information present (measured by the number of idea units) (d ) lexical information per clause (measured throu gh lexical density analysis), (e ) vocabulary complexity (measured by analyzing the frequen cy of the unique words used ), (f) rhetorical soundness, (g ) presentation and d evelopment of main ideas, and (h ) overall language use. The first five measures yielded continuous scores, whereas M easures 6 through 8 were presented with ordinal scores. Thus, two different kinds of st atistical tests were used to compare the two groups based on the type of scores (continuous v s. ordinal). The pre treatment Writing S cores 1 5 were analyzed by applying five exact t tests in order to make inferences about the difference between the means o f the two groups (ACMC vs. SCMC) and evaluate how similar the two groups were in terms of writing skills. The Null Hypotheses for each test were that the group did not differ si gnificantly. The Pre treatment Writing S cores 6 7 were using three non parametr ic Mann Whitney U tests. Again the Null Hypotheses for each test were that the group did not differ significantly. Analysis of the Post treatment Paragraphs The post treatment paragraphs were analyzed in order to find out if there was a difference betwee treatment writings based in respect to the eight proposed textual measures presented earlier. The post treatment data

PAGE 176

150 was analyzed using analysis of covariance (ANCOVA) method of statistical analysis for the five contin uous measures used in the study while the three ordinal measures were analyzed applying three separate Mann Whitney U tests. When performing the five ANCOVA tests, using according statistically significant effects was high, even though if each of the analyses were performed separately or fewer ANCOVA tests were conducted, at least one of the effects could be detected (Onwuegbuz ie & Levin, 2005). Thus, in order to avoid this limitation, the Three Step Approach for Testing Multiple Univariate Hypotheses proposed by Onwuegbuzie and Levin (2005) was applied. This led to the analysis of the pulled z scores from all variables in order to compare the ACMC and SCMC groups. Analysis of Covariance When using an ANCOVA an analysis of variance (ANOVA) is combined with regression analysis by adding a concomitant variable (covariate). The goal of utilizing analysis of covariance is to attemp t to control statistically for differences that existed before the treatment was implemented for the ACMC and SCMC groups before the treatment ( Onwuegbuzie & Daniel, 2003). In this study, each ANCOVA test aimed to answer one quantitative research question, and the concomitant variable used in all tests treatment mean scores. The decision to apply ANCOVA for the analysis of the five continuous scores was driven by the fact that this type of statistical analysis allows for reductio n of the bias and clarification of the relationship between the dependant and independent variables through allocating a percentage of the variance in the dependent variable to a potentially confounding variable ( Loftin & Madison, 1991) In other words, in this case it allowed

PAGE 177

151 for the adjustment of differences in the writing skills of the participants based on the assessment of pre treatment writings. The application of this inferential statistical procedure was viewed as appropriate for the current study based on the fact that the ACMC and SCMC groups were assigned randomly (Henson, 1998). The following null hypotheses were tested: H 0 Aj Sj = 0, where A presents the ACMC group, S presents the SCMC group, and j presents the writing score used in the tes t. The dependent variable ( Y ) for each ANCOVA test were the specific continuous treatment writings, while the covariate ( X ) was the corresponding writing score obtained after grading the particip pre treatment writings. Thus, the five ANCOVA tests that were applied in order to answer quantitative Research Questions 1 through 5 were the following: (a ) Question 1: T unit ANCOVA with Y = post treatment T unit score, and X = pre treatment T unit score; (b ) Question 2: idea unit mean length ANCOVA with Y = post treatment idea unit mean length score, and X = pre treatment idea unit mean length score; (c ) Question 3: idea unit number ANCOVA with Y = post treatment idea unit number score, and X = pre trea tment idea unit number score; (d ) Question 4: lexical density ANCOVA with Y = post treatment lexical density score, and X = pre tre atment lexical density score; and (e ) Question 5: vocabulary complexity score ANCOVA with Y = post treatment vocabulary complexity score, and X = pre treatment vocabulary complexity score. The use of the pre treatment scores as covariates was expected to improve the accuracy of Y prediction in each of the ANCOVA tests (Glass & Hopkins, 1995). In ANCOVA, the average varianc e within groups, MS w was reduced to approximately MS W (1 r w 2 ), where r w presents the within group correlation between the Y and the X for

PAGE 178

152 each ANCOVA test applied. The adjusted and unadjusted means of each group in each test was studied and reported in order to account for the credibility of ANCOVA findings: a considerable difference between the adjusted and unadjusted means affects the credibility of the ANCOVA (Glass & Hopkins, 1995). Mann Whitney U Test The ordinal nature of the scores obtained with the multiple trait rubric called for a different statistical analysis (Glass & Hopkins, 1995). These three ordinal scores, were analyzed applying three separate non parametric two tailed Mann Whitney U tests. The Null Hypotheses for each of the tests were that there was no difference between the means of the two groups. Pulled Score Analysis The researcher took into account the fact that in this study multiple tests of variance were performed in order to answer the research questions. Thus, in order to avo id the high probability of inflating the Type I error (rejecting the H o when it is true) for each statistical test a Dunn Bonferroni ew ew value was ew tests in the study and .05 is the pre specified Type I erro multiple univariate tests, posed serious limitations to the study. When performing the five ANCOVA tests, the possibility of not findin g statistically significant effects was high, even though if each of the analyses was performed separately or a fewer ANCOVA tests were conducted, at least one of the effects could be detected (Onwuegbuzie & Levin, 2005). In order to avoid this limitation, the Three Step Approach for Testing Multiple

PAGE 179

153 Univariate Hypotheses proposed by Onwuegbuzie and Levin (2005) was applied. In the first step, each univariate hypothesis was tested for statistical significance. Then, the third step was applied for the statis tically non significant outcomes. Onwuegbuzie and Levin (2005) suggest three different scenarios for Step 3. In this study, the univariate test of averaged standardized outcome variables described by them was applied. First, each of the eight outcome meas ure score was standardized ( z scores were calculated) making sure that the z scores on each measure were similarly oriented (i.e., better performance was presented by lower z scores). Then, the z scores for each variable was summed and averaged across meas ures, thus yielding a composite z score for each case. Further, an independent samples t test of group mean difference was performed in order to compare the ACMC and SCMC groups in terms of the average composite measure. The Null Hypothesis tested was tha t there is no significant difference between the z score means of the synchronous and the asynchronous groups. Qualitative Data Analysis Procedures The qualitative stage of the study followed a multiple case exploratory study design (Yin, 2003). Two sub s tages were implemented: a group case study stage with 60 participants and an instrumental multiple case study sub stage with eight participants identified based on their writing gains. The researcher described the educational setting based on the data col lected from the non student participants through informal interviews in the Qualitative stage, Step 1. Further, based on her observations and detailed narratives created immediately after each observation in the Qualitative stage, Step 2, the researcher de laboratory session. The first two steps allowed for providing rich background of the

PAGE 180

154 educational settings in which the participants acquired their writing skills. Furthe r, in Step 3 the researcher conducted CMC interviews with the 60 participants in the group case study and in Step 4, the researcher built eight individual instrumental case studies. Sub stage I: The Group Case Study In the group case study, multiple sourc es of data were used. Data were obtained through (1) IEP documentation, (2) field observations, (3) CMC interviews with the treatments writings. In addition, the debriefing sessions with th e non student participants in the study further class interactions. These multiple sources allowed for collecting rich data from the approached educational setting and bui lding the individual cases. In the following sections of this chapter, the data sources and data analysis procedures supporting the answers of the research questions posed for the group case study stage are described. In order to answer Research Question 1 which addressed the secondary sources, namely, IEP documentation and CMC logs. Resear ch Question 2 inquired about the patterns of dyadic interactions used by language learners in CMC environments. In order to answer this research question, the primary source used was the writing interactions, the data analysis of this primary source was informed by secondary sources: IEP documentation, observations, and student interviews. Finally, in order to answer Research Question 3 which aimed to address the factors that influence the CMC interactions, four qualitative da ta sources were

PAGE 181

155 interviews, and (d) non student participant data. In the following sections of this chapter, the procedures of data analysis are presented. The IEP Doc umentation Analysis The IEP documentation was used to provide a description of the educational environment and situate the group case study into a specific setting and to inform the data analysis and interpretations related to the research questions posed for this qualitative stage. Further, this source of data informed the design of the educational tasks and refined the focus the field observations. This source of data allowed the researcher to become more familiar with the data collection setting and to preserve the naturalistic nature of the study. Several documents were considered from this source: (a) Placement Test Outcomes Reports, (b) Intermediate Level (III) Writing Class Goals and Objectives, (c) the writing class textbook (Boardman & Frydenberg, 2002), and (d) the IEP CALL curriculum. The outcomes of the placement test used in the IEP (Appendix 1B & C) provided a basis for purposeful selection of participants of the appropriate level of proficiency. The first step was for the researcher to analyz e the writing class goals and objectives (Appendix 1 A) in order to assure that the goals of the research project do not contradict with the overall goals and objectives of the study. The researcher then discussed the research project goals and the writing goals and objectives with the non student participants, namely, the curriculum designer and the 2005 Fall semester writing teacher. Further, the researcher reviewed the writing textbook selected by the IEP curriculum designer to be used during the two se mesters of data collection and in collaboration with

PAGE 182

156 the Fall 2005 writing teacher created a tentative schedule according to which the various writing topics would be introduced to the class. This tentative schedule permitted planning for the individual st ages of the study. Finally, the CALL curriculum (Appendix 2) provided information about the way computer skills were targeted in the IEP and how the computers were incorporated in the language learning process. After a careful study of this curriculum, th e researcher who was also the CALL consultant for the IEP in the Fall semester of 2005 and the curriculum designer assured the consistency of the use of computers, as planned by the researcher with the CALL goals and objectives outlined in the CALL curricu lum. Field Observations Analysis As outlined earlier in this chapter, field observations were conducted on a regular basis and mainly targeted the computer laboratory sessions of the writing class. The journal narratives, which were based on the field note s taken during the observations, were analyzed line by line by the researcher. The goal of this analysis was to highlight the general emergent themes which would focus the data analysis and presentation and would allow identifying and describing systematic ally the dynamics of the writing class settings. After these themes were identified and verified with the non student participants, the researcher reviewed the narrative journal again identifying sub themes related to each of the general themes. Following is a detailed description of the outlined above data analysis process. The steps followed for data collection and analysis related to the class observations were: (1) creation of field observation notes during each observation, (2) transcribing and refinin g field observation notes immediately after each observation, (3)

PAGE 183

157 debriefing the content of the field observation note transcriptions with the non student participants on a regular basis, (4) refining field observation notes after the debriefing sessions, (5) creating narrative journals based on the observation notes for each individual session, (6) debriefing aspects of the journal narratives when needed with the non student participants, (7) collapsing separate observation narratives into a narrative jou rnal, (8) uploading the narrative journal to NVIVO, (9) identifying general themes, (10) identifying hierarchy of related sub themes, (11) generating reports using NVIVO (v. 2.0) (Qualitative Solutions and Research Pty., 2001), (12) using the generated rep orts for each general theme and related sub themes for writing the research report, and finally (13) debriefing the research report with a non student participant. For the graphical representation of this process refer to figure 3.5.

PAGE 184

158 Figure 3.5 Data Analy sis Flowchart: Field Observations As it is evident from the Data Analysis Flowchart displayed above, the process of the field observation data sources analysis was multistage and recursive. It took place during the data collection period and was extended through the data analysis period. After the field notes of each observation were transcribed, their discussion was part of the regular debriefing meetings with the writing teachers and additional details or

PAGE 185

15 9 clarifications, when necessary, were added. This process of recursive analysis and non student participant member check assured the data credibility and conformability ( Lincoln and Guba, 1985; Patton, 2002) Finally, after the composition of the research report, the report was discussed with the writing teacher who took part in both semesters in which data were collected in order to obtain a final member check of data report and interpretation. The total number of lines in the narrative journal text was 1 187. In order to assure systematic analysis of t he narrative journal text body which would allow the identification of the general themes and sub themes as well as their relation, the analysis was performed using qualitative data analysis software NVIVO (v. 2.0) (Qualitative Solutions and Research Pty., 2001). The software allowed for marking the general themes and the identified sub themes to the general themes. In addition, it allowed for marking portions of the narrative journal text to be used as illustrative examples, insert notes during the theme i dentification and create hierarchy of sub themes related to the general theme which were later displayed in the reports generated by the software. For example one of the general themes was working with software The following sub themes related to this the me were identified: use of ICA2, use of L.E.C.S., and use of Microsoft Word Further, for the first sub theme level use of ICA2 supporting lower level sub themes were identified: ICA activities, student participation, student postings, and difficulties T he ability to create such network of themes and display them in reports assured the systematic data analysis. A sample narrative journal report, generated after NVIVO (v. 2.0) (Qualitative Solutions and Research Pty., 2001) analysis is provided in Appendi x 6.

PAGE 186

160 The Treatment CMC Interactions Analysis The CMC interactions of the participants during the completion of the pre writing CMC discussion task served as a primary data source for answering Research Question 2 : What patterns of dyadic interaction do par ticipants manifest during the asynchronous and synchronous CMC interaction process? The pre writing CMC interactions of each dyad were available at the servers hosting the applications used for the CMC interaction: ICA2 (Nicenet, 1998) and L.E.C.S. ( Kanto Gakuin School Corporation, n.d. ). In order to identify the dyadic interaction pattern types and how they were influenced by the CMC writing exchange. The model for us ed pattern identification and analysis was the model of dyadic interaction proposed by Storch (2002). Storch (2002) propose the following patterns of dyadic interactions of second dominant, passive, novice, expert, collaborator These patterns, mutuality and equality of interaction (the graphical representation of the model is provided in Figure 4.1). Dominant is a pattern of interaction in which a person ta kes an authoritarian stance and appropriates the task. The passive pattern refers to interaction in which the participant has limited contribution to the discussion due to choosing to play a more passive role in the interaction process. The novice pattern is evident when an expert pattern takes place in the same interaction there is a sense of encouragement and collaboration between the participants with one of them being a leader. The collaborator pattern is evident in interaction situations in which the re is equality in the process of task completion and both participants work together through all stages of the task.

PAGE 187

161 The total of 30 CMC interaction logs (17 ACMC and 13 SCMC) were analyzed in order to identify the dyadic interaction for each case. The an alysis was performed line by line for each dyad log. The evidences for stances of interaction that would specify each in the data analysis are presented in Table 3. 3.

PAGE 188

162 Table 3.3 Stances of Interaction Sorted by Dyadic Pattern Dyad Pattern Type Dyad Interaction Stance Expert Novice more control; encourages the novice; offers assistance Col laborative/Collaborative work together on all parts; joined problem space; alternative views offered and both partners; cohesion of interaction (repeat & extend); nega tive & positive feedback Dominant Passive contribution; speech directed to the self; little assistance offered passive and/or subservient; fewer contributions or chal lenges; does not propose challenges; little assistance sought Dominant/Dominant inability to reach consensus; division of labor; few requests or collaborative completio ns

PAGE 189

163 In order to assure the credibility dependability and confirmability of the qualitative data, the researcher analyzed each interaction four times in a period of two weeks. The third and the fourth time of data analysis resulted in 100% intrarater agr eement in terms of dyadic pattern identification. Further, 12 of the CMC logs (40%) were analyzed by two independent raters. The agreement with both raters the level agreement was high 98% with rater one and 100% with rater two. The CMC logs analyzed f or patterns of interactions were used to initiate the partially ordered matrix for each case (Miles & Huberman, 1994). This matrix was further developed considering other sources of data presented in this chapter. The fields of the matrix are presented in Table 3.4.

PAGE 190

164 Table 3.4 Partially Ordered Matrix Fields: Group Case Study Participant CMC Mode Pulled z score Age Gender Education Interaction Pattern Interaction Pattern Interview Trend 1 Interview Trend 2 1 2 ACMC SCMC 0.02 0.08 2 5 19 M F College College Collaborator Dominant Collaborator Dominant Theme 1 Theme 2 Theme 1 & 2 Theme 2

PAGE 191

165 CMC Student Interviews The student ACMC interviews collected during the on going qualitative stage used as a medium of communication ICA2. The logs of the 18 interview questions and answers were available for each of the participant. The questions were grouped in a way with a specific trend. An example of such trend could be views on writing ; this trend was addressed in questions 5, 6, and 7 (refer to Table 3.2 for question types and trends representation). Within the trends, the themes that emerged in the interview data were identified and the research rep ort presented in Chapter IV was organized. For the purposes of data organization and analysis, the interviews were collapsed in one document. NVIVO (v. 2.0) software (Qualitative Solutions and Research Pty., 2001) was used for the data analysis. The follo wing steps were undertaken: (1) each individual log was analyzed within the proposed trends and the emergent themes for each interview trend were identified, (2) the themes were refined using the tree nod feature of NVIVO (v. 2.0), (3) interview report for each case was generated, and (4) the partially ordered matrix initiated during the Dyadic pattern of interaction analysis was further developed using the interview trends and themes. Thus, the partially ordered matrix contained the data analysis results pertaining to (1) the dyadic pattern of interaction exercised by each participant during the treatment CMC pre writing interaction, and (2) the interview trends and the emerged themes. The matrix was further extended including (3) demographic data for each participant: age, gender, and education, (4) the CMC mode of interaction of each participant, and (5) the pulled z score difference between the pre treatments and the post treatments writings

PAGE 192

166 were included. Table 3.4 presents the fields of the partially o rdered matrix, note that the interview trends are marked as Trend 1 or 2; the full list of trends is presented in Table 3.2. These trends were used to organize and present the Computer Interview Outcomes section in Chapter IV. The partially ordered matrix was created using Excel spreadsheet and vertical fields during the data analysis. The researcher strived to move from lower level concepts presented by the themes for eac h of the trends to higher level theorizing. In order to achieve this higher level theorizing, the comparisons of the emerging themes were applied systematically for each participant and across all participants in the group case study (Patton, 2002). Sub s tage II: The Instrumental Case Studies This section presents the steps undertaken for data analysis of the instrumental case studies, a detailed description of the analysis process for each case is provided in Chapter V. For the individual instrumental ca se studies, after the qualitative participants were identified upon the quantitative and qualitative stage data analysis, the eight individual cases were selected based on the pulled z scores obtained during the quantitative data analysis. As stated earlie r in this chapter, Extreme/Deviant sampling technique was applied (Kemper et.al., 2003). This technique allowed the identification of the extreme cases in the study. The actors of these cases were the students who gained the most and the least in terms of writing skills demonstrated in their first paragraph drafts written after a CMC pre writing discussion (treatment) as compared to the pre treatment draft written after a face to face pre writing interaction. Eight high and low gain actors were selected, two of

PAGE 193

167 each gain level and CMC mode of pre writing interaction. The researcher analyzed the CMC pre writing interaction logs and the first drafts of the case study participants; the goal was to identify the interaction patterns they demonstrated and link t hose patterns to the intertextual connections between the logs and the first drafts produced by the participants. Further in this section the specific steps of data analysis are described; the graphic representation of these steps is provided in Figure 3.6 The analysis aimed to answer the research questions posed for this sub stage of the study: Research Question 1: How do peers participate in asynchronous and synchronous pre writing CMC interactions? Research Question 2: (a) How do they use the specific i deas and language generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas? In order to assure the credibility dependability, and confirmability of the data analysis the re searcher reviewed the identified themes and intertextual connections several times refining the analysis. The last two review rounds resulted in 100% intrarater agreement; in other words, in the last round no changes were made to the data analysis as perfo rmed in the previous round. Further, a trained interrater reviewed the analyzed data confirming that the themes and intertextual connections identified by the researcher High inter rater agreement was achieved.

PAGE 194

168 Figure 3.6 Data Analysis Flowchart: Instru mental Case studies Analysis of observation narratives and interview records In order to elicit data related to the targeted case studies, the researcher first examined the observation journal narratives and pulled out the sections which pertained to th e actors of the instrumental case studies. These observation data were analyzed using NVIVO (v. 2.0) software (Qualitative Solutions and Research Pty., 2001) in order to

PAGE 195

169 identify aspects that would support the presentation of the actor. NVIVO reports were generated. Further, the participant record of each actor was pulled from the partially ordered matrix created during the Sub stage I: Group Case Study The interview themes identified in the matrix were examined. The third step was to relate these themes w ith the overall body of observation data and further examine the themes within the context of the g and CMC were presented. An ordered matrix was initiated based on the outcomes of this data analysis (Appendix 11). The CMC Interactions Analysis The next phase of the instrumental case study data analysis entailed the CMC logs generated during the pre w riting task as a primary source of data. The CMC log of each individual actor was analyzed in order to identify the descriptive characteristics of the CMC pre interaction ident ified during Sub stage I and explain it within the context of the overall views related to the CMC interaction task as shared in the interview. This analysis resulte d in the identification of concrete themes related to the specific aspects of the interaction. These aspects were: (1) number of posted messages, (2) length of postings, c task as shared during the interviews with the researcher. Considering these aspects, the ordered matrix was extended with additional fields.

PAGE 196

170 The Paragraph Inter textuality Analysis writing interaction in which the actor and his/her partner participated during the pre writing stage. The comparative analysis of the paragraph and the pre writing log was p erformed line by line. The intertextual analysis for each of the cases was performed on three separate levels. These levels included: (1) lexical level, (2) organizational level, and (3) textual level, with idea units being used as a unit of analysis. The intertextuality on a lexical level was analyzed based on the matching distinct lexical item (DLI) scores obtained during the quantitative stage of the study. For the purposes of the analysis, the following definition of DLI was used: the same lexical item s which appeared one or more times in either the CMC interaction or the post treatment paragraph were considered to present one DLI. The matching DLI sores represented the lexical items that were used both in the pre writing discussion and the following it draft. The consideration of the DLI scores obtained by the 60 participants in the Qualitative Sub stage I and quantitative stage of the study allowed the researcher to make suggestions about the strength of the pre writing discussion and the post treatmen t paragraph connections on a lexical level for each of the instrumental case study actor as compared to each other and to the overall language learning community of which this actor was a member. The organizational level analysis was conducted through lin e by line comparison of the texts of the CMC interaction log and the draft produced after this interaction. The comparison aimed to reveal the connections of the pre writing discussion organizational suggestions to the draft organization. In addition, the analysis accounted for the

PAGE 197

171 sequencing of ideas in the CMC discussion and whether the same or similar sequence was evident in the draft. This analysis allowed the researcher to suggest possible connections between the pre writing discussion and the post tre atment paragraphs on an organizational level. Finally, the textual (idea unit) level analysis of intertextuality was performed through segmenting the pre writing discussion texts and the post treatment writing into idea units. The following working defini tion of an idea unit was used in the analysis: an however, if a complement or re strictive relative clause is present or indirect question or indirect quotation is used, these belong to the idea unit presented by the main clause. for example: New York, Hou ...), and participial clauses (past and gerund) were considered to be separate idea units. When analyzing CMC discourse production, single phrases indicating agreement, disagreement, understanding, etc. were considered as separate idea un its unit from the pre writing discussions was compared against the idea units from the draft texts. The intertextual connections between the discussion and draft idea units were interpreted and new idea units composed in the draft were identified. During the analysis of the intertextual connections, the ordered matrix was further late the matrix horizontal and vertical fields during the data analysis.

PAGE 198

172 Comparing the Instrumental Cases The final step of the data analysis was to compare the cases accounting for the identified themes, interaction patterns, and intertextual connections. This comparison was performed first within each CMC mode/gain level dyad, then within each CMC group. Finally, a cross case comparison was performed. The partially ordered matrix allowed the researcher to have easy access to the data analysis outcomes in the process of describing case comparison. Combining the Quantitative and Qua litative Analysis The results of the quantitative stage of the study were considered during the qualitative stage of the study in order to support the multiple case analysis. In addition, the analyses of the qualitative and quantitative data revealed new a spects of the studied phenomenon which prompted the researcher to extend the quantitative stage with additional analysis. This additional analysis is addressed in detail in the following chapter. Thus, the quantitative and the qualitative findings complem ented the discussion by providing information about trends in writing skills acquisition demonstrated in the be situated within a richer context through offering multiple dimensions to the description of the writing skills acquisition processes and provision of a richer basis for data presentation and conceptualization. This in turn supported the researcher to accomplish a higher level of theorizing based on the o verall findings of the study.

PAGE 199

173 Chapter Summary This chapter presented a description of the study design, flow, and data analysis procedures. Reasons for applying sequential mixed method design were provided, and justification for combining qualitative and quantitative techniques was discussed. The instruments used in both stages were described along with the role of the researcher in the study. The quantitative and qualitative data analysis procedures used in the study were described in order to explain ho w the different sources of qualitative and quantitative data were approached.

PAGE 200

174 CHAPTER IV: THE GROUP CASE STUDY STAGE AND THE QUANTITATIVE STAGE ANALYSES AND RESULTS Introduction This study aims to investigate intermediate composition process from the perspective of the Writing as a process approach. More specifically, the processes that occur during the first two pre writing stages of the five stage process: the pre writing and drafting stage (Grabe & Kaplan, 1996) were in the focus of the investigation. The analysis accounted for the social factors such as learning and task writing task, as well as for certain deep and surface textual feat ures as presented in students were examined, this examination was informed by socio cultural theory, more specifically the dyadic interactions of the participants w ere examined within the context of the model of dyadic interaction proposed by Storch (2002). In this chapter, the results of the qualitative and quantitative data analysis are addressed and discussed. The qualitative data were collected during the pre an d post quantitative stage of the study as well as during the nested quantitative stage, more specifically during the treatment. The quantitative data were collected during the nested quantitative stage, more specifically during the pre treatment, treatment and post treatment sessions.

PAGE 201

175 In the first section of the chapter Qualitative Analysis: Sub stage I the computer laboratory sessions are presented and discussed. Further, the CMC treatment pre writing discussions are presented in the context of interacti on patterns. The last qualitative section outlines the outcomes of the CMC interviews conducted during the semester. A discussion of the qualitative data analysis outcomes is presented at the end of the Qualitative analysis section. The goal of these secti ons is to build a background within which the results of the quantitative stage and the instrumental case studies, discussed in Chapter V, are presented, as well as to answer qualitative questions 1 3 which were: (1) he role of CMC in the process of establishing their writing skills? (2) What patterns of dyadic interaction do participants manifest during the asynchronous and synchronous CMC interaction process? (3) What are the factors that influence the CMC pre writin g interaction process? How do these factors influence the interaction process? In the subsequent section of this chapter, the results of the nested quantitative stage are provided. Further, the research questions posed for the nested quantitative part are answered and discussed. Finally, the additional findings of the research are presented. First, the additional quantitative research question along with the methodology for this additional textual analysis are described. Next, the results of the textual a nalysis are presented within the context of the overall nested quantitative section. The chapter is concluded with the discussion of the overall findings.

PAGE 202

176 Qualitative Data Analysis: Sub stage I Computer Laboratory Sessions: Behavioral Patterns and Class Dynamics The student behavior and class dynamics of the computer laboratory writing class sessions were observed, and these observations were documented. During 14 sessions of the fall semester of 2005 and eight sessions of the spring semester of 2006, th e researcher assumed the role of observer. A detailed observation journal was kept; the notes from this journal were subsequently analyzed. When analyzing the computer laboratory observations notes, the following general themes emerged: (1) Class dynamics, (2) Class Management Issues, (3) Task Completion, and (4) Working with Software. Class Dynamics The two sections of the Level III Writing class whose face to face interaction was closely observed during Fall 2005 had distinctly different dynamics in thei r regular classes which were transferred to the computer laboratory class. One of the sections was composed mostly of continuing students; they had an excellent relationship with one another even prior to the beginning of the Fall 2006 semester. The writin g instructor, who class and outside of class. This was also evident from the research Thirteen instances of clear positive attitude demonstrated by students were identified in the observation notes; they included various behavioral trends that indicated motivation, readiness to work, and positive attitude towards the wri ting class. Students were energetic and expressed their readiness to work in the beginning of many of the computer laboratory sessions observed. They were mostly relaxed during the

PAGE 203

177 class sessions. The low anxiety was demonstrated by active participation w hen teacher student discussion was involved, mostly in the beginning of the computer laboratory sessions. The students were active in providing their opinion when class discussions were held, they were not reluctant to ask questions and request assistance when they needed it. In addition, they quickly accepted the new student in the section and made her feel part of the group starting with the first computer class session observed. However, in this section, a few instances of students experiencing stress w ere observed. In the beginning of the semester, the new student expressed reluctance to participate in the assigned task which involved posting a reply to the welcome message of the teacher using ICA2. Although she demonstrated good computer skills in term s of Internet navigation and typing, she was reluctant to proceed with the task. She asked the teacher about the purpose of registering to ICA2. Almost immediately, one of the continuing students who was familiar with the courseware volunteered to guide he r through the first steps of the task. Another student demonstrated stress during the treatment timed writing session and had difficulties completing the assignment. He spent the majority of the time composing his first sentence. However, he was not reluct ant to teacher and told her that he cannot organize his ideas. Due to the fact that this was the controlled post treatment writing session, the teacher did not provide h im with specific help but reminded him about the requirements of the task. Another trend that the researcher noticed in the class dynamics during computer laboratory face to face discussions was that some of the students in the class would lose their atte ntion focus during class discussions. This behavior was observed when the

PAGE 204

178 discussion took place immediately before an assigned writing task. Some of the students would attempt to start using the computer before the task was fully presented; others would st art reviewing the notes they took previously in the face to face class sessions; still others simply would not participate in the discussion and would attempt to open websites not related to the task. Compared to the first section, the students of the seco nd section of the Fall 2005 Level III Writing class appeared more alienated. This was evident by their behavior during face to face classes and transferred to the dynamics of the computer laboratory sessions. This observation was confirmed by their writing instructor as well. During a debriefing session, the teacher shared with the researcher that, in the beginning of the third week of classes, there was a loud argument between two of the male students in the class; these men did not talk to each other for the rest of the semester. Two of the six young men in the group were mostly communicating with each other during the breaks between classes; they seemed not to care about the other students in the class. Usually, before computer lab class sessions, the mal e students would sit in front of their computers either browsing the Internet, listening to music or chatting with friends from their countries. The women formed two distinct groups that would be together during the breaks. During the short class discussi on sessions in the computer class, there were several students who volunteered their answers; however, they mostly addressed the teacher, rather than engage in an exchange with one another. Typically, the students of this section were mostly quiet during t hese discussions and waited for the teacher to call

PAGE 205

179 that depicted a more lively contribution to a class discussion of two students when one volunteered an answer, and t his answer was supported and further extended by another student. There were instances when some of the students openly expressed boredom or frustration during task completion: they would attempt to access sites not related to the task, to listen to music using the computer or a portable MP3 players, or demonstratively abandoning the task pushing aside the keyboard and staring in front of them. The researcher did not notice the same strong contrasts between the sections of Lever III during the Spring 2006 semester. One of the reasons for this might be that there were fewer class sessions in which the researcher was involved only as an observer. However, none of the writing instructors indicated that there was an exceptional closeness or alienation in their sections. In addition, for the majority of the students this was their first semester at the IEP; there was a large group of students from the same country who knew each other before coming to the program. These students tended to stay together most of th e time. Class Management The class management in a computer laboratory class had its own specifics related to the fact that for the majority of the tasks, the computer was the main instrument. In the Spring semester of 2006, the computer laboratory sessio ns took place in two different computer laboratories. In the small computer laboratory used in the Spring 2006 semester, the computers were was set up in a way that the monitors were hiding the e front of the class. It is possible that this setting created a sense of privacy and limited the ability of the instructor to observe the actions of the computer users. Although the computers in the large

PAGE 206

180 computer class were set up in a way so the monitor screens did not hide the students from the teacher, students also seemed to feel their working places in the computer laboratory more private as compared to a regular classroom, this assumption was based on the observations of more frequent attempts to en gage in activities not directly related to the assigned task. More specifically, the researcher identified five behavioral trends related to class management during different computer laboratory classes: (1) multitasking using either an instant messenger o r an Internet site, (2) keeping portable MP3 players or computer headphones on during task completion, (3) beginning the work on the computer preliminarily, (4) taking notes using Microsoft Word when the teacher explicitly asked to close all computer appli cations, (5) temporary or permanent abandonment of assigned task. These trends were observed during the work of different sections of the writing class and were confirmed by the instructors as well. There were multiple attempts of multitasking during the computer laboratory classes which mostly involved using chat software and Internet sites not related to the class activities. While most of the students would close their extraneous windows on the computer once the instructor requested this, others would s imply minimize them or hide the bottom task bar so the minimized windows were not visible unless the cursor was positioned in the bottom area of the screen. The researcher documented nine specific instances of such behavior in her observation journals; thi s behavior was noticed by the instructors as well, and an explicit policy that would forbid the use of unrelated websites during the computer laboratory class sessions was stated by each of the four instructors teaching the writing class during the Fall 20 05 and Spring 2006 semesters.

PAGE 207

181 In addition to browsing Internet sources not related to the assigned activities, there were three students who would routinely stay with i pod or computer headphones on during task completion and sometimes even during class di scussion. This behavior was repeated during several computer laboratory sessions. When the instructor asked the students to take the headphones off, they would comply with the request; however, they would attempt this behavior again in subsequent computer laboratory sessions. Interestingly, this was a trend observed only in the section of the Fall 2005 writing class in which the students had alienated relationships. The researcher noted seven specific instances when students would start working on a writin g task before the task was fully presented by the instructor and discussed with the class. During these class discussions, students also attempted to take notes using Microsoft Word, although the instructors discouraged this; three of the observed students used Microsoft Word for note taking, and all of them had strong typing skills. Finally, it was observed that some students abandoned the assigned task or hurried to complete it. There were eleven specific instances of such behavior documented by the resea rcher. Five of these instances were observed: one during a small group discussion and four during class discussion. The researcher contributed this higher number of preliminary task completion or task abandonment during class discussion to the way the comp uters were positioned in the laboratory partially hiding the students applications not related to the task. Further, in one case, during individual work on the comput er, the task was temporarily abandoned because the student did not know how to perform an action on the computer that would take him to the assigned online material.

PAGE 208

182 He had difficulties positioning the cursor exactly on the link posted on ICA2, and after m ultiple attempts quit trying but did not request assistance. In four instances, during individual task completion, the students shifted to a software application that was not related to the task (Instant Messenger, the Internet, Microsoft Word, and Windows Sound recorder); two students engaged the teacher in a conversation not related to the writing class prior to beginning work on the task or during the task completion; two students abandoned a writing task for which the teacher did not request submission at the end of the class. When asked why they stopped writing, both answered that they would complete the task later. Even though speculating if the task was fully completed is difficult, in three of the observed instances of preliminary task abandonment, t he researcher checked the task outcomes were not revised. Working with Software During the pre quantitative stage of the study, the students were trained to use tw o software applications: ICA2 and L.E.C.S. These applications were used for the completion of the treatment task. In addition, students worked extensively with Microsoft Word to compose and format their writings. In order to assure that the participants wo uld become familiar with these applications and that they would get the necessary online communication skills, several activities that incorporated these applications were assigned during the pre quantitative stage. The trends demonstrated during the comp letion of these activities were observed and analyzed by the researcher.

PAGE 209

183 ICA When working with ICA2, students generally experienced very few problems. In addition to the user friendly design of the application, the reason for the fast acquisition of this courseware functions could be attributed to the fact that this application was used in other classes as well for the reason that it was the suggested courseware at the IEP. However, there were several trends that were observed as students were workin g with ICA2. One of the first activities performed on ICA2 included exploring web resources related to language learning; the links to these resources were posted by the teacher in the Link Sharing section of ICA2. The second part of the task was for the students to post links that they had found useful. None of the students posted their own links. In informal conversation, three students shared that the reason they did not post was that they did not have the time they found the sources posted by the tea cher to be very interesting, and they spent most of the time reading them, which left them with very little or no time to search for other links. In addition, they were not sure that they could find language learning sources on the Internet of the same qua lity as the ones provided by the instructor. Multiple tasks required students to post on the conference section of ICA2. The researcher observed that the students would type their postings directly into the application although ICA2 does not offer spell c hecking and the students at this level have significant difficulties with spelling. This difficulty was also confirmed by the interview outcomes: students reported spelling to be one of the difficulties they face when writing in English. Only one of the ob served students used Microsoft Word to compose her posting and then copied and pasted it into the ICA2 conferencing section.

PAGE 210

184 Out of the 103 students enrolled in the writing classes during the Fall 2005 and Spring 2006 semesters, only three were observed t o experience difficulties navigating through ICA2 after the initial training was completed. Two of them had difficulties finding the resources links posted on the Link Sharing section; the third one had difficulties opening links posted on the Link Sharing section. Another minor difficulty was that initially students would not fill in all the registration information when signing in their postings. This problem was i dentified and addressed by the teachers. L.E.C.S. The students completed several synchronous communication assignments using L.E.C.S. The majority of the questions asked by the students during the completion of these assignments were related to language a nd vocabulary. In the first class that used L.E.C.S., there were two occasions of students not being able to start their synchronous communication on L.E.C.S. because they did not post in the same L.E.C.S. chat group. This problem was related to the way th e L.E.C.S. functions for creating chat groups were set up: each group was automatically assigned a consecutive number, and account when providing instructions for chat gr oup access for the other writing sections and especially during the treatment task completion. In addition to the above, during the pre quantitative stage, the researcher documented one instance of students assigned to the same chat dyad to attempt communi cating face to face after they realized that they were sitting close to each other. When asked at the end of the class for the reasons for attempting to talk to each other rather than to chat during the task completion, the students said that it took them too long to type their responses. Although these two

PAGE 211

185 students had relatively good typing skills (2.5 and 2 on a 3 point scale), they perceived it to be time consuming to complete the task through chat discussion. This theme emerged again in the student int erviews: when sharing their opinions about the treatment pre writing CMC task, four students explicitly stated that CMC is time consuming. Another trend related to typing was that students would be impatient while o instances of complaints to the teacher about slow posting were documented by the researcher during the pre quantitative stage. Other students chose to multitask, switching between the chat window and a website not related to the chat. During the pre quan titative stage, three people were observed multitasking; however, the researcher believes that this number might be higher considering that multitasking was one of the class management issues observed by the researcher and shared by the teachers. Microsoft Word While most of the students were familiar with Microsoft Word and had at least beginning typing skills, there were two students, one who was enrolled in the fall of 2005 and the other in spring 2006 who, in addition to their very limited computer ski lls, had no experience working with text processing applications. However, these students progressed relatively quickly with their computer skills due to the intensive exposure to computers at the IEP: most of the classes at the IEP have at least one compu ter laboratory session per week. The researcher noticed that most of the students would prefer writing by hand rather than on the computer, if they had the choice. If the writing task was assigned to be completed on a computer, students would draft their writings by hand and then start working on a Word document. This could be an indicator of their perception of the

PAGE 212

186 computer as a tool used in more advanced stages of the writing process. Students would prefer printing texts they were assigned to read and ed it or paraphrase, rather than working with the texts directly on the computer. However, three students were observed to attempt using Microsoft Word for note taking during the very few and short teacher centered presentation in the computer class. All of t hese students were proficient in their use of Microsoft Word. The function of Microsoft Word that students were observed to use the most was spell checking, several of them used the Microsoft Word Thesaurus, and one was observed to use word count. These ob servations were confirmed by the interview outcomes as well: the Microsoft Word features most frequently reported to be used during the post treatment task were spell check followed by Thesaurus. All of them were expected to format their texts using functi ons of Microsoft Word related to formatting. When working with computers the students demonstrated different levels of computer skills. Some of the students were more proficient and had considerable experience using computers while others (fewer) had to s tart developing their computer skills at the IEP. Based on the observations conducted during the pre quantitative stage, the researcher concluded that the less computer skills proficient students were forming their computer skills quickly during the first weeks of the semester. By the fifth week of the semester, when the study experiment took place, they were familiar and worked comfortably with the three applications used in the study: ICA2, L.E.C.S., and Microsoft Word.

PAGE 213

187 The Treatment CMC Interactions CMC Discussions: Specifics The CMC pre writing discussions were examined and the outcomes of the discussions performed in the two modes of CMC were compared in terms of language produced and patterns of interactions used in these discussions. The followin g descriptive characteristics of language specifics were examined: (1) number of turns, (2) mean length of turns, (3) length of discussions measured with average number of total words, (4) number of distinct lexical items (DLIs), (5) mean length of word st rings between end of sentence punctuation. Descriptive characteristics of CMC interactions When examining the CMC discussions, it was noted that the CMC interaction that took place during the pre writing discussion differed in several ways depending on t he CMC mode. It could be inferred that the synchronous communication in general was more interactive than the asynchronous. The SCMC discussions were more dynamic and involved the generation of more language with a higher number of partner turns and distin ct lexical items used by the partners. The texts produced during the SCMC discussions were generally longer: the SCMC discussion mean length was 410.46 words, while the ACMC was 272.76 words. In the SCMC discussions, students used more distinct lexical it ems (DLIs) the mean number of DLIs was 92.15 as compared to the ACMC discussions which had mean number of 66.24 DLIs. However, the sentences produced during the ACMC discussions were easier to detect due to more consistent use of end of sentence punctuat ion and

PAGE 214

188 complete sentences, while the SCMC group used more utterances consisting of phrases or incomplete sentences and the punctuation was often irregular. The average length of word strings between end of sentence punctuation marks or end of posting wer e computed. This computation revealed that the ACMC group used longer strings of words between end of sentence punctuation marks or end of posting, with an average of 12.66 words, whereas the number of words in the SCMC discussion utterances, signified by end of sentence punctuation (period, exclamation or question mark, three dots) or end of posting (without a punctuation mark), was 6.92 words. In addition, the exchange of students who participated in ACMC and SCMC discussions differed by the number of po stings. For the ACMC discussion, the mean number of postings was 5 per dyadic interaction while in the SCMC discussion it was much higher 46.46. As compared to the SCMC, ACMC postings used longer chunks of text. The average length of ACMC postings was 5 4.98 words per turn, whereas the average length of the SCMC postings was 8.85 words. The sentences in the ACMC postings were mostly organized in short sequences that sometimes resembled paragraphs presenting a topic sentence and supporting ideas and other times were ideas outlines often formulated in complete sentences and presenting the topic, the controlling idea, and supporting ideas for the paragraph, while the SCMC participants did not organize their ideas in a paragraph like or outline formats. Thus, the pre writing discussions, when performed in ACMC mode, shared characteristics with outlining and producing a first draft, while the pre writing discussions performed in a n SCMC mode tended to concentrate on generation

PAGE 215

189 of ideas related to the writing top ic. Table 4.1 presents the descriptive statistics of ACMC and the SCMC discussions outlined above.

PAGE 216

190 Table 4.1 Pre writing Discussion per Dyadic Interaction: Descriptive Statistics n Mean Number of Turns Mean Length of Turns Mean Number of Words Mean Numb er of DLIs Mean Length of Sentence ACMC 17 5 54.98 272.76 66.23 12.66 SCMC 13 46.46 8.85 410.46 92.15 6.92 Patterns of Dyadic Interactions The dyadic interactions were examined in order to determine the specifics of the interaction patterns for each g roup. The researcher used the model of dyadic interaction proposed by Storch (2002). This model was discussed in detail in Chapter 3. Whereas Storch created and described the model within the context of face to face dyadic interactions, in this study it wa s applied to CMC dyadic interactions. The researcher found some similarities as well as differences between the dyadic patterns of interaction in face to face environments reported by Storch (2002) and the CMC interaction patterns. The modified model along with the main indicators manifested in each pattern of interaction in the CMC environment is presented in Figure 4.1 and discussed further.

PAGE 217

191 Figure 4.1 Dyadic Interaction and CMC Patterns Indicators, modified from Storch (2002) High Mutuality Low Equ ality Expert more control encourages the novice offers assistance Novice position repeats little contribution Collaborative/Collaborative work together on all parts joined problem space alternative views offered and discussed willing to engage with each leading resolutions acceptable for both partners cohesion of interaction (repeat & extend) Negative & positive feedback High Equality Dominant authoritarian appropriates the task long monologues little attention to contribution speech directed to the self little assistance offered Passive passive and/or subservient fewer contributions or challenges does not propose challenges little assistance sought Dominant/Dominant both contribute inability to engage w ith each inability to reach consensus division of labor few requests or collaborative completions Low mutuality were considered and identified. Th e most dominant trend of interaction observed was the high mutuality and high equality trend which led to a collaborative pattern of interaction. A total of 12 pairs demonstrated this trend as the main trend of interaction. Whilst this

PAGE 218

192 type of interaction was predominant in the SCMC discussion with nine out of 13 pairs using mainly collaborative interactions (69.23%), for the ACMC discussion it was a much less prevalent type of interaction: only for three out of 17 pairs (17.65%) was this interaction patter n identified as prevailing. The interaction pattern observed with the highest number of ACMC dyads was high equality and low mutuality which led to Dominant/Dominant discussion for eight dyads out of 17 (47.06%) this pattern of interaction was recognized as the prevailing one. Table 4.2 presents the patterns of interaction used in each CMC environment and the number of dyads for whom a specific pattern of interaction was prevailing. Table 4.2 Dyadic Patterns of Interaction by CMC Group Identified in Dyad ic Pattern ACMC Dyads SCMC Dyads FTF (Storch, 2002) and CMC interactions Collaborative/Collaborative 3 9 Dominant/Dominant 8 0 Dominant/Passive 1 1 Expert/Novice 0 1 CMC interactions only Expert/Dominant 0 1 Collaborative/Dominant 5 1 Based on the figures presented in Table 4.2, it could be concluded that the distribution of dyadic patterns across the two environments was different. Low mutuality seems to be more prevalent in an ACMC environment nine dyads demonstrated a type of interaction th at had low mutuality. However, the majority of these dyads were highly

PAGE 219

193 equal in their discussion, thus demonstrating a pattern of interaction in which both participants dominated the discussion. Further in this section, the observed patterns of interaction are illustrated with CMC logs from the treatment CMC transcripts. The collaborative dyads Log 4.1 presents an example of a collaborative interaction pattern in an ACMC environment. This CMC log is a part of the treatment CMC interaction between Anna, a f emale student from Russia, and Rashid, a male student from Saudi Arabia. Several trends of collaborative pattern of interaction were identified in the interaction of this dyad that are representative for the ACMC collaborative dyads. These two students we re able to create a common discussion space. Examples that illustrate the common discussion space of the dyad are evident in lines 2 11: Anna stated her view about the pictures (lines 2 and adding detail (lines 9 12). He took the next step and directed their discussion further by asking Anna to elaborate on her previous posting (lines 15 16). In two minutes he reposted his request (lines 18 20) adding the name of his partner to the greeting line (line 18). This could be interpreted as taking time to read his posting and to edit it in order to show partner consideration. Next, Anna posted a rather lengthy string of sentences (lines 22 34). Storch (2002) indicates lengthy contribution to discussions as an in dicator of opinion: (1) she started it with addressing her partner (line 22), (2) she directly went into providing the information requested by the partner (lines 23 25), ( 3) she was constructive in terms of progressing through the task and suggested a focus topic (line

PAGE 220

194 28), (4) she asked Rashid to provide his opinion about a specific topic for the paragraph and stated her suggestion (lines 26 34). In this last posting, she addressed multiple paragraph organization, including a topic sentence and supporting ideas. Line 26 contains explicit request for opinion; further, in line 28, the ph showing that she was expecting her partner to contribute t o her paragraph construction ideas. Finally, she ended her posting with a question to her partner (line 34). Thus, the interaction of this dyad shows a joined problem solving space within which the dyad members attempt the task completion. CMC Log 4.1 ACM C Collaborative Discussio n FROM: Anna (02/19/06 11:47 AM GMT 06:00) 1 I think two pictures on the bottom shows a lot of people live in THe USA.But it too 2 simple I mean they are big differense betven Alaska and NY area but it not for 3 paragraph 4 The sekond t wo pictures about people in America have different home, different 5 neiborhood becouse they came from different country seesecond picture. 6 The last two pictures explane this county has big city on bouth costs. 7 FROM: Rashid (02/19/06 3:03 PM GMT 06:00) 8 The picture describe that the U.S is very big including alascka and hawaii and very 9 populated .People live in big cities as well as townhomes or small villages. It also 10

PAGE 221

195 describes that it has a variety of mixed people from different countries, that most 11 pe ople live in big cities such as NY Los angeloes and Chicago. 12 FROM: Rashid (02/19/06 3:08 PM GMT 06:00) 13 HI MY FRIEND ......... 14 I HAVE TO ASK YOU ABOUT WHAT YOU MEAN BY THE DIFFERENTS 15 BETWEEN ALASKA AND N.Y ? IS IT THE AREA OR THE CULTURE ? 16 FROM: R ashid (02/19/06 3:10 PM GMT 06:00) 17 HI MY FRIEND ANNA......... 18 I HAVE TO ASK YOU ABOUT WHAT YOU MEAN BY THE DIFFERENTS 19 BETWEEN ALASKA AND N.Y ? IS IT THE AREA OR THE CULTURE ? 20 FROM: Anna (02/19/06 5:32 PM GMT 06:00) 21 hai,my frend! 22 I think alaska and Havai have so different clumat and besouce of that it people have 23 realy different life and climate have changed peple. 24 But I think we have to consentrate more people came from different country. 25 Rashid, what do you think about pictures, and we have ( 2) main and supporing 26 ideas. 27 I think topic people in the USA. 28 Main melting pot. 29 1.The basis culture and language from England. 30 2.People came from different country also most not from England. 31

PAGE 222

196 3.They live, work togever and also they are merrid. 32 Mayby topic can be America has own culture. 33 What about people? 34 Log 4.2 illustrates a collaborative interaction pattern in a n SCMC environment. Mina, a female student from Colombia, age 28, and Dimo, a male student from Saudi Arabia, age 28, were the participants in this exchange. As compared to the ACMC sample, the interaction was more dynamic, the postings were higher in number and shorter in length. Lines 1 11 present their exchange of ideas about the pictures; Mina described one picture and her partner took over a nd provided the description of the next. They demonstrated their consideration of each other in several different ways: Dimo ested that he was sharing his opinion, implying that fourth picture (he shared it in lines 3 and 4). When discussing the fifth picture (lines 9 and 10), she addressed she based her conclusions. Further (lines 12 23) they bounced ideas back and forth about high and the p articipants demonstrated high consideration of each other. CMC Log 4.2 SCMC Collaborative Discussio n Mina: The third one, represent the towns, how the cities are organized, 1 the houses ... but in small cities... 2 Dimo: and in the other pictures i see people from diffrent country( the 3

PAGE 223

197 have diffrent flags) they came to the U.S. 4 Dimo: the third one i think is the origine country of the immigrants 5 Dimo: small country and village 6 Mina: The fourth one represent all the imigrant people who live in USA, 7 this people are from all the world 8 Mina: Te fifht one is like the big cities ... you see! there are a lot of big 9 buildings and a big bridge 10 Dimo: yes you right 11 Mina: And the last one is about the geography .... what do you think about the 12 last one? 13 Di mo: what about the last one what you think? 14 Dimo: maybe it show the big stat or citys where the most immigrants 15 lives 16 Mina: yeah, it maybe is the important cities for USA .. 17 Mina: and what do you think about the People of USA? 18 Dimo: i think that the p ople in U.S. is a mix of many diffrent origine 19 Mina : I think people in USA are from all over the world...the people 20 came to USA looking for a better life and opportunities that maybe in they own 21 country they couldn't find. 22 Dimo: that's true 23

PAGE 224

198 In both ACMC and SCMC interactions examined in the study, the collaborative pattern demonstrated communication during which learners were engaging in the seemed to be acceptable for bot h members of the collaborative dyads. The Roschelle, 1993) by approaching together each part of the task and engaging with e and flowed logically. Although in the ACMC discussions, the postings were longer as compared to the SCMC and sometimes one posting would address several aspects of the task, the participants Further, they positioning these interactions within the quadrants defined by the two continuums of the dyadic interaction model, mutuality and equality (Storch, 2002), the collabora tive dyads demonstrated high mutuality and high equality. The dominant/dominant dyads There were eight dyads that demonstrated (2002) model, both of the participants in t hese dyads were identified as dominant. The high equality/low mutuality pattern was the predominant pattern in the ACMC interaction eight out of 17 dyads performed the pre writing task demonstrating a Dominant/Dominant pattern. Thus, 47% of the dyads who communicated via ACMC demonstrated this type of interaction pattern. Log 4.3 presents an example of Dominant/Dominant interaction in an ACMC environment.

PAGE 225

199 CMC Log 4.3 ACMC Dominant/Dominant Interactio n FROM: Shin (10/14/05 9:24 AM GMT 06:00 1 Yes, I'M t oo. 2 FROM: Shin (10/16/05 4:14 PM GMT 06:00) 3 I think that this is a my topic sentence;In the U.S, most people are living in the city 4 and people came from different countries and have different appearance. 5 Supporting ideas; 1.Most people live in city because they can get a lot of 6 advantage. 7 2.people have different appearance European, Asian,Africian. 8 3.people experience many kinds of culture. 9 FROM: Sebastiano (10/16/05 4:56 PM GMT 06:00) 10 ok, here is what i think. 11 for the topic sentence i think is better this: 12 topic: now the people in the united states live as a friendly and united country with 13 all the ethnic groups without segregation. 14 for main idea: we should talk about the way of life of us people. 15 and supporting ideas: we can talk about all the cultures in US and the most 16 important cities, and also we can talk about the life in the suburbs and farms 17 FROM: Sebastiano (10/16/05 7:22 PM GMT 06:00) 18

PAGE 226

200 according to the images we can talk about all the people in the whole country, we 19 cant talk about too about the lifestyle in suburbs and farms and also in big cities, 20 and also teh msot important caracteristics of this big cities, we cant talk about the 21 "melting pot" the compilation of all the cultures 22 FROM: Shin (10/17/05 10:02 AM GMT 06:00) 23 Hi! 24 Could you tell me more about "melting pot"? 25 FROM: Shin (10/19/05 8:12 PM GMT 06:00) 26 You have a lot of idea. Thank you for your suggestion. Iwill write in my paragraph 27 such as population in theU.S,i mmigration,multiculture,urbanis 28

PAGE 227

201 The participants in this interaction were Sebastiano, an 18 year old male student from Colombia and Shin, a female student from Korea, 38 years of age. Both, Sebastiano and Shin contributed their ideas but did not show any signs of being engaged with each e was only one request from Shin for clarification (line24) which was ignored by Sebastiano. They approached the pre writing task completion separately. There was no evidence of attempt for creating a joined problem space: both of them stated explicitly th at they post their own ideas (lines 4, 10, 11, 26 27). In her last posting, Shin her paragraph using her ideas and rephrased the ideas she posted earlier. However, she he used only his own ideas. There was one hint of disagreement during the online discussion expressed by Sebastiano (line 11) which was ignored by Shin, and it did not trigger a discussion. The expert novice dyad During the treatment, only one dyad demonstrated a communication pattern of a high mutuality but low equality. The members of this dyad were Temi, an 18 year old female student from Japan, and Tiana, a 25 year old female student from Colombia. The students preformed their pre writing discussion in an SCMC environment. While in their discussion there were several collaborative exchanges, the main trend was for Tiana to lead the discussion and guide her part ner through the task completion. This guidance was readily accepted and actually requested by Temi during the discussion. Following is the CMC log of the pre writing discussion of these two participants.

PAGE 228

202 CMC Log 4.4 SCMC Expert Novice Discussion Tiana: LE TS START 1 Temi: ok.what do you think about 1)?? 2 Tiana: I think that US is a big country that has many people of different 3 nationalities 4 Temi: i think that where people should live. 5 Temi: what ideas we should include? 6 i don't know we have to think about paragraph include topic sentence? 7 Tiana: what do you mean? 8 Tiana: NOW 9 Temi: what should we discuss? 10 Tiana: OK, The topic is the people of the US, so we have pictures that represent 11 that idea 12 Tiana: so, i think that we can talk a little bit 13 Temi: ok a little bit?! cool!!! 14 Tiana: of what do you think about pictures, what they represent, 15 Tiana: jajja 16 Temi: so, i think that these pictures meanning is about immigration. i mean many 17 people immigrated here and people do somethings. 18 Tiana: ok, im agree. I think that the pictures show that the two more important 19 cities for inmigrants are NY and L.A 20 Tiana: What about the houses? 21

PAGE 229

203 Tiana: like some live in houses and another in buildings jajajja? 22 Temi: i wanna ignore about houses because i don't know!!! ha ha ha no, just 23 kidding. OK let me think please wait a few minute. you can sleep 24 Temi: maybe this picture meaning is before and after... that's all. what do you 25 think? 26 Temi: top of the left's picture is immigration, and middle of the right's picture is 27 i mmigration too. and two more pictures is NY and LA right? 28 Tiana: maybe is like some people in US live in big cities and another in small cities 29 Temi: ok i agree. so. we should include immigration important cities, big cities and 30 another small cities. 31 Ti ana: NO, I THINK that the right picture in the second line represent the people 32 where many american came from 33 Tiana: the three represent, the cities where inmigrants entered to US: L.A, NY, 34 GULF OF MEXICO 35 Temi: okOKokOK!!! your opinion is better. i agree !!! 36 Tiana: THE houses I THINK LIKE, DIFERENT KIND OF CITIES WHERE THE 37 PEOPLE LIVE 38 Tiana: BUIDINGS OR HOUSES 39 Temi: buiding of houses is what? 40 Tiana: LIKE SOME AMERICAN PEOPLE LIFE IN BIG CITIES IN LARGE 41 BUILDING 42 Tiana: AND ANOTHERS LIVE IN SMALL CITIES IN BIG HOUSES, AND 43 QUIET PLACES 44

PAGE 230

204 Tiana: (do you understand what i mean?) 45 Tiana: NOW WE CAN START THE BRAINSTORM ABOUT PEOPLE IN US 46 Temi: i got it! 47 Tiana: HOW THEY ARE, WHERE THEY LIVE, HOW THEY ARE 48 Temi: i think that they have a dream. 49 they wanna co me here. they get visa and they can live here.. 50 Tiana, who took the expert role, was very active throughout the whole discussion and, in certain moments, she even dominated the discussion (lines 32 44). However, her consideration of the partner was evident ; thus, her interaction was high in terms of mutuality. She initiated the discussion (line 1), encouraged her partner to post and provided her opinion (lines 8, 14, 20 21), offered guidance for task completion (lines 14, 44), and made sure that the ideas she shared were clear to her partner (line 43). On the other hand, Temi assumed a different role in the discussion. Although she was not passive in the process of task completion, her contribution was rather limited. She quickly abandoned her ideas and ag reed with her partner rather than expanding and 6, 9, 24, 26). Further, in the interview, Temi indicated that she relied on her partner to evaluate her ideas and she based If my Thus, although the mutuality in this pair work was clearly visible, the equality was low, with one of the

PAGE 231

205 participants (Tiana) taking the role of the expert and the other (Temi) the role of the no vice. The dominant/passive dyads During the pre writing discussion, two dyads demonstrated low mutuality and low equality patterns. One of these two dyads completed the pre writing task in an asynchronous and the other in a synchronous CMC environment. Thus, this pattern of interaction comprised 5.88% of the ACMC discussions and 7.69% of the SCMC. While these dyads were situated within the same level of mutuality and equality, their specific interaction patterns were somehow different. In the ACMC dyad, only one of the members participated in the task by posting his opinion, whereas both of the SCMC dyad members posted on the chat forum during the pre writing session. The participants in the ACMC Dominant/Passive dyad were two male students from Korea age 21 and 25. One of them (Kyou, 25) posted a lengthy description of the pictures assigned for pre writing discussion. gue in which he addressed his his posting and immediately provided answers to them (lines 4 7, 9 11). The frequent use of singular and plural first person pronouns (l ines 2, 4, 5, 9, 13, 23, 24, 33) along with the 15), suggested a rather authoritarian tone. The other participant did not post any discussions. During the post treatment writing sessio n, the researcher noted that he opened the ICA2 section of his dyad and read the discussion before he started writing. In the CMC interview, he remained

PAGE 232

206 passive as well and did not answer the questions related to the CMC pre writing interaction and did not explain the reason for not completing the pre writing task. Log 4.5 presents this ACMC Dominant/Passive discussion. CMC Log 4.5 ACMC Dominant/Passive Interaction FROM: Kyou (02/19/06 11:07 PM GMT 06:00) 1 In these 6 pictures, I will say: 2 First of the pi ctures 3 We can see what America really is. Everybody talks about what American is, but 4 do we really know the America? This picture shows us the answer. This small map 5 6 Alaska, or small like Hawaiian Islands. 7 Second of the pictures 8 9 to know the America is a big country including many races or cultures, so you can 10 see a lot of different styles in Amer 11 Third of the pictures 12 I see the big city beside the sea, but I am not sure what the city is. I have been in 13 San Francisco and this city is like it, but some people said that is New York. But 14 his picture just shows us a history. Everybody knows 15 America was colonized before, so the first areas colonized was coasts. And this 16 picture tells us these areas have been these big cities. 17 Fourth of the pictures 18

PAGE 233

207 This is also a big city like third picture, 19 sea. The meaning is the inland areas grew to big cities after the coasts, like 20 Chicago, or Denver. 21 Fifth of the pictures 22 If you just see this picture, I guess you will say this means American from a lot of 23 countr ies. But I saw one of them is Canada, so I think there is the different 24 meaning. I think it wants to tell us America welcomes any country to come here, 25 and people like to come to America, too. So it became to present. 26 Sixth of the pictures 27 New York, Los An geles, Chicago, Houston, they are the biggest cities in America, 28 but they are throughout America. Form east to west, north to south, they are in 29 30 adaptable country for people. 31 Topi c sentence: 32 I will tell you what Americans are from my point. 33 The participants of the SCMC pair with Dominant/Passive pattern of interaction were Faraz, a 19 years old student from Saudi Arabia, and Kamil, a 24 years old student from Kuwait. As it is evide nt in the SCMC interaction log (CMC Log 4.6), Faraz progressed through the task without asking or offering any assistance to Kamil. His postings followed one after another (lines 5, 7 11) without any indication that he distracting remarks (lines 6, 13). The interaction pattern demonstrated by Faraz, based on

PAGE 234

208 passive. He contributed very little to the discussion; four out of his five postings actually were either distracting and not related to the task (lines 6 and 13) or presented a general question (lines 2 and 12). Only one posting (lines 3 4) contained text that was remotely related to the assig ned task. He stated that he did not agree with anything Faraz posted (line 13) but did not clarify his opinion and did not seek assistance from his partner. Log 6 presents their interaction. CMC Log 4.6 SCMC Dominant/Passive Interaction Faraz: hi 1 Kamil: y whats up 2 Kamil: well. i think the first picuter tell us there are many people wanna moved 3 from them countrey or travel to somewhere 4 Faraz: in the firs picture it's seem that the populatin of usa incease 5 Kamil: yeah maybe who`s know :) 6 Faraz: the secound picture it's show the amazing desing for the buldings and how 7 it's organize 8 Faraz: the 4th picture it shows that in USA there are diffirent nationalities which 9 cause diffirents culture. 10 Faraz: the final picture shows the map of USA and the impotant city t here. 11 Kamil: hey whats up 12 Kamil: first of all i`m disagree :P foe every thi nk 13

PAGE 235

209 Emerging Patterns When analyzing the CMC interactions for the current study in the context of patterns. These expectations were based on the fact that the model was created with a consideration of face to face interactions, whereas the interactions analyzed in the current study were preformed in a CMC environment. While most of the dyadic interact ion quadrants, two interaction trends which would not fit the model emerged from the data. The disengagement pattern The researcher labeled the first pattern as di sengagement. This pattern was evident in cases when one of the participants, after contributing to the discussion demonstrating the patterns specified by Storch (2002), abandoned it either temporary or permanently before it was completed. Thus, this patter n was identified as secondary to the main dyadic patterns indicated by Stroch (2002). The researcher differentiated this pattern from the passive or the novice ones because it was observed with participants who contributed substantially to the discussion and at the same time the disengagement strongly influenced discussion dynamic. Examples Anna posted a question to her partner, but he never answered it abandoning th e interaction (CMC Log 4.7). Interestingly, this pattern was observed in 12 (71%) of the ACMC dyads, whereas only one dyad (7.7%) of the SCMC group demonstrated it (CMC Log 4.6). However, it should be noted that L.E.C.S. does not provide time tags to the chat postings; thus, the researcher assumes that there is a possibility that during the SCMC

PAGE 236

210 interaction, there were some undetected instances of temporary disengagement from the task. Dyads with mutuality and equality mismatches both members of the dyads were assumed to occupy the same quadrant of the Mutuality/Equality continuum; this assumption was further confirmed by the dyadic face to f ace interaction data analysis performed by Storch in the model construction. When applying the model to CMC interaction data, the researcher noted that in some of the cases, the two members did not occupy the same quadrant of the Mutuality/Equality contin uum. In other words, they demonstrated patterns of interaction that did not match each other on the level of Mutuality. A total of seven dyads demonstrated a mutuality mismatch. Six dyads comprised a collaborative (high mutuality) partner with a dominant ( low mutuality) partner. In one dyad, a participant who demonstrated an expert (high mutuality) pattern of interaction worked with a partner who utilized a dominant (low mutuality) pattern. The ACMC group had the highest number of mutuality mismatch dyads five (29.4%), whereas only two SCMC dyads (15.4%) demonstrated such mismatch. Collaborative/dominant dyads With the ACMC pairs the trend was that one of the participants was approaching the task collaboratively, whereas the other assumed a dominant ro le entirely or almost entirely ignoring the partner. Log 4.7 illustrates such an interaction. The two students participating in the exchange were Isabella, a 28 years old female student from Italy, and Sam, a 46 year old student from Haiti. In this interac tion, Isabella posted one long message which mainly stated her opinion about the topic. There were two attempts on her side to set a collaborative discussion. In line 26 she addressed the

PAGE 237

211 partner with a question about his opinion and in line 27 she used th implying that her partner could provide his opinion. However, the overall tone of her posting was dominant: except for the above mentioned instances, there were no other requests addressed to the partner or any other signs of expectati ons and recognition related around her ideas proposing a different suggestion for paragraph organization (lines 32 46). 34) suggestions were expected. Finally, he explic itly invited Isabella to make her further contribution to the task discussion (line 47). However, Isabella never revisited the discussion (evidence of disengagement). In all five of the ACMC dyads with a mutuality mismatch, the secondary interaction patter n of disengagement was observed with one of the partners. In four out of the five ACMC dyads, the participant who abandoned the discussion was the one who demonstrated a dominant pattern of interaction. CMC Log 4.7 ACMC Collaborative/Dominant Interaction FROM: Isabella (10/17/05 9:36 AM GMT 06:00) 1 Hello Sam sorry for my delay. 2 These are my ideas: 3 1 picture: describe the density in the 50th states if America: people in the state of 4 America are uniform distribuited and so in Haway island. In Alaska the majority 5 density of the people is in the south part of island that is the farest from Pole. 6

PAGE 238

212 2 picture: a lot of the people live in the big cities, in high and essential skyscrapers 7 with a lot of levels. 8 3 picture: other people lives in comfortable and wit h few levels houses in the 9 suburbs far from the frenzied live of the big cities. 10 4 picture: in the US the people have a lot of origins. They came from England, 11 Ireland, Scotland, Germany, Spain, Mexico, Ireland, France, China, Sweden, 12 Vietnam, Canada and Korea. 13 5 picture: a lot of cities are built near the sea and often long bridges link 2 different 14 part of these cities. 15 6 picture: the most densely populated cities are New York on the Atlantic sea and 16 Los Angeles and San Diego on the Pacific coast. 17 ORGANIZ ATION: 18 These 6th photo can be correlated so: 19 First of all I would put the idea of the first picture: people in the state of..... 20 second I would put the idea inthe 6th picture: the most densely.... 21 third I'd put the idea of 4th picture: in the US the peop le.... 22 fourth I'd put the idea of 2nd picture: a lot of people live.... 23 fifth I'd put the idea of 3th picture: other people live in comfortable... 24 sixth I'd put the idea of 5th picture: a lot of cities are built.... 25 What do you think???? 26 The topic sentence could be: The people of the us can be described by the 27 geographical distribution and origins. 28 Concluding idea: Mixture of race in spread territory ( in order to enlarge!!!!). 29

PAGE 239

213 FROM: Sam (10/19/05 11:07 AM GMT 06:00) 30 Hi, Isabella. You have done a tre mendous work. However, your picture's ideas are 31 almost the same except some little change. so I don't need to rewrite them again. 32 the only difference is the organization of the ideas. Here how I would organize 33 them : 34 1. I would put the idea of the 4th pict ure telling about the race and origin of U.S 35 2. I will put the idea of the 1st picture : about density and geographical distribution. 36 3. ''''''''' ' ' ''''''''''''''''''''''''''''''6th picture : the most populous cities 37 4. " " " " 5th picture : telling about the situation of the construction and so 38 on. 39 5. " : " " " " " "2th picture telling about High level people. 40 6. " " " " ?": " "" 3th picture, telling about the people living in the surburb 41 According to me ,the topic sentence should be ""The people of U.S. con be 42 distributed by diferrent ways. 43 topic sentence wiil be : people in U.S. 44 controlling ideas : different was to describe it. 45 Then we will have : ST1 '1st way is by race and Origin 46 ST2: description by density and g eographical distribution 47 ST3: Living situation of the people. 48 so you tell me about what do you think. 49

PAGE 240

214 I n the synchronous discussion, only two dyads demonstrated a mutuality mismatch. In one of the dyads, the participants demonstrated a dominant and an expe rt patterns and in the other a collaborative and a dominant patterns. These discussions are presented in CMC logs 4.8 and 4.9. The interaction of the SCMC dyad that demonstrated collaborative/dominant mutuality mismatch was different from the ACMC ones. While for each of the ACMC dyads with a mutuality mismatch, the dominant participants simply ignored their partners disengaging from the task as soon as they perceived their contribution to the task to be completed, the dominant behavior of the SCMC partic ipant was different. Isa, a 22 year old male student from Morocco, was the SCMC dyad member who demonstrated a dominant behavior while working on the treatment task with his partner, Mia, a 40 year old female student from Colombia. Mia was willing to work together with Isa, providing her views on the task (for example lines 11, 16, 26 27, 31), striving to keep the interaction orderly and cohesive (lines 28, 40 41, 57 58), and expressing interest in her to have an authoritative demeanor from the beginning of the discussion. He chose the pseudonym Isa rather than his real name and refused to introduce himself in the beginning of the discussion (line 4); further, he typed using all capital letters which, a ccording to the rules of various chat communities is a sign of dissatisfaction, anger or pointing out that the information must be noticed by communicate with family and friends, thus it could be assumed that he would be informed about the basics of chat etiquette. However, even if it is assumed that he was not familiar with this particular detail, the tone of many of his postings tended to be authoritarian (for

PAGE 241

215 examp le line 4, 5, 37). He attempted to command his partner rather than to constructively contribute to the task completion frequently giving directions (lines 7, 9, 12, 14, 37). Mia sensed this and at some point she stated that Isa was rude (line 33). However, he ignored this remark. In her interview, Mia shared that she was highly dissatisfied by the way the online discussion progressed: CMC Log 4.8 SCMC Collaborative/Dominant Interaction Mia: Hi everybody 1 ISA: HI MIA 2 Mia: Who is Isa, I want to know 3 ISA: U DON T NEED 4 ISA: LET START WORKING MIA 5 Mia: Ok. U.S. has lots of people who live in big cities 6 ISA: WE HAVE TO TALK ABOUT THE PEOPLE IN US 7 Mia: C'mon 8 ISA: HAVE U READ THE PROMPT ON THE YELLOW SHEET? 9 Mia: yes 10 Mia: Some of them live in cities and some of them in country 11 ISA: OK LET'S GO 12 Mia: zzzz 13 ISA: FIRST OF ALL WE HAVE TO TALK ABOUT THE DIFFERENT 14 STATES IN US 15

PAGE 242

216 Mia: many people are from other countries 16 ISA : THERE ARE 51 STATES IN USA 17 Mia: 50 18 ISA: OK 19 ISA: AND PEOPLE IN US ARE DIVIDED BETWEEN THESE 50 STATES 20 Mia: what else? US has big cities like NY, Chicago, Los Angeles, San Diego 21 ISA: THEN WE WILL TALK ABOUT THE BUSINESS BUILDINGS AND THE 22 SKYSCRAPERS IN US BECAUSE MOST OF PEOPLE WORK THERE 23 ISA: YES WE CAN ALSO TALK ABOUT THE BIG CITIES IN US SUCH AS 24 NYC LOS ANGELES 25 Mia: People in the US are very polite, like to make much money and to spend too 26 mucha 27 Mia: Yes, I wrote about that at the beggining 28 I SA: OK 29 ISA: DO U UNDERSTAND THE THIRD PICTURE? 30 Mia: Now, they make big engineering buding like bridges 31 ISA: I THINK IT SHOWS THE AMERICAN'S HOUSE 32 Mia: Please, don't try to say I'm not intelligent. You're rude. 33 ISA: YES I AGREE 34 ISA: WE CAN SAY THAT 35 I SA: GOOD JOB MIA 36 ISA: CONTINUE... 37 ISA: WHAT ABOUT THE FOURTH PICTURE? 38

PAGE 243

217 ISA: HELLO...??? 39 Mia: Ok. we can begin with organize the paragraph. Yes. Here live people from 40 around the world 41 ISA: I THINK THAT IN THE FOURTH PICTURE SHOWS THAT THERE ARE 42 ETHNIC D IVERSITY IN USA 43 Mia: Are ypu agree? Do you want to add something else? 44 ISA: OK LET'S ORGANIZE OUR PARAGRAPH 45 ISA: WE START WITH THE MAIN IDEA OK? 46 Mia: The topic sentence could be: The ethnic diversty of people of the United 47 States What do you think? 48 ISA: YES I AM OK BUT WE HAVE TO WRITE A SENTENCE 49 ISA: IT COULD BE : THE PEOPLE OF THE USA FORM AN ETHNIC 50 DIVERSITY FOR SEVERAL REASONS 51 ISA: REPRESENT INSTEAD FORM 52 Mia: ok 53 ISA: WHAT DO U THINK? 54 Mia: the supproting ideas we can describe eah picture 55 ISA : I THINK WE WILL USE THE PICTURES FOR DETAILS OR EXAMPLES 56 Mia: Yes. I said that. If you're talking about controlling idea "several reasons" 57 could be. 58 ISA: THE SUPPORTING IDEA COULD BE: USA IS THE FIRST POWER IN 59 THE WORLD THANKS TO THE PEOPLE WHO LIVES THERE 60 ISA: THEN WE CAN TALK ABOUT THE PICTURES 61

PAGE 244

218 Mia: I think we can write about how are Americans and their values 62 ISA: ENGINEERING BUILDING SKYSCRAPERS ... 63 Mia: Not only write about fisical sources, but also write about thougth, believes, 64 ISA: YES WE CAN THESE KIND OF EXAMPLES 65 ISA: DO U THINK WE ARE READY TO START WRITING OUR 66 PARAGRAPH 67 Mia: yes. 68 ISA: SO WE CAN PRINT 69 Mia: Isa, but we have to organize the entire paragraph. What about the concluding 70 sentence? 71 ISA: IT COULD BE : FINALLY THE USA IS ST ILL THE F IRST POWER IN 72 THE WORLD 73 The dominant/expert dyad The participants in the dyad that demonstrated dominant/expert patterns were John, a 26 year old student from Korea, and Mariana, a 34 year old student from Colombia. In their interaction, John was the more active participant. He had a total of 14 postings, and, in most of these postings, there was a clear evidence of task consideration. His interaction had several of the indicators of the expert pattern: he initiated the task completion (line 2), co ntributed his views about the pictures he had to discuss (lines 9 14, 20 23, 28 31), encouraged Mariana to stick to the task (lines 6, 7, 8, 38, 40), and directed her to the handout presenting the images related to the task (lines 18, 19). He did not ignor

PAGE 245

219 by her (lines 25 26, 34), although they were only marginally related to the task; however, after that he immediately returned back to the task topic. Mariana, on the other hand, used the discussion space one sidedly as a forum to share her own personal experience (lines 4 5, 15 17) or general ideas about the topic without specifically referring to the images that were part of the task (lines 24, 27, 32 33, 35, 39, 41) or to the idea to guide her through the task and to show her the information sources she needed to use contribution; her focus was on her own ideas and experiences which prevented her from responses. For example, her post about polite and rude behavior of Americans (lines 4 5) was not directly r elated to the task; therefore, her partner first asked for clarification of her information sources (line 6), but this request was ignored. Further, John encouraged her to post more (lines 7, 8); he continued with his postings related to the task (lines 9 her posting directly continued her thought from lines 4 5. Therefore, the little attention y to relate to this contribution, suggested a dominant pattern of interaction; whereas John patterns correspond with the high end of the mutuality continuum.

PAGE 246

220 CMC Log 4.9 SCMC Dominant/Expert Interaction JOHN: sorry i'm late 1 JOHN: let's start 2 Mariana: JOHN > sure 3 Mariana: I have heard that the people in the U.S are very polite but some peop le are 4 rude. 5 JOHN: where R U? 6 JOHN: hey 7 JOHN: hi 8 JOHN: i think that we must include "melting pot" 9 JOHN: a lot of people come to USA 10 JOHN: america is surrounded ocean 11 JOHN: that mean that other country people enter america more easy.. .. 12 JOHN: easier than other country... 13 JOHN: so usa is open place...so we must include open things of american... 14 Mariana: yes,because I work in a restaurant as a hostes and some people ask me if 15 they can have the witress they like and, I tell them I can not because it is in turns. 16 then they get very apset.some of them are not patience. 17 JOHN: look at the picture.... 18 JOHN: yellow paper 19 JOHN: usa houses have no fence 20 JOHN: 3rd picture 21

PAGE 247

221 JOHN: also there are many high building 22 JOHN: here what are we find ing this picture 23 Mariana: I don't think the same because, the american embasy is very strict with us. 24 JOHN: but other embasys also closing to us 25 JOHN: because we must protect their enemy.... 26 Mariana: people can come easy to usa but, by occean. 27 JOHN: big building...we can use this by idea... 28 JOHN: they like to show something ...like confidence 29 JOHN: to other country... 30 JOHN: so they built big building.... 31 Mariana: I think people come to the usa is because they have the oportunity to 32 come here. 33 JOH N: right...i agree that 34 Mariana: people from usa like to be on time to appointments. 35 JOHN: o.k 36 JOHN: any opinion? 37 JOHN: that isn't involved six picture... 38 Mariana: they like to respect the lines in the ban and others places. 39 JOHN: we have to discuss involved six picture 40 Mariana: they are very organize. 41 Class Dynamics and Patterns of Interaction As it was stated in the Class Dynamics section of this chapter, two of the Level III Writing class sections participating in the study during the fall semeste r of 2005 had very

PAGE 248

222 different dynamics in their face to face classes and these dynamics were evident during the computer laboratory sessions as well. The researcher compared the patterns of interaction of these dyads. A total of seven dyads was examined: fo ur dyads were composed of participants from the section that demonstrated closer relationship and willingness to collaborate with each other, and three dyads from the section that had a more alienated relationship. The comparison of dyadic patterns of int eraction of the two fall 2005 writing class sections revealed different distribution of the patterns. Three of the dyads from the more collaborative section of the writing class demonstrated patterns of high mutuality: two dyads also had high equality inte raction with both members being collaborative and one dyad had low equality with one member being an expert and the other novice. One dyad had a mutuality level mismatch with one member being dominant and the other collaborative. None of the three dyads composed of students from the more alienated section of the writing class demonstrated a high mutuality pattern: two of the dyads had a mutuality level mismatch with one member being dominant and the other collaborative and the other dyad demonstrated a low mutuality level with both participants being dominant. Although it is difficult to attribute the difference between the patterns of the who had full data set were ACMC group dyads, while two of the dyads from the writing discussion, it could be suggested that there is a possibility that the class dynamics had some influence on the dyadic pattern of CMC interactio n.

PAGE 249

223 Computer Mediated Interview Outcomes The researcher conducted interviews with all students from Level III Writing class in Fall 2005 and 2006. Only the interviews of the students who had a complete treatment data set (pre treatment and post treatment w riting samples and CMC pre writing treatment) were considered for analysis. Five of the participants who completed all three assignments related to the treatment were reluctant to answer the interview questions despite multiple reminders. Thus, 55 intervie ws were analyzed. The structured interviews were conducted asynchronously using ICA2 as a medium. During the first semester, the questions were posted on an ICA2 space dedicated for each participant; in the second semester collection, due to the high numb er of participants, the researcher posted the question in the ICA2 space dedicated for each Level III Writing section and received the answers via the send a personal message function of ICA2. The follow up questions and requests for clarifications were se nt as personal messages. Because of the asynchronous nature of the interview medium, the participants had the liberty to avoid answering some of the questions; thus, 21 out of the 55 interviews were partially incomplete with one or more questions left unan swered. However, the researcher took these interviews into account because they provided the study. In order to provide more comprehensible presentation of the theme frequencies, the number of students/answers for each theme are reported along with what percentage of total answers this number presents. The interview outcomes revealed several important trends related to the factors that influence the CMC pre writing in teraction process. These trends concerned previous

PAGE 250

224 computer related experience, attitudes towards writing, computer mediated communication, and task perception. The interview questions are presented in table 4.3.

PAGE 251

225 Table 4.3 Interview Questions Outline Qu estion Type Question 1 4 Background (a) age; (b) country of origin; (c) native language; (d) educational background. 5 Opinion For you as a language learner, how important is it to develop good writing skills in English? Please explain your answer. 6 Current Experience What is the most difficult thing for you when you write in English? 7 Current Experience What is the most enjoyable thing for you when writing in English? 8 Current Experience How comfortable are you when working with computers? 9 Past Experience How were computers used at your previous school(s)? 10 Past Experience Did you use computers to communicate with other people (chat, email, list serves) before this class?

PAGE 252

226 Table 4.3 (Continued) Interview Questions Outline Questio n Type Question 11 Opinion You have chatted with L.E.C.S. (the program you used to interview a partner) and also posted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? 12 Feelings How did you f eel completing the discussion of the pictures? How was it different from when you discuss things in class? 13 Current Experience Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please explai n your answer: What was very helpful? What was distractive or not helpful?

PAGE 253

227 Table 4.3 (Continued) Interview Questions Outline Question Type Question 14 Opinion Do you think CMC activities are good for learning how to write better? Please explain yo ur opinion. 15 Current Experience Did you include in your writing some of the ideas that your partner shared with you? 16 Current Experience When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictu res? 17 Current Experience Did you find that your typing skills hindered your discussion on line? 18 Current Experience Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your messages?

PAGE 254

228 Atti tudes Towards Writing Three of the questions (5 difficult and enjoyable aspects of writing. Question 5 ( For yo u as a language learner, how important is it to develop good writing skills in English? Please explain your answer .) was answered by 44 students. Question 6 ( What is the most difficult thing for you when you write in English?) was answered by 45 students Question 7 ( What is the most enjoyable thing for you when writing in English?) was answered by 43 students. Six distinct themes related to the importance of developing good writing s kills were identified in the interviews of the students: (1) writing for further studies, (2) writing for communication; (3) writing as being related to other language skills, (4) writing related to future and current career, (5) writing as part of everyda y life, and (6) writing to share personal experiences. The highest number of students (39, 88.64%) perceived that good writing skills are important because of the necessity to communicate effectively. There were 16 students (36.37%) who related their acade mic goals to the formation of good writing skills. While eight students (18.18%) indicated that writing is important for their future or current career. For example in the answer of one of the In my opinion the importance to be agood writer that its give you alot of benefits in alot majors in your life for example if you wana apply for a jop u must pic the properate words in your C.V ,also if u wana apply collage or university they required alot off homeworks and projects to do so u will need alot of academic vocobulary and the most simple thing if you wana say your theory in

PAGE 255

229 any subject in life or to write a letter for formal place spelling and punctuation is preserved unchang ed.) Writing was closely related to the development of other language skills such as reading and speaking by 12 students i think its very important because i am learning english as a langua ge so i can not seperate speaking that good writing skills are important for everyday life. In the answers of these two students, this theme was combined with at least o ne another theme such as importance of writing for academic and career purposes. For example, in her interview answers, one participant combined the theme of writing as an important communication medium and the theme of the importance of writing in everyda transmite aned communicate have better communication skills and improve my perfomance not only in my studies but also in the real life When sharing their diff iculties with writing in English, most of the students pointed out features related to the surface aspect of writing such as grammar, including translation from L1 (23, 51.11%), spelling (26, 57.78%), and vocabulary (27, 60 %). For example, one of the stud ents shared the following about his difficulties related to writing when I write in English, I am the most difficult translation for me from korea to English because English has oppositely construction sentence the m ost difficult thing is writing passive voice sentences posted a short answer to the question about the difficulties he experienced when writing spiling and gramer

PAGE 256

230 Few students expressed awareness of difficulties related t o text organization and clear representation of ideas (6, 13.33%). One of the students shared the following in her The most difficult thing is write correctly. Using the appropiate words, transitions and with a good organization g she addressed both grammar difficulties and difficulties related to text organization. The enjoyment from writing triggered various response from the students with the highest variety of themes. However, it is important to note that the highest number of students (18, 41.86%) avoided answering this question. Which could be contributed to the difficulty to think of positive aspects related to writing or having difficulties formulating them. Learning how to write better in English and control the text w as a theme shared by 10 (23.25%). One of the students explained his positive attitudes towards writing in the following way: [are enjoyable for me when writing in English] I like Another stressed that she enjoyed when she can c To see it ready and to see my ideas on paper Seven students perceived as most enjoyable writing about personal experiences To describe something about my experience The most thing i like in writting that i can put every thougt in the paper If I can delivery my idea to the people, and I can effect to the people so I make change

PAGE 257

231 I can learn more and more vocabulary and let the other know what is my opinion Receiving feedback (2, 4.65% ) was one of the themes. The students who shared can have a beedfack of my writting, and that is a good sign to continue writing.. My bigest joy in writing eng lish is that I realize that I'm in the way to reach my goal by practicing, for I know that as soon as I'm writing things and my teacher is correcting me Mastering vocabulary was yet another theme related to positive aspects of writing in English. One of when I write in English, I am the most enjoyable thing is leaned the new word (4.65%) of the interview answers. There were several themes related to positive aspects of writing which were identified in the interview answers of individual students: chatting (1, 2.33%), being creative (1, 2.33%), writing a coherent text (1, 2.33%), and the challenge of writing in English (1, 2.33%). There were also three students (6.98%) that expl icitly indicated that nothing is enjoyable for them about writing, and one (2.33%) that stated that the only enjoyable thing having to do with writing was to write his name using Roman script. Computer and CMC Related Experiences It appeared that all of the students who answered the question about their computer experiences (42), except for three (7.14%), had somewhat established computer skills prior to the semester in which the data were collected. This conclusion was made based on their answer to Quest ion 9 ( How comfortable are you when working with computers? ).

PAGE 258

232 The majority of the students (39, 92.86%) answered that they were either comfortable or very comfortable when working on a computer, that they love and prefer working on a computer. One of this I will [be] uncomfortable if no computers can be used anymore computers were used for various academic tasks such as presentations, homework completion, software design (Q uestion 9), 10 (4.6%) did not use computers in their education prior to enrolling in the ELI. Seven (15.28%) of the students who had not used computers for academic purposes identified themselves as being comfortable or very comfortable; they used verbs su like love computers. Almost all students (39, 95.12%) who answered Question 10 ( Did you use computers to communicate with other people (chat, email, list serves) before this class?) shared that they had use d computer mediated communication to stay in touch with friends and family. In addition, three students (7.32%) also stated that they had used CMC for professional communication, whereas none of the participants had used CMC as a medium for the completion of educational tasks. There were three (7.32%) students who indicated low level of comfort when working on computers; two (4.88%) of these students also indicated that they had not used CMC prior to the semester in which data were collected. The third stud ent with low computer skills was a continuing student; thus, she acquired some CMC skills at the IEP in the previous semester as part of her studies. CMC Interaction and Task Perception When asked about their perception of the pre writing CMC task, 50 s tudents answered. The majority of them (42, 84%) perceived the task to be helpful for their

PAGE 259

233 On my opinion, it is helpful. When I discussioning activity, I can get more information and differe nt opinions for other people the on line discussion activity was very helpful to me in writing my paragraph, because most the ideas from the pictures have already been discussed between us together so it made it more easy for writing the paragraph Four of the students (8%) recognized explicitly the positive role of their partners; the fact that the discussion promoted the generation of more ideas and better understanding of the topic was mentioned by seven students (14%). Two (4%) students stated that the CMC format was something that made the task better. Two (4%) students mentioned that the CMC format of the discussion allowed them to revisit the discussion and see errors. One student (2%) admitted that her perception was that sh e and her partner generated fewer ideas but the discussion was useful for the completion of the writing task. Three of the students (6%) indicated that they found the task useful but time consuming. Two students (4%) expressed positive attitude towards the CMC discussion but stated that they would prefer to talk about the task. There were eight students (16%) who stated that they did not like the pre writing CMC discussion. Four of them were from the ACMC group and four from the SCMC group. In her intervie Talking about pictures was a little bit difficult. The problem does not concern the way of talking about it (on ESLS with a computer) but it concern the pictures themselves that were not easy to understand Anyway, I prefer to talk about pictorial things in front of a person because I can use physic interactions to explain what I thing to my partner

PAGE 260

234 Three of the participants 6% (two from the ACMC group and one from the SCMC group) perceived the topic discussion with a partner to be unnecessary because they would prefer to work on the writing task alone. One of them (1%) added that the was what he would value more. Two students (ACMC, 4%) stated that they would prefer to talk rather than post messages. For I was uncomfortable because I can not use computer very well. In class I can n ot speak if I want. In computer I have to One of the students (SCMC, 2%) did not feel comfortable during the task completion because of her partner who showed disrespect during the discussion. Lastly, one student (ACMC, 2%) stated that the task was borin g, and there was not enough time to discuss it. A total of 49 students answered the question which addressed the inclusion of answered that they used the ideas of their partners. Eleven of those students (22.45%) explicitly stated that they used the interaction log during their writing. Some students (6, 12.24%) reported that they used only their ideas when writing. Five of these students completed their discussion async hronously via ICA2 and four of them demonstrated a dominant dyadic pattern, which constituted 16.67% of the dominant were collaborative participants (6.90% of the collaborative participants). Thus, the group an ACMC mode of communication (five out of six) and there were more dominant (4) than collaborative participants (2) amo ng them.

PAGE 261

235 CMC Interaction Tools and Typing When asked if they used anything additional during the CMC discussion an electronic dictionary, the Internet or a regular dictionary the majority of the students answered negatively (23 62,79%). One studen t from the SCMC group (4.35%) explicitly stated that there was no time for using anything additional to support the discussion, another from the ACMC group (4.35%) stated that, in general, he used a dictionary but did not use it for this task. The rest of the students (91.31%) who stated that they used nothing to support their discussion did not provide any additional comments. From the ACMC group, nine students used dictionaries, either online, electronic or hard copy, whereas in the SCMC group the number of students using dictionaries was seven. When asked if the CMC interaction was slowed down or hindered by their typing skills, the majority of the ACMC participants 14 out of the 23 students who answered this question (60.87%) stated that their inte raction was not hindered. Nine (39.13%) shared that they had difficulties with typing. Six out of the 21 SCMC participants who answered this question (28.57%) indicated that typing was a problem during their chat pre writing session. The Low Computer Skil ls Proficiency Students The three students one male, Sam (46) from Haiti, and two females, Shin (38) from Korea, and Anna (40) from Russia who indicated lower computer skills were older than the majority of the students. The mean age of the participant s was 25.55. These lower computer proficiency students shared that they felt as if they were in the process of I was too scared when I was

PAGE 262

236 working with computers, today in ELI I really happy to do exercise with computers because I hope improve my computer skills with computers was similar with the other two students. All three of them participated in the extracurricular computer workshops offered in the program; in addition Anna shared in a personal conversation that she asked her husband, who was a computer programmer, to show her how to use various computer functions. All three were part of the ACMC treatment group. They had an overall positive percepti on of the CMC pre writing task. Anna and Sam explicitly indicated their positive experience during the task completion. They felt that the CMC pre writing interaction helped them to prepare for their writing, and Anna indicated that the pre writing task h ad also a positive influence on the development of her computer skills. Shin stated that she did not feel comfortable when completing the task because of the computer. Nevertheless, she found that sharing ideas with her partner helped her later to write he r paragraph. These students demonstrated different degree of gain as identified based on the difference of the pulled z scores of the eight measures used for the paragraph analysis of the pre treatment paragraph, completed after a face to face pre writing discussion, and the post treatment paragraph, completed after the CMC pre writing discussion. In her post treatment writing scores, Shin had the lowest gain among all participants. This gain was 6.06133, with group mean gain score 0.144566553. Her partne r, a student who indicated a high comfort level with computers, had a positive gain (1.1454). Anna also had a negative gain (

PAGE 263

237 fifth among all participants in the study, while his partner had the highest gain of all participants. Anna and Sam demonstrated collaborative patterns of dyadic interaction during the ACMC pre writing task completion. They participated in the interaction with ideas and propose their own original ideas. Shin, however, politely recognized her raph based on Qualitative Stage Summary of Results When analyzing the qualitative data, three qualitative questions were addressed: (1) of CMC in the process of establishing their writing skills? (2) What patterns of dyadic interaction do participants manifest during the asynchronous and synchronous CMC interaction process? (3) What are the factors that influence the CMC pre writing intera ction process? How do t hese factors influence the interaction process? Following is the discussion of each research question based on the outcomes of the qualitative data analysis. Qualitative Research Question 1 a CMC in the writing process. Based on the themes identified in the interviews, most of the students found that the pre writing CMC discussion supported the writing of their first paragraph drafts. This support w as qualified in various ways. The students perceived that they generated ideas and understood better the topic of the written assignment due to the

PAGE 264

238 ideas into their paragraphs. The students also shared that the written format of the discussion supported their paragraph writing. This format provided them with the opportunity to review the discussion and see their errors. Some of the students explicitly stated that they revisited the CMC interaction logs while writing their paragraphs. However, there were students who perceived as more appropriate to incorporate only their own ideas into the paragraph. The majority of these students participated in an ACMC discussion and demonstrated dominant patterns of interaction. Four main reasons for the CMC pre writing being unnecessary were pointed out in the interviews. The first two were directly related to the format of the discussion: students experienced difficulties formulat ing ideas in a text based environment and the text based mode of communication was time consuming. These two opinions were prevalent among the students who expressed negative attitude towards the pre writing discussion. The other two reasons were not relat ed directly to the CMC format of the discussion and were also shared by a lower number of students: teacher feedback was identified as more valuable than interactions with peers, and writing was defined as an individual activity that does not require peer collaboration. Patterns of Dyadic Interaction Qualitative Research Question 2 focused on the patterns of dyadic interaction in a CMC environment as identified based on the analysis of the CMC pre writing discussion logs. The patterns of dyadic interaction observed in the CMC discussions in this study corresponded with the model of dyadic interaction proposed by Storch (2002). However, based on the outcomes of this study it could be concluded, that the patterns of dyadic interaction were strongly influenced by mode of interaction. The synchronous discussion

PAGE 265

239 promoted opportunities for more collaboration, whereas the asynchronous discussion, although providing opportunities for collaboration, led to more dominant patterns. Further, there were interactions in w hich students would demonstrate a different level of different level of engagement was more evident in the ACMC group. Another specific of the CMC interaction also observed mostly in the ACMC group was that students would participants would post only once and would not engage in further discussion. This disengagement pattern could be attributed to the less dynamic nature of the ACMC interaction as well as to the less structured environment in which the interaction took place. Factors Influencing the CMC Pre writing Interaction In this section, Qualitative Research Question 3 is addressed. In or der to identify the factors that influenced the CMC pre writing interaction process, several sources of data were considered: prior and post treatment computer laboratory observations, the dyadic patterns of interaction identified based on the CMC logs, an d interview outcomes. Several factors that influenced the CMC interactions were identified: (1) the use of computer as an interaction medium, (2) CMC mode of interaction, (3) partner interaction, and (4) opinion of the nature of writing. Discussion of each one is presented below. The Use of Computer as an Interaction Medium During the computer laboratory classes, almost all students who participated in this study demonstrated sufficient computer skills; in the interviews, they also reported being exposed t o computers and CMC. Most of the students used computers to prepare

PAGE 266

240 presentations, write papers, and work on projects. Further, they were involved in CMC interactions for personal reasons predominantly in their native language; only a few of them had used CMC for work related tasks. None of them reported using CMC for academic tasks requiring problem solving. Thus, whereas their computer skills were somehow developed, the lack of experience related to problem solving via CMC may have affected their interact ion. However, more specific interview questions addressing this issue as well as in depth analysis and comparison of interaction of students with and without CMC problem solving experience would shed more light on this particular problem. In their intervie ws, some students shared that they experienced difficulties related to typing skills as well as difficulties expressing themselves in a text based environment. These difficulties made the CMC interaction more time consuming and challenging. Further, during the computer laboratory observations, the researcher noted that many students would attempt to work on the computer only after taking handwritten notes, which suggests that they were less likely to use the computer in the early stages of the writing proce ss. There were three participants in the study who had limited computer skills. The analysis of their interactions showed that two of them were able to engage in collaborative pre writing discussion via ACMC. The third student, who also performed her inte raction in an ACMC dyad, demonstrated a dominant pattern. The two students who demonstrated collaborative patterns of interaction also expressed positive attitudes towards the CMC pre writing discussion task, whereas the third student, who demonstrated dom inant discussion pattern, stated that she did not like the task and would

PAGE 267

241 prefer talking about the pictures. Therefore, based on these particular cases, it could be suggested that even relatively short exposure (in this case four weeks) to computer based t asks performed in an academic environment and additional support through extracurricular workshops could be useful for students and would allow them to perform and benefit from pre writing and writing tasks using computers. The Mode of CMC Interaction The mode of CMC interaction was the most dominant factor that influenced the pre writing discussion. The ACMC and the SCMC interactions analyzed in this study differed significantly in terms of language quality and quantity. These interactions differed in ter ms of dyadic patterns as well. The language produced by the participants in the two modes of interaction differed in several ways. The ACMC mode triggered postings that used longer strings of words often organized in complete sentences, whereas in the SC MC interactions, the electronic utterances were much shorter and did not always comprise complete sentences. However, the number of unique words in ACMC interactions was lower as compared to the SCMC interactions. Further, the SCMC interactions generated l onger discussions, with higher number of posting, therefore the learners had more opportunities to generate and interact in the target language. When working in a n SCMC mode, students demonstrated a stronger tendency to collaborate with each other, where as the ACMC mode led to more dominant patterns of interaction. In addition, the interaction patterns of the same type were demonstrated in a different way depending on the mode of CMC. The collaborative students in ACMC mode attempted to contribute to the discussion with longer postings demonstrating

PAGE 268

242 presenting their opinion, and directing the task. In the SCMC discussions, these moves were frequently spread out through separate postings. The dominant students who communicated via ACMC would more often abandon the discussion space after posting once, whereas the dominant students from the SCMC group would make several postings but would demonstrate inability to engage in to a discussion with their partner. No Expert or Novice patterns were observed with the ACMC postings. It might be inferred that they were transferred to Dominant and Passive patterns due to the less dynamic nature of the ACMC interaction, thus creating mu tuality mismatch. On the contrary, in the SCMC interactions, the immediate contact with the discussion partner in addition to the mutual time space presented a higher participation demand and partner consideration, which led to fewer instances of mutuality mismatch. Based on these outcomes, it could be suggested that the asynchronous discussion was less likely to trigger a dynamic exchange with active negotiation as compared to the synchronous interaction; as a result, it is possible for students who pa rticipate in this type of interaction to be less susceptible to considering the contribution of their partners. Further, students who were dominant in their ACMC discussions might have been more students who participated in the ACMC discussions wrote longer sentences, frequently organized their postings in an outline or even a paragraph format, which allowed them to approximate the paragraph organization and support.

PAGE 269

243 Partner Interaction Style Th way the CMC discussion task was completed. The inability or the unwillingness of one of the partners to engage in the interaction resulted in interactions that would not devel op completely the topic. These interactions were less coherent and tended to demonstrate unequal levels of partner engagement. Further, in their interview students shared that Opinion on t he Importance of Writing In the interviews, the majority of the students shared that good writing skills were important for the reason that writing was a communication medium. This seems to relate writing CMC discussion task. There was no relation found between patterns of dyadic interaction and attitudes towards writing in the group of students who expressed overall positive opinion about writing. However, there was a relation between the dyadic pattern o f interaction and the less prevalent view shared by students on writing as an individual experience that does not students who expressed this view did not engage in a discuss ion with their partners and demonstrated a dominant pattern of interaction. In the following sections of this chapter, the results of the quantitative stage are presented. They are discussed in the context of the specific quantitative research questions po sed in the study. Further, the additional analysis and findings triggered by the nature of the CMC discussions identified in the qualitative stage are presented and discussed.

PAGE 270

244 Quantitative Stage: Analysis and Results In this section, the results of the te xt analysis of the post treatment paragraphs are presented and discussed. The texts produced by students were analyzed based on eight criteria, namely: (1) syntactic complexity (measured by calculating the mean length of T units), (2) amount of information present in a single focus (measured by mean length of idea units), (3) quantity of overall information present (measured by the number of idea units) (4) lexical information per clause (measured through lexical density analysis), (5) vocabulary complexit y (measured by analyzing the frequency of the unique words used), (6) rhetorical soundness, (7) presentation and development of main ideas, (8) and overall language use (the last three criteria were assessed using a multiple trait rubric). The inferential each of the first five criteria, analysis of covariance was performed with covariates the treatment paragraphs produced be fore the treatment CMC pre writing discussions on a different topic. The last three criteria were analyzed using Mann Whitney U test. As stated in Chapter III, in order to assure independence of the scores, the unit of analysis for each of the above state d tests was a dyad. For each dyad the mean scores for every measured criterion were calculated. These mean scores were then compared accounting for the mode (asynchronous versus synchronous) of the CMC pre writing discussion of the assigned topic and using as a covariate the corresponding scores obtained from their pre treatment writings.

PAGE 271

245 Comparing the ACMC and the SCMC Groups A total of 60 participants that worked in 30 dyads were considered for the study. Although 103 students were enrolled during the two semesters in which the data were collected, complete data sets were obtained for 30 dyads, 17 ACMC dyads and 13 SCMC age, and gender. This comparison was performed in order to assure that the two groups did not differ significantly. Following are the results of this background data analysis. 2004). Group Comparison: Age The part icipants in the current study ranged in age from 17 (1) to 46 (1). The mean age was 25.55 with median ages 25 (8) and 26 (8). The majority of the participants were in their late teens or early twenties (31). The two groups were compared in terms of their age. In order to perform this comparison, a two tailed t test was conducted. The mean age of the ACMC group was 26.44 years ( SD SD = 4.07). The Null Hypothesis was that age means of the two groups did not di ffer significantly. The test failed t t (58) = 1.34, p = .19. The Mean age difference was 2.06 at 95% confidence interval ( 1.05, 5.14). Based on these results, the researcher concluded that the probability of one of the groups to h aving a higher mean age as compared to the other is high. In other words, the two groups did not differ significantly in terms of their mean age.

PAGE 272

246 Group Comparison: Educational Background The two groups were compared also in terms of their educational back ground. According to the application requirements at the IEP, all prospective students should have at least a high school diploma. Based on the interview outcomes, the researcher identified three distinct groups of participants: (1) students who have gradu ated from high school; (2) students who completed their high school education and had at least one semester of college education, and (3) students who had obtained their college degree its equivalent) and/or were working prior to their enrollment at the IEP. The descriptive statistics revealed that the highest number of participants (31) already had their college or higher degree earned prior to enrolling to the IEP. The participants who were in the process of working on their college degree were 17, and the ones who enrolled in the IEP with a high school diploma were 12. Table 4.4 presents the participants in terms of their education and group assignment. Table 4.4 ation level and group assignment Group Assignment ACMC SCMC Total High School Diploma Only 6 6 12 College in Progress 10 7 17 College Degree Obtained 18 13 31 Total 34 26 60

PAGE 273

247 In order to proceed with further statistical comparison of the groups, the r esearcher had to collapse educational level one (high school) and two (college) into one Pre College/College. This decision was made due to the small sample size. The decision to collapse these two levels was based on the fact that the majority of the st udents who already received their college degree were professionals with starting or already established careers; thus, it was viewed that the college level group would be more similar to the high school level group in terms of their educational background Consequently, two levels were considered for further comparison analysis of the ACMC and the SCMC groups: (1) Pre College/College and (2) Post College level. The groups were compared in terms of the odds of a participant of a particular educational level to be assigned to a particular group (ACMC or SCMC). The results of the odds ratio analysis showed no statistically significant difference between the two groups in terms of educational level (OR = 1.6154; 95 % CI = 0.57, 4.54). Further, the groups were with the test, the following assumptions were reviewed, and the researcher concluded that they were met: (1) random sample from population, (2) no single data point equals zero; (3) all observations ind at the wide 95% confidence interval for relative risk, with an obtained two sided p = .43, there was no association of student education level and CMC group assignment. In other words, the tw o groups did not differ significantly in terms of education level. Group Comparison: Gender was a higher number of male participants in the current study: 36 (60%) of the

PAGE 274

248 p articipants were male. The groups were compared in terms of the odds for a participant of a particular gender to be assigned to a particular group (ACMC or SCMC). The results of the odds ratio analysis showed no statistically significant difference between the two groups in terms of gender (OR = 0.6364; 95 % CI = 0.22, 1.8070). The number of participants, as assigned randomly to ACMC and SCMC groups and consequently to dyads, is presented in Table 4.5. Table 4.5 Dyad Assignment by Group ACMC Group Dyad Assi gnment SCMC Group Dyad Assignment Mixed Gender Same Gender Total Mixed Gender Same Gender Total Males Females Males Females 10 6 1 17 8 3 2 13 Next, the groups were compared in terms of the odds of a participant to be assigned to a mixed or the same gender dyad for the ACMC and the SCMC group. The results of the odds ratio analysis showed no statistically significant difference between the two groups in terms of assignment to a mixed or the same gender dyad (OR = 0.8929; 95 % CI = 0.20, 3.91). with the test, the following assumptions were reviewed, and the researcher concluded that they were met: (1) random sample from population, (2) no single data point equals zero Test, at the wide 95% confidence interval for relative risk, with an obtained two sided p =

PAGE 275

249 1.00, there was no association of dyad gender composition and CMC group as signment. In other words, the two groups did not differ significantly in terms of gender composition at a group and a dyad level. Group Comparison: Pre Treatment Writing Score The pre treatment paragraphs written by the students from the ACMC and SCMC grou ps were compared in order to evaluate how similar the two groups were in terms of writing skills. Five of the proposed measures were presented with continuous scores; thus, they were analyzed using five separate exact t tests. These measures were as follo ws: (1) syntactic complexity (measured by calculating the mean length of T units), (2) amount of information present in a single focus (measured by mean length of idea units), (3) quantity of overall information present (measured by the number of idea unit s) (4) lexical information per clause (measured through lexical density analysis), (5) vocabulary complexity (measured by analyzing the frequency of the unique words used). Three of the measures, namely: (6) rhetorical soundness, (7) presentation and deve lopment of main ideas, (8) and overall language use used ordinal scores; they were analyzed applying three separate non parametric two tailed Mann Whitney U tests. The descriptive statistics for each of the pre treatment measures are presented in Table 4.6 Following are the results of the proposed comparison of the two groups.

PAGE 276

250 Table 4.6 Pre treatment scores: Descriptive Statistics Measure n M SD Skewness Kurtosis 1 T units Scores (Syntactic Complexity) ACMC 17 11.75 2.90 1.20 1.19 SCMC 13 1 1.89 1.31 0.19 0.69 2 Mean length of idea units (amount of information present in a single focus) ACMC 17 7.72 0.99 0.41 1.38 SCMC 13 7.65 0.63 0.07 1.26 3 Number of idea units (quantity of overall information present) ACMC 17 17.65 8. 86 2.31 6.87 SCMC 13 18.15 5.60 0.47 1.56 4 Lexical density (lexical information per clause) ACMC 17 4.32 0.80 0.58 0.02 SCMC 13 4.74 0.82 0.22 1.28

PAGE 277

251 Table 4.6 (Continued) Pre treatment Scores: Descriptive Statistics Measure n M SD Skewne ss Kurtosis 5 Vocabulary complexity (frequency of the unique words used) ACMC 17 0.43 0.14 1.15 1.54 SCMC 13 0.36 0.08 0.80 0.56 6 Rhetorical soundness ACMC 17 13.58 23.44 SCMC 13 18.00 23.44 7 Presentation and development of main ideas ACMC 17 12.82 22.72 SCMC 13 19.00 22.72 8 Overall language use ACMC 17 15.88 22.94 SCMC 13 15.00 22.94

PAGE 278

252 The ACMC and SCMC groups were treated as two separate populations. The goal of the exact t test was to confirm or rej ect the Null Hypothesis. Failing to reject the Null Hypothesis would indicate that the means of the two groups do not differ significantly. Before conducting the t test, the researcher carefully examined the following assumptions and concluded that they w ere met: (1) independence of the groups, (2) normal distribution of the sample means, and (3) equal or similar variance of the two groups. The random assignment of the participants to either ACMC or SCMC group and the consideration of the mean pair scores rather than individual scores assured the compliance with the first assumption. The Shapiro Wilk W at the alpha=.05 level allowed to the researcher to conclude with 95% confidence that there was sufficient evidence that the means were normally distributed for both groups for three of the measures: (2) mean length of idea units (ACMC W=.82, SCMC W=.28), (4) lexical density analysis (ACMC W=.49, SCMC W=.45), and (5) vocabulary complexity (ACMC W=.07, SCMC W=.64) However, two of the measures revealed non nor mal distribution. For measure (1), mean length of T units, the scores of the ACMC group were non normally distributed (W=.03), while for the SCMC group the distribution was normal (W=.92). For measure (3), number of idea units, both groups had non normall y distributed scores (ACMC W=.0007, SCMC W=.03). After examining the data, the researcher identified extreme scores for the ACMC group for both measures: mean length T units and number of idea units. These extreme scores were examined, and it was concluded that there was no error in the score calculation. The normality of the data sets was again tested without the extreme observations which showed that their exclusions led to normal distribution: for measure (1), mean

PAGE 279

253 length of T units, ACMC group (W=.10), and measure (3), number of idea units, both groups had normally distributed scores ACMC W=.23. However, there was one more non normal distribution identified: for the SCMC group on measure (3) number of idea units, W=.03. The researcher examined the data, and did not find any outliers or extreme observations. The researcher decided to proceed with the t test with this data set. This decision was based on the following considerations: the detected non normal distribution was only in one of the groups, and the normality was slightly below .05. The overall data set distribution was normal (W=.07). In addition, the t test is considered to be relatively robust to normality violation (Glass & Hopkins, 1995). Thus, the independent two tailed t tests for each me asure with non normal distribution was performed with and without the ACMC extreme observations, and in both cases, the results were not statistically significant. Therefore, the researcher decided to proceed by reporting the results of the t tests includi ng the extreme observations. This decision was made based on the following: (1) the two groups were perceived to be two separate populations for this particular test, (2) the t tests on these measures conducted without the extreme observation did not chang e the way the Null Hypotheses were interpreted, and (3) the t test is perceived to be robust to violation of normality (Glass & Hopkins, 1995). Finally, the probability F values for each of the proposed measures were obtained; which allowed the researcher to decision which variance to use when interpreting the outcomes. The F values are reported in Table 4.7 along with the results of the t tests for each proposed measure.

PAGE 280

254 Th e researcher did not anticipate either of the two groups to have higher scores on any of the proposed measures; thus, five two tailed independent t tests were performed. Each of the t tests failed to reject the Null Hypothesis. Thus, no significant differe nce between the means of the proposed measures for the ACMC and SCMC groups was detected. The outcomes of the five t tests conducted are presented in Table 4.7.

PAGE 281

255 Table 4.7 Pre treatment t test: Statistics and Outcomes Measure DF Null Hypothesis, Pr obability P, Variance 1 T units Scores (Syntactic Complexity) ACMC SCMC 23.5 Failed to reject ( p =.86) .08 Unequal 2 Mean length of idea units (amount of information present in a single focus) ACMC SCMC 16 Failed to reject ( p =.81) .1 1 Equal

PAGE 282

256 Table 4.7 (Continued) Pre treatment t test: Statistics and Outcomes Measure DF Null Hypothesis, Probability P, Variance 3 Number of idea units (quantity of overall information present) ACMC SCMC 28 Failed to reject ( p =.8 5) .11 Equal 4 Lexical density (lexical information per clause) ACMC SCMC 28 Failed to reject ( p =.17) .89 Equal 5 Vocabulary complexity (frequency of the unique words used) ACMC SCMC 26 Failed to reject ( p =.10) .05 Unequal

PAGE 283

257 The group compar ison as based on the three ordinal measures (6) rhetorical soundness, (7) presentation and development of main ideas, and (8) overall language use used ordinal scores and was performed using the non parametric Mann Whitney U test. For each of the Mann Whitney U tests conducted, the Null Hypotheses failed to be rejected at 95% confidence level. Following are the p values obtained in the comparison of the two groups on each of the proposed measures: (6) rhetorical soundness p =.18, (7) presentation and development of main ideas p =.06, and (8) overall language use p =.79. The researcher concluded that there is no significant difference between the two groups when their pre treatment draft paragraphs were compared by the eight measures proposed for the study. Therefore, it was concluded that if any changes are detected in the comparison of the post treatment writing scores of the two groups, they can be attributed to the treatment. The following section presents the post treatment data analysis and resu lts. Post treatment Data Analysis and Results The post treatment data was analyzed using analysis of covariance (ANCOVA) method of statistical analysis for the five continuous measures used in the study while the three ordinal measures were analyzed apply ing three separate Mann Whitney U tests. possibility of not finding statistically significant effects was high, even though if each of the analyses were performed separatel y or fewer ANCOVA tests were conducted, at least one of the effects could be detected (Onwuegbuzie & Levin, 2005). Thus, in order to avoid this limitation, the Three Step Approach for Testing Multiple Univariate

PAGE 284

258 Hypotheses proposed by Onwuegbuzie and Levin (2005) was applied. This led to the analysis of the pulled z scores from all variables in order to compare the ACMC and SCMC groups. The following sections present each of the listed analyses and the obtained results. ANCOVA Analysis of Post treatment Wri tings In ANCOVA, the analysis of variance is combined with regression analysis by adding a concomitant variable (covariate). The goal of utilizing analysis of covariance is to attempt to control statistically for differences that existed before the treatme nt was implemented for the ACMC and SCMC groups ( Onwuegbuzie & Daniel, 2003). In this study, five separate ANCOVA tests were administered for the five continuous variables. In order to avoid the high probability of inflating the Type I error (rejecting the H o when it is true) for each statistical test a Dunn Bonferroni ew ) was used (Onwuegbuzie & Levin, 2005) Each test aimed to answer one quantitative research question. The following Null Hypotheses were tested: there is no difference between the means of the ACMC and SCMC groups when compared by the proposed measures. For each ANCOVA test, the following assumptions were examined carefully: (1) the independence of the observations, (2) equal or similar variance of the two groups, and (3) norma l distribution of the sample means (Glass & Hopkins, 1995). The independence of observations was assured through the random assignment of participants to either the ACMC or SCMC group within each class section that participated in the study. Further, the p articipants were randomly assigned to pairs which performed the pre writing discussions. In order to obtain final independent score, for each pair the post treatment score on each measure was calculated. This score was used for statistical analysis. The

PAGE 285

259 pr obability F values for each test were examined and equal or similar variance of the groups was confirmed. The normal distribution of the sample means was assured through the analysis of the Shapiro Wilk W values for each of the measures. This analysis reve aled that three out of the five measures had values lower than the SAS default level was/were affecting the distribution. The extreme scores were eliminated from the data set; this led to normal distribution of the data sets as suggested by the analysis of the data set; for the three measures with non normal distribution, the values wi th and without the extreme data sets are listed.

PAGE 286

260 Table 4.8 Descriptive Statistics: ANCOVA Datasets Measure (with /without extreme observations) n M SD Distributi on 1 T units Scores (Syntactic Complexity) 30/29 ACMC (with /without extr eme observations) extreme observation # 15 excluded 17/16 12.77/12.26 2.58/1.57 Non normal/ Normal .0002/.10 SCMC 13 12.78 1.1 2 Mean length of idea units (amount of information present in a single focus) 30/29 ACMC (with /without extreme ob servations) extreme observation # 15 excluded 17/16 8.57/8.36 1.07/0.64 Non normal/ Normal .006/.92 SCMC 13 8.05 0.83

PAGE 287

261 Table 4.8 (Continuing) Descriptive Statistics: ANCOVA Datasets Measure (with /without extreme observations) n M SD Distribution 3 Number of idea units (quantity of overall information present) 30/28 ACMC (with /without extreme observations) extreme observation # 7 and 13 excluded 17/15 18.35/16.07 7.35/3.74 Non normal/Normal .008/.56 SCMC 13 19.36 4.90 4 Lexical density (lexical information per clause) 30 ACMC 17 4.32 5.20 Normal .09 SCMC 13 4.74 5.40 5 Vocabulary complexity (frequency of the unique words used) 30 ACMC 17 0.40 0.11 Normal .45 SCMC 13 0.06 0.13

PAGE 288

262 The ANCOVA was condu cted with both data sets: the complete data set and the data set from which the extreme observation was eliminated. Neither ANCOVA test showed statistically significant difference between the groups; thus, it could be concluded, that the non normal distrib ution of the data did not affect the outcomes of the test. Further, the data of the tests performed with the full data sets are presented. In addition to the assumptions described above, before running the analysis of covariance (ANCOVA), one more assumpt ion was examined: the assumption of no interaction between the regression lines of the continuous, and categorical predictor variables. The regression lines for each group were studied and it was concluded that the parallel regression line assumption is tr ue. Based on this, the researcher concluded that it was appropriate to conduct the ANCOVA tests for the five continuous variables. ANCOVA results: syntactic complexity The first quantitative research question posed in the current study was: What is the di fference in the syntactic complexity present in the post treatment paragraphs of the students who participated in the synchronous versus asynchronous CMC pre writing discussion? The ACMC and the SCMC group results were compared based on their post treatmen t T unit scores, using as a concomitant variable their pre treatment T unit scores. The descriptive statistics for the measure are presented accompanied further by the results of the ANCOVA analysis. The examination of the descriptive statistics of the tw o groups in terms of their T unit scores, revealed that in the post treatment paragraphs the mean length of the T Units for the ACMC group was only slightly lower (12.77) than the mean length of the T units for the SCMC group (12.78), with a difference of 0.009. The standard deviation of the two groups was also similar: for the ACMC group it was 2.58, whereas for the SCMC

PAGE 289

263 group it was 1.53. There was also a similarity between the two groups in terms of their minimum scores: the lowest mean length the T uni ts produced by ACMC participants was 10.28 words, whereas for SCMC it was 10.58 words. However, the difference in the longest mean length T units between the two groups was much higher for the ACMC group it was 20.86 whereas for the SCMC group it was 16. 02 words. The results from the ANCOVA revealed that, when an adjustment was made for pre treatment scores of T unit mean length, there was no difference between the groups on T unit mean length post treatment scores ( F (28)=0.07, p =.88). Therefore, the Null Hypothesis proposed could not be rejected, and it was concluded that the two groups did not differ in terms of syntactic complexity as measured with T unit mean length scores. These results are presented in Table 4.9. ANCOVA results: information present i n a single focus The second quantitative research question posed in the current study was: What is the difference in the amount of information present within a single focus of the post treatment paragraphs of students who participated in synchronous versus asynchronous CMC pre writing discussion? The ACMC and the SCMC group results were compared based on their post treatment Idea Units Mean Length (IUML) scores, using as a concomitant variable their pre treatment IUML scores. Further, the descriptive statis tics for the measure are presented below followed by the results of the ANCOVA analysis. The descriptive statistics of the two groups revealed that in the post treatment paragraphs the mean length of the idea units for the ACMC group was only slightly hig her (8.571) than the mean length of the idea units for the SCMC group (8.057), with a difference of .514. The standard deviation of the two groups was also similar: for the

PAGE 290

264 ACMC group it was 1.075, while for the SCMC group it was .831. There was also a sim ilarity between the two groups in terms of their minimum IUML scores: the shortest mean length for the idea units produced by ACMC participants was 6.82 words, and for the SCMC participants it was 6.38 words. However, the difference in the longest mean len gth idea units between the two groups was much higher for the ACMC group it was 11.975 while for the SCMC group it was 9.635 words. The results from the ANCOVA revealed that when an adjustment was made for pre treatment scores of IUML, there was no diffe rence between the groups on IUML post treatment scores ( F (28)=2.08, p =.1606). Therefore, the Null Hypothesis proposed could not be rejected and it was concluded that the two groups did not differ in terms of amount of information present within a single fo cus as measured with IUML scores. These results are presented in Table 4.9. ANCOVA results: quantity of the overall information The third quantitative research question posed in the current study was: What is the difference in the quantity of the overall information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? The ACMC and the SCMC group results were compared based on their post treatment Idea Units Number (IUN) sco res, using as a concomitant variable their pre treatment IUN scores. The descriptive statistics for the measure are presented below followed by the results of the ANCOVA analysis. The descriptive statistics of the two groups revealed that in the post trea tment paragraphs the mean number of the idea units for the ACMC group was only slightly lower (18.35) than the mean length of the idea units for the SCMC group (19.39), with a

PAGE 291

265 difference of 0.3. The standard deviation of the ACMC group was higher (7.35) t han the standard deviation of the SCMC group (4.90). There was also a similarity between the two groups in terms of their minimum IUN scores: the lowest number of idea units produced by ACMC participants was 11.5 words, and for the SCMC participants it was 10 words. However, the difference in the highest number of idea units between the two groups was much higher for the ACMC group it was 36.5 while for the SCMC group it was 29 words. The results from the ANCOVA revealed that when an adjustment was made f or pre treatment scores of IUN, there was no difference between the groups on IUML post treatment scores ( F (28)=0.16, p =.67). Therefore, the Null Hypothesis proposed could not be rejected, and it was concluded that the two groups did not differ in terms of quantity of the overall information present as measured with IUN scores. These results are presented in Table 4.9. ANCOVA results: lexical information The fourth quantitative research question posed in the current study was: What is the difference in the lexical information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? The ACMC and the SCMC group results were compared based on their post treatment Lexical Density (L D) scores, using as a concomitant variable their pre treatment LD scores. The descriptive statistics for the measure are presented further followed by the results of the ANCOVA analysis. The descriptive statistics of the two groups revealed that in the po st treatment paragraphs the mean of the LD scores for the ACMC group was only slightly lower (5.20) than the mean length of the LD scores for the SCMC group (5.40), with a

PAGE 292

266 difference of 0.21. The standard deviation of the ACMC group was higher (0.95) tha n the standard deviation of the SCMC group (0.83). There was also a similarity between the two groups in terms of their minimum LD scores: the lowest lexical density score of the ACMC participants was 4 points, and for the SCMC participants it was 4.2 poin ts. The difference in the highest LD scores between the two groups was also close for the ACMC group the highest LD score was 7.73 while for the SCMC group it was 6.5. The results from the ANCOVA revealed that when an adjustment was made for pre treatmen t scores of DL, there was no difference between the groups on DL post treatment scores ( F (28)=0.64, p =.54). Therefore, the Null Hypothesis proposed could not be rejected, and it was concluded that the two groups did not differ in terms of lexical informati on present as measured with DL scores. These results are presented in Table 4.9. ANCOVA results: vocabulary complexity The fifth quantitative research question posed in the current study was: What is the difference in the vocabulary complexity present in t he post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? The ACMC and the SCMC group results were compared based on their post treatment vocabulary complexity (VC) scores, using as a conco mitant variable their pre treatment VC scores. The descriptive statistics for the measure are presented below followed by the results of the ANCOVA analysis. The descriptive statistics of the two groups revealed that in the post treatment paragraphs the m ean of the VC scores for the ACMC group was only slightly higher (0.46) than the mean length of the VC scores for the SCMC group (0.40), with a difference of

PAGE 293

267 0.06. The standard deviation of the ACMC group was higher (0.14) than the standard deviation of t he SCMC group (0.11). There was also a similarity between the two groups in terms of their minimum VC scores: the lowest lexical density score of the ACMC participants was 0.30 points, and for the SCMC participants it was 0.31 points. The difference in the highest VC scores between the two groups was also close for the ACMC group the highest VC score was 0.74 while for the SCMC group it was 0.58. The results from the ANCOVA revealed that when an adjustment was made for pre treatment scores of DL, there wa s no difference between the groups on DL post treatment scores ( F (28)=0.35, p =.22). Therefore, the Null Hypothesis proposed could not be rejected, and it was concluded that the two groups did not differ in terms of vocabulary complexity as measured with VC scores. These results are presented in Table 4.9.

PAGE 294

268 Table 4.9 ANCOVA Results Measure Pr>F DF Null Hypothesis, Varia nce T units Scores (Syntactic Complexity) 0.07 28 Failed to reject (p=.88) Equal Mean length of idea units (amount of information present in a single focus) 2.08 28 Failed to reject (p=.16) Equal Number of idea units (quantity of overall information present) 0.16 28 Failed to reject (p=.67) Equal Lexical density (lexical information per clause) 0.64 28 Failed to reject (p=.54) Equ al Vocabulary complexity (frequency of the unique words used) 0.35 28 Failed to reject (p=.22) Equal Mann Whitney U Test Three of the proposed eight measures used ordinal scores. They were analyzed applying three separate non parametric two tailed Mann W hitney U tests. Following are the results of the proposed comparison of the two groups for each measure. Rhetorical soundness The ACMC and SCMC groups were compared in terms of their mean dyad scores presenting the evaluation of rhetorical soundness of e ach post treatment writing. The Null Hypothesis for the test was that there was no difference

PAGE 295

269 between the means of the two groups. The test failed to reject the Null Hypothesis at 95% confidence level with an obtained p =.89. Presentation and development The ACMC and SCMC groups were compared in terms of their mean dyad scores presenting the evaluation of presentation and development quality of each post treatment writing. The Null Hypothesis for the test was that there was no difference between the means of the two groups. The test failed to reject the Null Hypothesis at 95% confidence level with an obtained p =.22. Overall language use The ACMC and SCMC groups were compared in terms of their mean dyad scores presenting the evaluation of presentation and development quality of each post treatment writing. The Null Hypothesis for the test was that there was no difference between the means of the two groups. The test failed to reject the Null Hypothesis at 95% confidence level with an obtained p =.65. Pulled Scores Analysis univariate tests, posed serious limitations to the study. When performing the five ANCOVA tests, the possibility of not finding statistically significant effects was high, even thoug h if each of the analyses was performed separately or a fewer ANCOVA tests were conducted, at least one of the effects could be detected (Onwuegbuzie & Levin, 2005). In order to avoid this limitation, the Three Step Approach for Testing Multiple Univariate Hypotheses proposed by Onwuegbuzie and Levin (2005) was applied. A univariate test of averaged standardized outcomes of the eight variables was used. First, each of the continuous outcome measure scores was standardized ( z scores were calculated) making s ure that the z scores on each measure were similarly oriented. For

PAGE 296

270 the ordinal scores, the percentile rank was first calculated and then converted to z scores. Then, the z scores for each variable were summed and averaged across measures on a dyad level, thus yielding a composite z score for each dyad. Further, an independent samples t test of group mean difference was performed in order to compare the ACMC and SCMC groups in terms of the average composite measure. The Null Hypothesis tested was that the re is no significant difference between the z score means of the synchronous and the asynchronous groups. The goal of the test was to either confirm or reject the Null Hypothesis; in other words, to infer whether the (Cody & Smith, 2006, p. 183) or would differ in the population of English language learners of intermediate level of proficiency. The researcher did not hypothesize a priori if any of the two compared means was greater than the other; thus, a two tailed t test was performed. Before conducting the t test, the researcher carefully examined the following assumptions and concluded that they were met: (1) the independence of the groups, (2) normal distribution of the sample means, and (3) equal or similar variance of the two groups. The random assignment of the participants to either ACMC or SCMC group and the consideration of the mean pair scores rather than individual scores assured the compliance with the first assu mption. The Shapiro Wilk W for the ACMC group was =.05. These results revealed that the ACMC group had a non normal distribution. After examining the extreme scores, the r esearcher concluded that there was no error in the score calculation.

PAGE 297

271 The normality of the data sets was again tested without the extreme observation and it was concluded that its exclusions led to normal distribution (ACMC group W=0.48). The independent two tailed t tests were performed with and without the extreme observations and in both cases, the results were not statistically significant. Therefore, this allowed the researcher to conclude that there is sufficient evidence that the test outcomes were not affected by the non normal distribution of the data. Further, the results of the full data set are reported. Finally, the probability F value of 0.32 was obtained; this F which allow ed to conclude that equal group variances and probability values must be used when interpreting the test outcomes. The descriptive statistics are presented in Table 4.10. Table 4.10 Pulled z scores: Descriptive Statistics Pulled Z scores n M SD ACMC 17 0.02 0.70 SCMC 13 0.08 0.52 For the pulled z scores scores, a two tailed t test was conducted to compare the means of the ACMC group ( M = 0.02, SD = 0.70) and SCMC group ( M = 0.08, SD = ignificance, t (28) = 0.32, p =

PAGE 298

272 .0.69. The mean difference was 0.93 at 95% confidence interval ( 0.57, 0.38). Based on these results, the researcher concluded, that the difference between the pulled z score means of the ACMC and the SCMC group are not stat istically significant. The outcomes of the t test are presented in Table 4.11. Table 4.11 Independent t test of the Pulled z scores: Statistics and Outcomes Pulled z scores DF Null Hypothesis: H 0 1 2 = 0, Probability P Variance ACMC SCMC 28 Failed to reject ( p =.69) 0.32 Equal Quantitative Results Summary The results obtained with analysis of covariance inferential statistics showed that there was no difference between the post treatment paragraphs of the ACMC and t he SCMC groups in terms the eight proposed measures: (1) the syntactic complexity (measured by calculating the mean length of T units), (2) the amount of information present in a single focus (measured by mean length of idea units), (3) the quantity of ove rall information present (measured by the number of idea units) (4) lexical information per clause (measured through lexical density analysis), (5) vocabulary complexity (measured by analyzing the frequency of the unique words used), (6) rhetorical soundn ess, (7) presentation and development of main ideas, and (8) overall language use. In other words, based on these results it could be concluded that the mode

PAGE 299

273 of CMC communication used for the pre writing discussion did not affect the writing outcomes in th e first draft A possible reason for these outcomes might be that the development of writing skills that would be detected in the surface (vocabulary, syntactical complexity) and deep textual level (presentation and development, overall language use, rhet orical soundness) is a long process, the changes over a short period of time are not dramatic and could not be detected when the comparison is performed on a small sample. With a larger sample size, differences between the groups might be detectable. Howev er, the researcher noticed quantitative and qualitative differences in the treatment texts produced by the students during their CMC interactions. The ACMC discussions involved the production of longer sentences but involved fewer postings as compared to t he ACMC. The mean number of turns for the ACMC discussions was 5; the ACMC discussions had a mean number of turns 46.46. They were shorter in terms of number of words. The mean length of the ACMC discussions was 272.76 words, whereas the mean length of the SCMC discussion was 410.46. Similar differences between These differences called for further examination of the post treatment language data in its connection with the lan guage data produced during the treatment. These connections were analyzed, and this analysis is presented in the following section. Additional Quantitative Findings Design of the Additional Quantitative Research In addition to the statistical analyses d escribed in the preceding sections of this chapter, the pre writing CMC interactions and the post treatment paragraphs were

PAGE 300

274 examined in order to find out the degree of intertextual relationship between the CMC interactions and the post treatment paragraphs This analysis was based on distinct lexical items that students used both in the pre writing discussion and the post treatment paragraph. The research question posed was: Do the ACMC and the SCMC modes of the pre writing discussion influence to a differe nt degree the texts produced after those discussions as measured by matching distinct lexical items? The texts of the pre writing discussions were normalized and the lexical items in the pre writing discussions, and the post treatment paragraphs were iden tified using the procedures described earlier in Chapter 3. Further, each identified lexical item was morphological paradigm are the same lexical item: for example, differ, d iffered, different, difference, differing, differently 65). Thus, if inflectional or derivational forms with the same root were detected, they were considered to be the same lexical item. Note that the lists of the most frequently used 1 999 words and 1000 2000 words compiled by Nation (2001), were already sorted by lexical items. writing discussions and paragraphs that were not found in the above men tioned lists were identified by the researcher using the text analysis software designed for the study; these words were compiled into a third list named Above 2000 Words List ; the words were classified by lexical items. Two independent inter raters verifi ed the accuracy of the Above 2000 Words List The inter rater reliability was higher than 99% with both inter raters. The few instances of discrepancies were discussed and corrected; these instances were based on

PAGE 301

275 human error rather than on disagreement abo ut the list structure. Further, the finalized Above 2000 Words List was added to the database and was compared against the 1 999 and 1000 2000 Most Frequently Used Words Lists (Nation, 2001) aiming to assure that none of the lexical items appeared in more than one list. Thus, the researcher obtained three lexical item lists: (1) 1 999 Most Frequently Used Words Lists compiled and published by Nation (2001), (2) 1000 2000 Most Frequently Used Words Lists compiled and published by Nation (2001), and (3) Abo ve 2000 Words List compiled by the researcher. These three lists were organized by lexical items and were further used to analyze the intertextual relationship between the texts produced during the treatment tre atment drafts. While in the previous analyses, the unique lexical items were considered, in other words each occurrence of a lexical item was identified and counted, in this analysis each lexical item was counted only once. For example, if the words house houses appeared one or more times in either the CMC interaction or the post treatment paragraph, they were counted as one distinct lexical item (DLI). In order to compare the degree of the pre writing discussion influence over the post treatment paragra phs for the two groups participating in the study, namely the ACMC group and the SCMC group, the following steps were performed. (1) The DLIs were identified for each of the CMC pre writing interactions; this allowed the researcher to obtain the overall CM C DLI score for each pair of participants. (2) The DLIs were identified for each of the post treatment paragraphs. (3) The DLIs in each post treatment paragraph were matched against the corresponding pre writing CMC discussion and the DLIs used in both the preceding pre writing CMC discussion and the post treatment

PAGE 302

276 paragraphs were identified obtaining a matching DLI score for each participant. (4) Further, the matching DLI scores were used to calculate the ratio of the matching DLIs to the overall CMC DLI s core in the corresponding pre writing CMC discussion. The formula used for the ratio calculation was: Matching DLI score x 100 = Matching DLI Ratio Score (1) Overall CMC DLI score Thus, for each of the participants a matching DLI ratio score was calculated. (5) The last step was to calculate the mean matching DLI ratio score for each pair in order to obtain independent matching DLI ratio scores that could be further used for statistical comparison.

PAGE 303

277 Figure 4.2 DLI Data Analysis Flow These steps were performed using the text analysis software designed for the study. In order to assure that the software scripts were running correctly, the researcher 1 Identify DLI in CMC discussions Overall CMC DLF score 2. Identify DLI in post treatment paragraphs. Match DLIs 3. Obtain Matching DLI Scores 4. Obtain Matching DLI Ratio Score Calculate ratio 5. Obtain Matching DLI Ratio Score mean for each pair Conduct statistical analysis Calculate Mean

PAGE 304

278 and the programmer independently calculated the scores for 10% of the data and compared them to the software outcomes. The results matched 100% which allowed the researcher to conclude that the scripts were running appropriately and all generated results were correct. The matching DLI ratio scores were analyzed compared through a t test. The Null Hypothesis tested was that there is no significant difference between the matching DLI ratio means of the SC MC and the ACMC groups. The goal of the test was to either confirm or reject the Null Hypothesis; in other d differ in the population of English language learners of intermediate level of proficiency. The alternative to the Null researcher did not hypothesize a priori if any of t he two compared means was greater than the other; thus, a two tailed t test was performed. Results of the Additional Quantitative Research Before conducting the independent t test, the researcher carefully examined the following assumptions and concluded t hat they were met: (1) the independence of the groups, (2) normal distribution of the sample means, and (3) equal or similar variance of the two groups. The random assignment of the participants to either ACMC or SCMC group and the consideration of the mea n pair scores rather than individual scores assured the compliance with the first assumption. The comparison of the Shapiro Wilk W=0.6772 for the ACMC group and Shapiro Wilk W=0.7125 for the SCMC groups with the SAS

PAGE 305

279 default level of significance, alpha=.05 allowed to conclude with 95% confidence that there was a sufficient evidence that the means were normally distributed. Finally, the probability F value of 0.1128 was obtained; this F value was greater than the SAS default level of significance, alpha=.05 which allowed to conclude that equal group variances and probability values must be used when interpreting the test outcomes. For the distinct lexical item (DLI) scores, a two tailed t test was conducted to compare the means of the ACMC group ( M = 40.42, SD = 13.66) and SCMC group ( M = 26.30, SD t (28) = 3.25, p = .003. The Mean difference was 14.11 at 95% confidence interval (5.23, d (1.20) signified a large eff ect size. Based on these results, the researcher concluded, that probability of obtaining higher matching DLI ratio score means for the ACMC group than for the SCMC group from a representative population under the same conditions was highly unlikely. Thu s, the intertextual connections between the pre writing interactions and the first drafts for the ACMC group participants were significantly stronger than for their SCMC counterparts. The descriptive statistics are presented in Table 4.12 and the outcomes of the t test are presented in Table 4.13.

PAGE 306

280 Table 4.12 Independent t test of the Distinct Lexical Items (DLIs): Descriptive Statistics N DLI M DLI SD DLI Skewness DLI Kurtosis ACMC 17 40.42 13.66 0.02 0.93 SCMC 13 26.30 8.63 0.42 0.42 Table 4. 13 Independent t test of the Distinct Lexical Items (DLI): Statistics and Outcomes DLI DF Null Hypothesis: H 0 1 2 = 0, Probability P, Variance ACMC SCMC 28 Rejected (p=.0030) 0.1128 Equal Summary of the Additional Quantitative Findings The results of the DLI scores t test suggested that, as measured by DLIs, the ACMC and the SCMC modes of the pre writing discussion influenced to a different degree the texts produced after these discussions. In the post treatment writings, the students who participated in the ACMC pre writing discussions used to a greater extent the DLIs from their pre writing discussions as compared to the SCMC group. Thus, it was concluded that the intertextual relation on a lexical level between the pre writing

PAGE 307

281 discussion performed via ACMC mode (treatment) and the writings that followed these discussions (post treatment) was stronger as compared to the same criteria for the SCMC group. This comparison shed some light on the intertextual relationships between these texts. It allowed inferring that the degree to which the CMC pre writing discussion influenced the first draft was affected by the mode of the pre writing discussion; more specifically, the ACMC pre writing discussion seemed to influence to a higher degree the first drafts. This conclusion is based on the assumption that the same DLI might have conveyed similar or Chapter Summary In this chapter, the data analysis and results of the quantitative and q ualitative stages of the study were presented. Further, based on the qualitative stage findings related to the nature of language produced in the two different modes of CMC in which the treatment pre writing task was performed, additional data analysis was proposed, and the outcomes of this analysis were presented. The discussion of the findings reported in this chapter as well as the implications and recommendations related to these findings are presented in Chapter VI: Summary, Discussions, Recommendation s, and Implications. The following chapter presents eight instrumental case studies of the eight students who benefitted the most and the least from the treatment task; these case studies aim to disclose the specifics of the investigated first stages of t he writing process.

PAGE 308

282 CHAPTER V: INSTRUMENTAL CASE STUDIES Introduction In this chapter, eight single instrumental case studies with unit of analysis intermediate ESL learners are presented. I approached the presentation and the analysis of my observat ions within the context of the writing as a process approach focusing on the first two stages of the five stage process: the pre writing and drafting stage (Grabe & ba ckgrounds, the environment, the opportunities for dyadic interaction during the pre writing stage, and the influence of the outcomes of these interactions were taken into account when presenting each case study. Thus, the dyadic brainstorming, the influenc e of computer mediated environment and the intertextual connections of the first drafts were investigated. My analysis was informed by socio cultural approach to interaction as well; the model of dyadic interaction proposed by Storch (2002) was used to int erpret the writing task completion. The cases were chosen using Extreme/Deviant sampling technique (Kemper et al., 2003). This purposive sampling allowed the identification of the most outstanding cases am ong the participants in the quantitative stage of the study ( n =60) based on writing gains as a result of the treatment. Thus, the students who demonstrated the highest and the lowest gains in their writings after participating in a CMC pre writing task wer e the actors of these case studies.

PAGE 309

283 As described in Chapter III: Methods, the gains were measured by comparing the pre treatment and post analysis, namely: (1) the syntactic complexity ( measured by calculating the mean length of t units), (2) the amount of information present in a single focus (measured by mean length of idea units), (3) the quantity of overall information present (measured by the number of idea units) (4) lexical inform ation per clause (measured through lexical density analysis), (5) vocabulary complexity (measured by analyzing the frequency of the unique words used), (6) rhetorical soundness, (7) presentation and development of main ideas, and (8) overall language use ( the last three criteria were assessed using a multiple trait rubric). In order to obtain comparable standard scores, for each of the above criteria z scores were calculated (Glass & Hopkins, 1995) ; this allowed the identification of the extent and the dire ction to which the scores of each participant for the eight proposed criteria deviated from the group distribution mean. Further, the pulled average pre and post treatment z scores for each of the 60 participants in the study were obtained; based on the d ifference between the post treatment and pre treatment pulled average scores the z score average gain was calculated and the case study actors were identified. Table 5.1 presents these participants and their pulled average z scores.

PAGE 310

284 Table 5.1 Instrumental Case Study Participants: Average Gains Achievement Level and CMC Mode Case Study Actor Z Score Average Gain Low ACMC Shin 1.218328443 Azad 0.744473159 High ACMC Felipa 1.783441878 Isabella 1.31275934 Average Gains: ACMC Group (n=34) 0.01355427 Low SCMC Ajwad 0.656827894 Kamil 0.636001647 High SCMC Kang 0.64492224 Sun 0.460250215 Average Gains: SCMC Group (n=26) 0.051795905 Average Gains CMC (n=60) 0.014764139 As is shown in Table 5.1, the eight students selected for the instr umental case studies are grouped into four pairs: (1) two students participated in an ACMC pre writing interaction and demonstrated the lowest pulled gain, (2) two students participated in an ACMC pre writing interaction and demonstrated the highest pulled gain, (3) two students participated in an SCMC pre writing interaction and demonstrated the lowest pulled gain, (4) two students participated in an SCMC pre writing interaction and demonstrated the highest pulled gain. Appendix 5.1 presents the complete l ist of z scores for the case study participants.

PAGE 311

285 The case study analysis and comparison aims to provide an understanding of how ESL students of intermediate level proficiency who had high and low gains in their post treatment writings participated in one pre writing CMC task and applied the results of this pre writing interaction in their paragraphs. The pre writing discussion task, presented in detail in Chapter III: Methods, provided the participants with six images united by the pictures and prepare to write a paragraph on the topic. The specific research questions addressed in the case studies were as follows: Research Question 1: How do peers participate in asynchronous and synchronous pre writing CMC interactions? Research Question 2: (a) How do they use the specific ideas and language generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these idea s? Thus, the pre writing CMC discussion, the text of the draft created based on this discussion, and the intertextual connections between the pre writing discussion and the following draft were the focus of each instrumental case study analysis. The inter textual analysis for each of the cases was performed on three separate levels. These levels included: (1) lexical, (2) organizational, and (3) textual level, with idea units being used as a unit of analysis. The intertextuality on lexical level was identi fied based on the matching DLI scores obtained during the quantitative stage of the study. The working definition of DLI was: the same lexical items which appeared one or more times in either the CMC interaction or the post treatment paragraph were conside red to present one distinct lexical item (DLI). These sores represented the lexical items that were used both in the pre

PAGE 312

286 writing discussion and the following it draft. The consideration of the DLI scores allowed me to make suggestions about the strength of the pre writing discussion and the post treatment paragraph connections on a lexical level. The organizational level analysis was conducted relating the pre writing discussion organizational suggestions such as topic sentence, supporting ideas, and concl usions, to the draft organization. In addition, I accounted for the sequencing of ideas in the discussion and whether the same or similar sequence was evident in the draft. This analysis allowed me to suggest possible connections between the pre writing di scussion and the post treatment paragraphs on an organizational level. Finally, the textual (idea unit) level analysis of intertextuality was performed through segmenting the pre writing discussion texts and the post treatment writing using language strin writing discussions was compared against the idea units from the draft texts. The intertextual connections between the discussion and draft idea units were interpreted and new idea units composed i n the draft were identified. The following working definition of idea units was used in the process of e verb phrase along with (Chafe, 1985, p. 106); however, if a complement or restrictive relative clause is present or indirect question or indirect quotation is used, these b elong to the idea unit presented by the main clause. Dependent clauses, appositives (including examples provided by the ...), and participial clauses

PAGE 313

287 (past and gerund) were considered to be separate idea units. When analyzing CMC discourse production, single phrases indicating agreement, disagreement, understanding, etc. were considered as separate idea units emoticon, etc. The intertextual connections were pres ented and interpreted in the context of following notation was used in the presentation of idea unit analysis of intertextuality: IUx x treatment writing, IUx* writing discussion), and IUx** writing discussion). In the analysis, I approached the idea units in each CMC pre w riting interaction based on their relation to the topic under discussion. If an idea unit was perceived to As you know most of Americans came from other countries non as you know Americans came from other countries. follow more precisely the connections. During the idea unit analysis and comparison, I found that the intertextual relations of the draft idea units and the id ea units from the pre writing were of different level of intensity. Some draft idea units were either identical or very similar to idea units from the discussion, while others were new. Thus, four levels which presented the

PAGE 314

288 strength of intertextual connect ions emerged from the data: high, medium, low, non existent The high level of intertextual connections signified idea units in the post units from the pre writing di scussion. They included the following categories: (1) and (4) own idea units f rom the discussion rephrased. Medium level included idea units with a relation to the pre writing discussion being less direct. The following categories were identified and considered in the analysis: (5) new paragraph idea units based on it(s), (6) new paragraph idea units based on own idea unit(s), (7) new categories for the low level of intertextual connections included new idea units more specifically: (8 ) new paragraph idea units which presented a development or development of own topic, (10) new paragraph idea units which presented a development Finally, the non existent intertextuality was identified in (11) new paragraph idea units which presented a development of a new topic not traceable in the discussion. To assure the consistency of the analysis, two complete case studies (25%) were reviewe d by a trained peer interrater. In addition, this interrater was also well informed about the specifics of the intertextual analysis and had recent experience in using idea unit as a unit of intertextual analysis. No significant discrepancies were identifi ed by the interrater in the reviewed case studies.

PAGE 315

289 In this chapter, each instrumental case is presented separately with a subsequent analyzing, and comparing a collec tion of cases rather than one or two separate cases would reveal better the processes under investigation. The findings of the study are not conclusive, due to the limited number of participants; however, taking into account the limited research related to the pre writing and first paragraph stages of the writing process in the field of ESL, I believe that analyzing one pre writing interaction produced by each of the extreme case actors and the first drafts that followed these interactions would shed some l ight on the processes occurring in the pre writing stage and their connections with the first draft writing. In my view, the findings will inform further research in this area as well as instructional practices related to the use of CMC in the ESL writing class. The Case Studies ACMC Actors, Low Gains Case Studies Case Study One: Azad I met Azad in the beginning of the Fall 2006 semester. He just arrived into the United States along with a group of Arabic students from Saudi Arabia and Qa tar. Due to the large number of students enrolled in the fall semester, it was difficult for me to have detailed observations of his class interactions; thus, my presentation of Azad is mainly based on the analysis of his computer interactions with other s tudents and the answers he provided during our ACMC interview. Azad was a 19 year old male student from Qatar. He graduated from high school where he mainly focused on studying science related subjects. In the ACMC interview,

PAGE 316

290 Azad was concise and straight to the point with his answers. He would avoid providing details or extending his answers after my follow up clarification questions. In addition, he avoided answering some of the interview questions, mainly those related to his opinion on writing. When I approached him during one of the breaks and asked him to post It is hard, teacher interview is provided in Appendix 9. When sharing with me his past educational experiences during our ACMC interview, Azad answered that it was mainly teacher centered, which he did not like. He s are kept postings, it appeared that he was quite comfortable with computers. While his computer use for educational purposes before coming to the U.S. was limited to a computer class only and he did not report using computer or CMC for problem solving tasks and research, he stated that he frequently used computers for chat and email. As with the majority of the students at the ELI, this most probably was one of the main means of communication with friends and family back home, although Azad did not specify this in his answer. The treatment interaction The treatment task, presented in Appendix 4, was completed by Azad in an asynchronous environment using ICA2 conferencing function The pre writing interaction task required a discussion of six related pictures united by the Following is the complete log of their ACMC pre writing interaction.

PAGE 317

291 Log 5.1 ACMC Pre writing Discussion: Azad and P artner Message 1 Azad FROM: Azad (02/17/06 6:26 PM GMT 06:00 1 I think that the first picture shows that the population of the U.S. are in the most 2 U.S. areas, in the north, south east and west... 3 4 Message 1 Peer FROM : Peer (02/20/06 1:57 PM GMT 06:00 5 Ok. I think first picture shows the population of United States are too many so they 6 don't need any more immigrants. 7 Message 2 Azad FROM: Azad (02/20/06 1:58 PM GMT 06:00) 8 I think the second picture shows that so me of the big buldings in one of the U.S. 9 cities. 10 Message 3 Azad FROM: Azad (02/20/06 2:02 PM GMT 06:00) 11 the third picture show the differents between Americans, because some of them live 12 in small houses and some of them in big houses. 13 Message 2 Peer FROM: Peer (02/20/06 2:03 PM GMT 06:00) 14 Yes, second picture means a city that have many buildings. 15 Message 3 Peer FROM: Peer (02/20/06 2:07 PM GMT 06:00) 16 Third one means residence that americans live. 17 Message 4 Azad FROM: Azad (02/20/06 2:08 PM GMT 06:00) 18 I think the fourth picture shows that the U.S. immigrants have been coming from 19 different countries like Spain, England, Korea ...etc. 20 Message 5 Azad FROM: Azad (02/20/06 2:10 PM GMT 06:00 21

PAGE 318

292 the fiveth picture is like the second picture. 22 Mes sage 4 Peer FROM: Peer (02/20/06 2:11 PM GMT 06:00) 23 I think the fourth mean a race diversity. As you know most of americans came from 24 other countries. Therefore there is a culture pluralisim in U.S. 25 Message 5 Peer FROM: Peer (02/20/06 2:13 PM GMT 06 :00 26 The last picture means location where America is placed and big city. 27 Message 6 Azad FROM: Azad (02/20/06 2:14 PM GMT 06:00) 28 Fainally, the sixth picture mean th size of the U.S. and hoe are the states are far 29 from each other. 30 Azad and hi s partner posted more times than other students who participated in ACMC pre writing interactions: the average number of turns for the ACMC group was 5, while for this dyad, it was 11. There is a possibility that this higher number of messages was a resul t of the discussion strategy they applied. Based on the time tags, it appears that all of the messages except for the first one were made within 16 min between 1:58 PL GMT and 2:14 PM GMT; thus, their interaction was synchronous although performed using a CMC tool designed for asynchronous communication. This affected the dynamics of the interaction: it shared features of both synchronous and asynchronous discussion: as A CMC message in this study and they consisted of one or two sentences, while the in all of the other ACMC interactions considered in the study each message would contain numerous sentences. Further the number of the messages posted by Azad and his partner w as higher. However, their messages also had features of asynchronous communication

PAGE 319

293 use of complete sentences, writing longer sentences, and posting two or three sentences within one message. During the pre writing discussion, Azad and his partner demons trated dominant patterns of interaction. Despite the numerous postings (six made by Azad and five by his partner), they did not show the ability to create a discussion space online. Azad, in his message 1 (lines 1 4) in the pre writing discussion indicated an attempt to set the discussion in a collaborative mode: he initiated the discussion stating his opinion about the first picture from the discussion task and invited his partner to express his opinion. However, this was the only posting that showed an at tempt to create a shared discussion space. Further, Azad proceeded with the description of the second and the third picture (Messages 2 and 3, lines 8 addressed different aspects of the first picture Their interaction consisted of postings of their own interpretations of the pictures which differed between the two participants; neither Azad nor his partner showed signs of considering, building on or extending the tribution to the discussion was similar in terms of quality and quantity: both of them addressed only part of the task (their interpretation of the pictures) and failed to discuss the future paragraph organization. Moreover, they posted a similar total num ber of idea units: Azad 16 and partner 15. The number of units: Azad 8, partner 11. In his interview, Azad shared satisfaction with the way he completed the ta sk and I did a

PAGE 320

294 writ ing discussion and the appreciation for the opportunity to share ideas, when asked if he used the discussion ideas when composing his paragraph, Azad stated that he used new ideas when writing I write new ideas Azad also perceived the t ask to be helpful for his writing skills development because it supported the improvement of his spelling. Despite the overall positive attitude and satisfaction with the task completion, Azad expressed concerns related to time limitation which prevented h It was not very helpful, bacause we did not have a time to write many things I accept this comment with some reservations, as it is evident from the time tags that most of the discussion too k place for only 16 minutes. My impression was that before their CMC session, each one of them prepared something to post, and they logged into ICA2 to post this prewritten material rather than to engage in discussion. The paragraph: intertextual connecti ons Based on the gain scores of his pre and post treatment paragraphs, it could be concluded that Azad did not benefit greatly from the ACMC pre writing discussion. The gain scores he obtained for his post treatment writing were negative: .656827894 as m easured by the difference of the pulled pre and post treatment z scores. On six of the eight measures, he performed worse in his post treatment writing demonstrating a negative gain (refer to Appendix 5.1 for complete list sc ores). The two criteria on which he improved his writing were the amount of information present in a single focus (measured by mean length of idea units) and lexical information per clause (measured through lexical density analysis). Further, I address the treatment paragraph written

PAGE 321

295 following the ACMC pre writing discussion in order to make inferces about the strategies he applied when using the pre writing discussion during the composition of his first draft. Durin g their discussion, the partners did not address paragraph organization, although this was explicitly requested in the task description. This again suggests that they did not invest enough time and effort into task discussion. The organization of st treatment paragraph showed that he reorganized the ideas shared during the further in ACMC Paragraph Comparison Matrix 5.1: Azad.

PAGE 322

296 ACMC Matrix 5.1 Paragraph Comparison: Az ad Message 1: FROM: Azad (02/17/06 6:26 PM GMT 06:00 I think IU1* that the first picture shows IU2* that the population of the U.S. are in the most U.S. ar eas, in the north, south east and west... IU3* How about you Partner?... IU4* The number of people of the U.S is very big IU1 IU3** [ Rephrased synonyms ] and there are many kinds or types of people. IU2 IU10** [ Rephrased synonyms & sy ntax changes ] First, there are many immigrants who have been coming from different countries IU3 IU13* [ Exactly the same own shortened ] and they are all of Message 1 FROM: Partner (02/20/06 1:57 PM GMT 06:00 Ok. I think IU1** first picture shows IU 2** the population of United States are too many IU3** so they don't need any more imigrants. IU4**

PAGE 323

297 ACMC Matrix 5.1 (Continued) Paragraph Comparison: Azad t) Message 2: FROM: Azad (02/20/06 1:58 PM GMT 06:00) I think IU5* the second picture shows IU6* that some of the big buldings in one of the U.S. cities. IU7* U.S. area. IU4 IU3* [ Own, rephrased syntax changes for cohesion transformation and shor tened drops PP] Second, in many U.S. cities there are some huge buildings IU5 IU7 *[ Own, rephrased. Synonyms, content change] which made it by Americans people. IU6 [ New, develops new Message 3: FROM: Abdulaziz Alemadi (02/20/06 2:02 PM GMT 06:00) t he third picture show the differents between Americans, IU8* because some of them live in small houses IU9* and some of them [live] in big houses. IU10*

PAGE 324

298 ACMC Matrix 5.1 (Continued) Paragraph Comparison: Azad Postings (by Idea Unit) Message 2 FROM: Partner (02/20/06 2:03 PM GMT 06:00) Yes, second picture means a city IU5** that have many buildings. IU6** Message 3 FROM: Partner (02/20/06 2:07 PM GMT 06:00) Third one m eans residence IU7** that americans live. IU8** topic extending own topic from IU7* ] Third, the people in U.S. are different, IU7 IU8* [ Own, rephrased syntax change of predicate center cohesion .] because some of them are rich IU8 [ New, develops new t opic extending own topic from IU10*] and they live in big houses, IU9 IU10* [ Own, rephrased synonym change and added verb ] and some of them are normal IU10 [ New, develops new topic extending own topic from IU9*] and live in regular Message 4: FROM : Azad (02/20/06 2:08 PM GMT 06:00) I think IU11 the fourth picture shows IU12* that the U.S. immigrants have been coming from different countries like Spain, England, Korea ...etc. IU13*

PAGE 325

299 ACMC Matrix 5.1 (Continued) Paragraph Comparison: Azad P Unit) Message 4 FROM: Partner (02/20/06 2:11 PM GMT 06:00) I think IU9** the fourth mean a race diversity. IU10** As you know IU11** most of americans came fr om other countries. IU12** Therefore there is a culture pluralisim in U.S. IU13** Message 5: FROM: Azad (02/20/06 2:10 PM GMT 06:00 the fiveth picture is like the second picture. IU14* houses. IU11 IU9* [Own, rephrased content change] finally, Mess age 5 FROM: Partner (02/20/06 2:13 PM GMT 06:00 The last picture means location IU14** where America is placed and big city. IU15** Message 6: FROM: Azad (02/20/06 2:14 PM GMT 06:00) Fainally, the sixth picture mean th size of the U.S. IU15* and hoe are the states are far from each other.[10] IU16*

PAGE 326

300 ACMC Matrix 5.1 (Continued) Paragraph Comparison: Azad Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color own idea unit 4. Blue Highlight Color

PAGE 327

301 writing discussion showed a rather low level of shared lexical items. The ratio of the distinct lexical items matching in the ACMC discussion a nd the paragraph to the overall CMC discussion distinct lexical items was 26.09, while the same ratio for the ACMC group was 41.77 and 34.30 for the ACMC and SCMC groups combined. This suggests a weak relation between the two texts that of the pre writin g discussion and the following draft. In order to reveal the dynamics of the intertextual connections of the two texts, these connections were further analyzed on an idea unit level. The content idea units were identified, the organizational phrases such I think were considered to be non content idea units. Azad did not use any of the non content idea units from his pre writing discussion in his paragraph. The total number of idea units Azad composed in his paragraph was 11. Only two of these 11 idea units (IU1 and IU2), (IU3** and IU10**) as stated in the discussion. They were rephrased by using the popul ation number of people race diversity many types or kinds of people the f ourth [picture] mean there are unchanged idea unit of his own from the pre writing discussion (IU 3 IU13*), which constituted 9.1% of the total number of idea units in his post treatme nt writing; however, he shortened it omitting the list of countries he provided in the discussion.

PAGE 328

302 rephrased idea units from his own postings. He rephrased five idea units usin g various rephrasing strategies (IU4 IU3*, IU5 IU7*, IU7 IU8*, IU9 IU10*, and IU 11 IU9*). The following specific rephrasing strategies were used: (1) syntax changes that affected the subject and were made to achieve paragraph cohesion [for examp le, in idea the people of the US they some of them they example he shortened idea unit4 by simplifying a phrase and omitting a list of clarifying examples: all of US areas big replaced with huge one of the US cities many US cities small houses regular houses rephrasing techniques were performed within the boundaries of the b orrowed idea units. He composed three new idea units (IU6, IU8, and IU10) as an extension of topics proposed in his own idea units, which constituted 27.27% of his writing. wit h his own ideas stated in the pre writing discussion, they were either idea units that were rephrased (45.45%) or new idea units influenced by the his own ACMC postings (27.27%). Only 18.18% of the idea units that Azad composed in his post treatment paragr

PAGE 329

303 Table 5.2 Paragraph idea units: Azad Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 1 0 2 5 0 0 0 0 3 0 0 11 Paragraph % 9.1% 0% 18.18% 45.45% 0% 0% 0% 0% 27.27% 0% 0% 100% Criteria Coding Scheme: (1) (2) Own verbatim (3) (4) Own rephrased (5) (6) New based on own i dea unit (7) (8) (9) New development of own topic (10) (11) New no traceable connections with pre writing discussion

PAGE 330

304 In tertextual connections: conclusions Based on this analysis, I would interpret paragraph was connected more strongly with his own ideas developed during the pre writi ng discussion; in addition to the rephrased own ideas, the three new ideas included in was marginal, and did not influence to a great extent the paragraph composition Thus, the dominant pattern, demonstrated in the pre paragraph as well: he mainly worked with his own ideas, rephrasing and extending them. writing discussion and the use of its outcomes were approached as if they were outcomes of an individual brainstorming rather than peer collaboration results. He used four different strategies for rephrasing: (1) syntax changes that affect the subject or the verb and were made to achieve paragraph cohesion, (2) idea unit clipping, (3) use of synonyms, (4) content change. When examining the paragraph idea units in which the intertextual connections with the discussion idea units were identifiable, I found that Azad did not extend any of the ideas units from his pre writing discussion into more than one idea unit in his writing. The comparison of the z scores obtained from the analysis of his pre treatment and post treatment paragraphs showed overall negative gains; in other words, in his post treatment paragraph, Azad performance was weaker. However, Azad improved on some of the criteria considered for textual analysis: it seems that, that the opportunity to work with the pre writing discussion outcomes rephrasing a lready composed idea units helped Azad to improve his ability to handle more information in a single focus and to present more lexical information in a single clause.

PAGE 331

305 Answering the Research Questions How do peers participate in asynchronous CMC interactio ns? During the asynchronous interaction, Azad and his partner failed to create a common discussion space. They demonstrated dominant patterns of dyadic interaction: both of them posted their own views on the pictures and the topic and did not engage in dis 16 minutes attempting to use ICA2 conferencing function designed for ACMC for a synchronous exchange. This resulted in a higher number of postings, although they were shorte r, shorter as compared with other ACMC interactions; however, these postings preserved the general specifics of the ACMC postings: low number of non content idea units and longer and mainly complete sentences. How do they use the specific ideas and languag e generated during these interactions in their writings? pre writing discussion were lower than the average. I could not detect any intertextuality on organizational level: t he partners did not discuss explicitly this aspect of the task and the sequencing of ideas shared during the discussion differed from the idea sequence in the post connections with th e pre writing discussion when the comparison was performed based on idea units. All of the paragraph idea units showed some intertextual connections on idea unit level: Azad did not generate any new idea units during his first draft composition, although, in the interview he stated that he composed new ideas when asked

PAGE 332

306 level intertextuality: he borrowed verbatim and rephrased idea units. However, the consideration of his par most of the idea units he rephrased were his own. His rephrasing techniques were performed within the boundaries of the idea units. He also developed some idea units based on his own topics addressed in the discussion; the se idea units had low level intertextuality.

PAGE 333

307 Case Study Two: Shin Shin was a 38 year old female student from Korea. In her country, she worked as a nurse. Shin first enrolled at the ELI in Spring semester 2005 and, based on her placeme nt test results, she was enrolled in Level III. I met Shin during the Spring semester of 2005. During this semester, as the computer assisted language consultant at the ELI, I was responsible for the administration of a computer diagnostic skills test. The goal of the test was to identify the students whose low computer skills may impede their performance at the ELI. Shin was one of the students who needed additional help with her computer skills. The outcomes of the diagnostic skills test showed that, alth ough being comfortable using email, Shin needed to develop skills related to formatting, saving files to a particular location, and typing; in addition, her Internet search skills were weak. During the workshops she attended, I remember her to be rather sh y and reluctant when completing the assigned computer tasks. Shin took a break during the summer of 2005, and, based on the outcomes of the Fall 2005 placement test, she was assigned again to Level III. In her interview, Shin shared that she graduated fro m a nursing school in which the classes were organized in a We have two things practice and theory Appendix 10). Her perception of writing was that being a good writer is important for Very important for study her perception of writing and development of writing skills, Shin perceived vocabulary to be the key for the d evelopment of good writing skills; the difficulty with writing that she

PAGE 334

308 personally encountered was sentence word order. She stated that something that would be enjoyable for her when writing in English would be writing about her personal experiences, such as her trip (to Korea). In her interview, Shin shared that she felt more comfortable when using computers after studying for two semesters at the ELI. When comparing the two applications she used during the training period L.E.C.S. and ICA2 Shin stat ed that she preferred ICA2. She explained this preference in terms of her weaker typing skills which did not bother her as much when posting on ICA2. In addition, I would say that the interface of ICA2 and the way it functions are similar to email, a means of CMC communication with which Shin felt more comfortable. Further in her interview, Shin shared her perception of the importance of developing good computer skills to be related to her profession. s Level III section. Based on my observation notes, during the face to face classes, Shin was routinely quiet. She would never volunteer an answer during class discussion but would almost always have one when the teacher asked her directly. When working in a small group setting, she limited desire to collaborate with partners during the computer laboratory small group activities as well. I observed two computer labora tory triad sessions in which Shin participated. During these sessions, she would sit further from the computer allowing her peers to interact with each other and have access to the computer. She would be the quiet observer who would provide her opinion onl y if asked directly. Thus, her pattern of class participation and small group interaction might be identified as being very close to what

PAGE 335

309 interaction: Shin almost never asked questions or sought assistance; her contribution to discussions was very limited and was not made unless it was requested by peers or the teacher. The treatment interaction During the CMC treatment session, based on the random participant assignment Shin was assigned to an ACMC discussion dyad. Her peer was an experienced computer user. In the pre writing discussion, Shin and her partner demonstrated a dominant interaction pattern. Following is the full transcript of their interaction. Log 5.2 ACMC Pre writing Discussion: Shin and Partner Message 1 FROM: Peer (10/14/05 9:22 AM GMT 06:00) 1 HI, IM THERE, IM ALREADY LOGGED IN 2 Message 1 FROM: Shin (10/14/05 9:24 AM GMT 06:00) 3 Yes, I'M too. 4 Message 2 FROM: Shin (10/16/05 4:14 PM GMT 06:00) 5 I think that this is a my topic sentence; In the U.S, most people are living in the city 6 and people came from different countries and have different appearance. 7 Supporting ideas; 1.Most people live in city because they can get a lot of 8 advantage. 9 2.peopl e have different appearance European, Asian,Africian. 10 3.people experience many kinds of culture. 11 Message 2 FROM: Peer (10/16/05 4:56 PM GMT 06:00) 12 ok, here is what i think. for the topic sentence i think is better this: 13

PAGE 336

310 topic: now the people in the united states live as a friendly and united country with 14 all the ethnic groups without segregation. 15 for main idea: we should talk about the way of life of us people. 16 and supporting ideas: we can talk about all the cultures in US and the most 17 important cit ies, and also we can talk about the life in the suburbs and farms 18 Message 3 FROM: Peer (10/16/05 7:22 PM GMT 06:00) 19 according to the images we can talk about all the people in the whole country, we 20 cant talk about too about the lifestyle in suburbs a nd farms and also in big cities, 21 and also teh msot important caracteristics of this big cities, we cant talk about the 22 "melting pot" the compilation of all the cultures 23 Message 3 FROM: Shin (10/17/05 10:02 AM GMT 06:00) 24 Hi! Could you tell me more about "melting pot"? 25 Message 4 FROM: Shin (10/19/05 8:12 PM GMT 06:00) 26 You have a lot of idea. Thank you for your suggestion. Iwill write in my paragraph 27 such as population in theU.S,immigration,multiculture,urbanism 28 During their ACMC pre writin g discussion, Shin posted four messages and her partner three, which was slightly higher than the average number of turns for the ACMC group (7). Their contribution to the discussion in terms of idea units number was similar: Shin composed 15 and her partn er 18 idea units. They started the asynchronous session with two messages (lines 1 4) whose goal was to check whether both of them could access the discussion space. The next two messages (lines 5 22) were made two days later. Shin addressed the task in her message; she provided her topic sentence and supporting ideas. She did not invite her peer to post his contribution to the task nor did

PAGE 337

311 she pose a question to him. There was a hint of disagreement that her partner expressed in for the topic sentence i think is better this better sentence which Shin posted in her previous message. Further in the interaction, Shin did not respond to this part of the message. She asked a question (lines 23 24) in regard to the melting pot final message (line 25 27), she posted a very brief outline of her future paragraph which d both Shin and her partner demonstrated a dominant pattern of dyadic interaction: they both contributed to the task completion but were unable to engage with each other contributions. When sharing her experiences about the treatment task, Shin stated that she felt uncomfortable because of her computer skills. She also stated that in a face to face discussion, she could remain silent, if she wished but in a computer di scussion, she had to post. During the discussion, she used an electronic dictionary. Her weak typing skills affected negatively the discussion, another reason for her to prefer the face to face discussion to a CMC one. She felt that sharing information wit h her partner helped her to ideas. The paragraph: intertextual connections treatment paragraph written after the pre writing ACMC session showed negativ e gain as measured by the difference of the pulled pre and post treatment z scores: 1.218328443. In fact, she had the lowest gain among all participants in the study. The only criterion in which her post treatment

PAGE 338

312 writing improved was presentation and de velopment (refer to Appendix 11 for the scores). Further, I address the intertextual connections that were identified between the first paragraph draft composed by Shin and the pre writing discussion in which she participated writing discussion on a lexical level were medium to weak. The ratio of matching distinct lexical items of her paragraph and the pre writing discussion was 34.85, which was lower than the ave rage ACMC group ratio (41.77) and very close to the average ratio of the ACMC and the SCMC group combined (34.30). In the pre writing discussion, Shin and her partner discussed the paragraph content and organization. The intertextual connections between th is discussion and her paragraph were traceable but rather weak. Shin followed the general pattern of organization she stated in her final posting. There were four parts population, immigration, multiculture, urbanism. fo ur were traceable in her paragraph. However, they were reordered, and immigration appeared both in the beginning merged with population description (new ideas) and in the third part of the paragraph, where it was a part of the multiculturalism presentation The full text of the pre are presented in Comparison Matrix 5. 2.

PAGE 339

313 ACMC Matrix 5. 2 Paragraph Comparison: Shin graph Ideas (by Idea Unit) Message 1 HI, IM THERE, IU1** IM ALREADY LOGGED IN IU2** The U.S includes various race [ own rephrased IU6* (1) synonymic phrase, (2) syntax for cohesion ] and people from different countries IU1 IU5* [ own rephrased: synta x for cohesion, two IU*combined ] Population of the U.S is over 281 million. IU2 [ new topic ] The U.S the third most populous country in the world. IU3 [ new topic ] Develop of industry are caused urbanization. IU4 IU15* [ own extended part of IU15*] Therefore, country sides exist only some small towns and houses. Message 1 Yes, I'M too. IU1* Message 3 I think IU2* this is my topic sentence IU3*; In the U.S, most people are living in the city IU4* and people came from different countries IU5* and h ave different appearance IU6* Most people live in city IU7* because they can

PAGE 340

314 ACMC Matrix 5. 2 (Continued) Paragraph Comparison: Shin get a lot of advantage. IU8* people have different appearance European, Asian,Africian. IU9 people experience many kinds of culture IU10* IU5 IU15** [ New IU15** ] Also, when people live in the city, IU6 IU4* [ own rephrased verb tens e cohesion marker added ] they can get a lot of advantage. IU7 IU5* [ own exactly the same, shortened conjunction dropped] Another thing is IU8 [ New, added for cohesion ] the U.S people have more convenient from industry. IU9 [ New] They can use many kinds o f transportation IU10 [ New ] and rising standard of living. IU11 [ New] According to U.S people, the U.S history is very important in the Message 2 ok, here is what i think. IU3** for the topic sentence i think IU4** is better this: IU5** topic: IU6* now the people in the united

PAGE 341

315 ACMC Matrix 5. 2 (Continued) Paragraph Comparison: Shin states live as a friendly and united country with all the ethnic groups without segregation. IU7** for main idea: IU8** we should talk about the way of life of us people. IU9** and supporting ideas: IU10** we can talk about all the cultures in US and the most important cities, IU11** and also we can talk about t he life in the suburbs and farms IU13** word. IU12 [ New ] Many people want to come to U.S. IU13 IU5* [ Rephrased own content changes ] It is immigration. IU14 [ New, based on own IU5*] People from various countries what have different culture. IU15 IU10* [ Rephrased own, extended with more detail ] So, They respected other culture IU16 IU7** [ ] and make a new culture. IU17 IU18** [ New ] The U.S as a

PAGE 342

316 ACMC Matrix 5. 2 (Continued) Paragraph Compar ison: Shin Message 3 according to the images we can talk about all the people in the whole country, IU14** we cant talk about too about the lifestyle in suburbs and farms and also in big cities, IU15** and also teh msot important caracteristics of this big cities, IU17** we cant talk about the "melting pot" the compilation of all the cultures IU18** IU18** ] where various racial and ethnic groups have been combined into one culture. IU19 IU18** ] In conclusion, the U.S is made more powerful by various people and culture. IU20 [ New ]

PAGE 343

31 7 ACMC Matrix 5. 2 (Continued) Paragraph Comparison: Shin gs(by Idea Unit) Unit) Message 3 Hi! IU11* Could you tell me more about You have a lot of idea. IU13* Thank you for suggestion. IU14* Iwill write in my par agraph such as population in the U.S, immigration, multiculture, urbanism IU15*

PAGE 344

318 ACMC Matrix 5. 2 (C ontinued) Paragraph Comparison: Shin Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color o wn idea unit 4. Blue Highlight Color

PAGE 345

319 paragraph with the pre writing discussion were stronger in her part of the discussion. Out of the total 20 idea units she composed in her paragraph, one idea unit (IU7), 5% of the total idea units of her post message (IU5*), and five idea units (25%) were rephrased own discussion idea units (IU1, IU4, IU6, IU13, and IU15). Thus, high level intertextual connections existed writing discussion with her partner; however, she was mainly using her own idea units at this level of intertextuality. Shin used different strategies when rephrasing her idea units: (1) use of synonyms different appearance various race unit 1, and in IDEA UNIT6 [for example she incorporated idea unit 5* and [includes] people from different countries the idea units from the paragraph idea unit 4, idea unit15 [for example, Shin extended people experience many kind of culture people from various countries what have different culture discussion into one large r idea unit was applied [ people from different countries (IU1) is a combination of idea unit6* and idea unit5*]; (5) content change [ (IU5*) was changed to many people want to come to U.S discussion, Shin mainly applied changes within the borrowed idea unit boundaries. Only one paragraph idea unit presented an attempt to move beyond the idea unit boundaries.

PAGE 346

320 In addition to th e rephrasing of her own idea units used in the pre writing discussion, Shin composed 13 new idea units (65%). Five of them she based on three of IU19. Shin developed t wo topics suggested by her partner: life style in suburbs and farms melting pot compilation of cultures (IU7**) prese rving only its key topic: respect and culture. One paragraph idea unit vocabulary choice. Thus, on a medium level of intertextuality, Shin considered some of ideas. Shin composed eight new idea units which did not have a direct connection with any of the idea units from the discussion. For example, idea unit12 developed a new subordinate topic related to culture; idea unit20 was the concluding sentence of t he paragraph and actually summarized several ideas shared during the discussion. One of the new idea units (IU 8) was added for establishing paragraph cohesion. In addition, some of these new idea units (IU2 and IU3) provided new facts to support the topic of population. Intertextual connections: conclusions On a lexical level, the relationship between treatment paragraph and the pre writing discussion she had with her partner suggested lower than ACMC group average intertextual connection. On organization level, there were some weak but detectable connections with the pre writing discussion. The connection of her paragraph with the pre writing discussion on an idea unit level showed that she established mainly high and medium level intertextua l connections and considered almost all of her own idea units. She extensively rephrased all of her own idea

PAGE 347

321 units that she included into her paragraph, except for one. The strategies she used for rephrasing of her own ideas units were: (1) syntax changes, (2) use of synonymic phrases, contribution was evident at a medium level of intertextual connection: Shin composed new idea units which were based on topics developed by her partner. Thus, Shin s much weaker. While the majority of her paragraph idea units (12 60%) showed intertextual connections with the pre writing discussion, in her paragraph, Shin composed eight idea units (40%) that were new and did not show any traceable connections with t idea units sorted by intertextuality type.

PAGE 348

322 Table 5.3 Paragraph Idea Units: Shin Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 1 0 0 5 5 1 0 0 0 0 8 20 Paragraph % 5% 0% 0% 25% 25% 5% 0% 0% 0% 0% 40% 100% Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. New based on own idea unit 7. units. 8. 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 349

323 Answering the Research Questions How do peers participate in a synchronous CMC interactions? During the asynchronous interaction, Shin and her partner failed to create a common discussion space. They both contributed to the task but did not comment or How do they use the specific i deas and language generated during these interactions in their writings? pre writing discussion were lower than the average. There were some detectable connections on organiz stronger intertextual connections with the pre writing discussion when the comparison was performed based on idea units. Many of the paragraph idea units showed some intertextual connections on idea unit level. Shin was able to establish high level intertextual connections with idea units mainly generated by her during the pre writing discussion, while the medium level rephrasing techniques were performed within the boundaries of the idea units. Only in one idea unit did she combine two idea units from the discussion, showing an attempt to move beyond the idea unit boundaries. ACMC Actors, Low Gains: Comparison Particip When comparing the two cases, I found that the two students had different backgrounds: they were coming from different countries and cultures, they were at

PAGE 350

324 different stages of their lives both academically and professionally, and they ha d different computer exposure and experiences. However, there were a few similarities in their pre writing interactions and the texts they composed after this interaction. These similarities are highlighted and discussed below. Although their attitude to wards the task overall was positive, in the interviews, Azad and Shin expressed some concerns about the ACMC treatment task. Azad perceived it to be time consuming, while for Shin the task was difficult because of her typing skills; thus, she indirectly su ggested that the task was time consuming as well. One additional aspect of the ACMC discussion that made Shin uncomfortable: unlike in class discussion, she felt that she could not remain silent during the ACMC discussion. However, both participants stated that sharing information with their partners was helpful. In addition, Azad perceived the task to be helpful to improve spelling. The Pre writing Interaction There were some similarities and differences in their interactions, paragraph writing, and inter textual connections between the pre writing interaction text and the text of their first paragraph draft. Although Azad and Shin valued the opportunity to share information with their partner, neither was able to engage in a collaborative discussion and de monstrate dominant patterns of interaction. In their pre writing interactions they

PAGE 351

325 Post treatment Paragraph Gains and Intertextual Connections Azad and Shin demonstrated low writing gains in their paragraph drafts comp osed after ACMC discussion. They demonstrated positive gains on limited aspects of their post treatment paragraphs. Azad showed ability to use higher amount of information in a single focus (as measured by mean length of idea units) and higher lexical info rmation per clause (as measured through lexical density analysis). The only aspect of writing Shin improved was presentation and development of the topic (as measured using multiple trait rubric). The intertextual connection of their paragraphs and ACMC di scussions on a lexical level was lower than the ACMC group average. This suggests that their use of vocabulary generated during the discussion was limited. Further, on an organizational level, the intertextual connections were weak (for Shin) to non existe nt (for Azad). his own idea units from the pre idea units, thus, demonstrating a medium level intertextuality. She also composed newly developed ideas not discussed in the pre writing stage. Thus, t he intertextual connections discussion. The two participants also used similar paraphrasing strategies when incorporating idea units from the ACMC discussion into their w riting: (1) syntax changes, (2) idea unit clipping, (3) synonymous changes, (4) content changes. In addition to these paraphrasing

PAGE 352

326 techniques, Shin combined idea units from the discussion into larger idea units in the text. Shin also attempted to develop t hemes which were only briefly mentioned by her partner but were not developed in the discussion. Answering the Research Questions How do peers participate in asynchronous CMC interactions? These case studies suggest that learners of different ages and c omputer literacy levels may become engaged in a dominant type of interaction during an ACMC contribution. How do they use the specific ideas and language generated during these interactions in their writings? The two low gain participants showed strong high level intertextual connections, but they mainly rephrased their own ideas. The common rephrasing strategies they used when including discussion idea units into their paragraphs were: (1) use of synonyms, (2) syntactic changes, (3) idea unit clipping, and (4) content changes. establishing intertextuality on a medium level. (Ref er to Appendix 5.5: Ordered Comparison Matrix for a table presentation of the case comparison.) ACMC Actors, High Gains Case Studies Case Study Three: Felipa Felipa is one of the continuing students who participated in the study. She had started as a Level I student in the Summer of 2005, two semesters before she participated in the study. Although she was inconsistent in answering my

PAGE 353

327 questions from the asynchronous interview, I decided to use her as one of the actors in the instrumen tal case study. This decision was based on the fact that Felipa demonstrated the highest gain among all students who participated in the study as identified by comparing differences between the pulled z score of the post treatment and pre treatment writing s. The information that I was unable to obtain from the interview was partially which she participated during the SCMC and ACMC training sessions as a source of infor mation. Felipa is a 23 year old student from Colombia. During the past year, she resided in the United States with her husband. The spring semester of 2006 was her third semester at the ELI, and she was making good and steady progress in the development o f her language skills. I had the chance to observe Felipa during face to face and computer laboratory sessions. Based on these observations, I could conclude that she was a diligent learner who enjoyed group work and could work productively in a small face to face group environment. message posted on ICA2, Felipa had specific goals for the writing class as well as a clear understanding of her weak sides with regard with writing. ( The full text of this response is provided in Appendix 5.7.) In her response to her writing teacher, she wrote: That i want to know everything about writing. sometimes i am comfuse with the vocabulary and the order of the words. i want to know about how i can write in past and present progresive. and also i what to know about punctuation. to the interview questions, I could conclude that Felipa was an experienced computer

PAGE 354

328 user. She reported using computers routinely for communicating with her friends and for shopping. She expressed a very positive attitude towards computers and CMC in I like use the chat yes I love chat with my family after the class because it is only my free time for use CMC and this is very important for me because in this way you never are feeling alone Appendix 5.6 for the full text of the interview.) In addition, based on the SCMC peer interview students performed as a training task before the trea was that CMC could support her language learning,. During this activity, when asked by her peer if she thought that chat and email in English could help her to learn the yes i think it's very good b ecause we can to thing more about what we are going to write In the beginning of her studies at the ELI, she befriended two other students (one male and one female) from her Level I class. They formed a very tight group which stayed together during all three of the semesters. During the semester in which the study took place, Felipa and her two friends were enrolled in the same section of Level III. They were always together during breaks, and, whenever they had the choice, they worked in the same group during class assignments. Felipa and one of her friends were randomly assigned to the same dyad for the treatment. During the observed class sessions in which Felipa participated, I noticed that she was very open and did not hesitate to ask questions or re quest assistance from peers and the teacher. The treatment interaction The contribution to the pre writing discussion of Felipa and her partner was very similar. Each one posted one long message in which the entire task was addressed. Their contribution was similar in terms of quantity as well. The total

PAGE 355

329 content idea units in her pre writing interaction posting was 26. The organization of controller idea controlling idea suggested that she Felipa, who posted approximately five hours after her partner, acknowledged her was not support ed by her interpretation of the pictures which was different from that of Log 5.3: ACMC Pre writing Discussion: Felipa and Partner Message 1 FROM: Partner (02/19/06 4:29 AM GMT 06:00 ) 1 The people of the U.S. 2 Describing the pictures. 3 Picture 1: 4 Territory information describing the density of population distributed in all states of 5 America. Besides the rapport density/territory seems to be different if we analyze 6 the state of Alaska and the island of Hawaii. While the people of the biggest part of 7 America are uniform distributed accordingly in Hawaii Island, in Alaska the 8 majority density of the people is concentrated in the south. 9 Picture 2: 10 If we compare with European cities the struct ure of American cities has been as a 11

PAGE 356

330 result the opposite process to think about how to live individual life. While the 12 business center of the American cities has a tendency to look similar every where, 13 outside down town the rural parts of land keep the pec uliar characteristics of the 14 American real dream. People live in an own houses with garden far from the ring of 15 business work. 16 Picture 3: 17 Most of American cities are crossed by river or sea therefore the bridge is a good 18 idea to unify all part of the city and gives it sometimes an interesting engineering or 19 architectonic characteristic. 20 Picture 4: 21 This is the real fascinating skyline of the American cities. In a relative short space 22 is concentrate in height most of the workers of the U.S. The skyscraper is not only 23 a high building but became a symbol of the power of American business. 24 Picture 5: 25 The Unites states a nation of immigrants. As consequence of that from the 26 beginning the U.S. has been seen a mix of culture and the ethnic diversity is 27 probably the most important thing. 28 Picture 6: 29 Two principal cost in exam: the Atlantic cost and the Pacific cost in relation with a 30 principal most populated cities in the U.S. New York City is a symbol of the 31 Atlantic Ocean. Moreover Los Angeles and San Diego are two i mportant cities for 32 different reason situated on the Pacific Ocean. 33 Topic and controller idea: 34

PAGE 357

331 Even though the United States has composed by large variety of enormous ethnic 35 diversity it conserve own identity. 36 Body: 37 Supporting C.I. by example of what is re cognizable nation identity of America. 38 On the other hand what does make the real America today? 39 Example based from the description of the 6 pictures. In each one pictur has 40 described America. 41 Message 1 FROM: Felipa (02/19/06 9:14 PM GMT 06:00) 42 Hi Pa rtner I am agree with you. I thing the topic is how live the people in the U.S. 43 Description. 44 Picture # 1 45 In the first picture I can see how the US although a very large country, one can see 46 how people live throughout. The country equally; no matter if it to the South North 47 east or West and even in Alaska or the Hawaiian Islands. 48 Picture # 2 49 *In the second picture I can see one of the US largest cities, like San Francisco 50 Miami, Los Angeles Atlanta Dallas which clearly shows large skyscrapers as well 51 as ver y dense population. 52 Picture 3rd 53 We can se a small town which are plentiful around the US; with lots of small 54 buildings and houses where people live in a more peaceful way with less stress and 55 perhaps happier [8] 56 Picture 4th 57

PAGE 358

332 In the 4th picture we can se e what seems to be a mixture of the different from all 58 over the world marching towards a common goal. 59 P icture 5th 60 On the 5th picture we can see what appears to be the golden gate bridge in NY 61 City with the large skyscrapers in the background. 62 You can als o see some of the large boats which navigate the Hudson River as 63 everybody know NY is one of the larger metropolis in the world for business and 64 finances. 65 Picture 6th 66 In the last picture we can see the main land which stretches from Los Angeles to 67 NY and from Chicago to Houston. 68 We can also see that the US has the Pacific Ocean and the West and the Atlantic 69 Ocean in the East. 70 We can also observe the proximity of Cuba and Mexico to the south. 71 Topic and controller idea. 72 U.S is a large mass of land in which people from all over the world live and work 73 together. 74 Based on these observations, I could conclude that the two partners were unable interaction. It seems that Felip discussion took place: there was no evidence from the discussion posts that he read n also showing sign of task disengagement. Thus, both of

PAGE 359

333 them disengaged from the task once they completed their first posts without showing any intention for further discussion. In the interview, when asked about the pre writing task, Felipa stated that she had a positive opinion about the task; however, she admitted that she felt confused in the the pictures from the task]. Thus, I would suggest that the participa tion in the pre writing message exchange helped her to understand the task better. She also shared that, when r understanding of the task. The paragraph: intertextual connections In her post treatment paragraph draft, Felipa showed gains on six of the eight proposed text analysis parameters as measured by the difference of the pre and post treatment z scores: ( 1) the syntactic complexity (measured by calculating the mean length of t units), (2) the amount of information present in a single focus (measured by mean length of idea units), (3) lexical information per clause (measured through lexical density analysis ), (4) rhetorical soundness, (5) presentation and development of main ideas, (6) and overall language use (the last three criteria were assessed using a multiple trait rubric). She demonstrated a negative gain on the following two parameters: (1) the quant ity of overall information present (measured by the number of idea units) and (2) vocabulary complexity (measured by analyzing the frequency of the unique words used). Felipa had the highest gain of the students in the ACMC treatment group as well as of al l 60 participants in the study with a difference of the pulled pre and post treatment z scores being 1.783441878.

PAGE 360

334 writing discussion were weak the ratio of matching dis tinct lexical items of her paragraph and the pre writing discussion was 27.27, which was lower than the average ACMC group ratio (41.77) as well as the average ratio of the ACMC and the SCMC group combined (34.30). However, I decided to look at the ratio o f matching distinct lexical items of her paragraph and the pre writing discussion considering only her writing. This decision was based on the intertextual comparison on idea unit level which revealed that Felipa did not use any of the idea units of her pa rtner; thus, she considered pre writing posting and her first paragraph, it appeared that the intertextual connections at a lexical level were close to the medium of th e ACMC group and higher than the medium of the whole groups as a whole intertextuality at the lexical level as medium to strong rather than as a weak. In the pre writing postings, Felipa and her partner each proposed their own paragraph organization. Felipa did not follow hers, only, the topic sentence of her paragraph closely corresponded with the topic she proposed for the discussion. She used many of the content idea units she posted during the pre wri ting discussion but reordered them extending some of the topics with new support. However, there was an intertextual connection between her message organization and her paragraph which was not related to the sequence of ideas: she incorporated in her parag raph either rephrased or unchanged cohesion idea units used to organize her ACMC pre writing. Presented below is a detailed discussion of the intertextual connections on the idea unit level between the pre writing discussion in which Felipa participated an d her post treatment paragraph.

PAGE 361

335 writing by Felipa in her post treatment writing, it appears that she did not use any of his ideas. She concentrated mainly on rephrasing, developing, and supporting further her own writing treatment paragraph with analysi s comments.

PAGE 362

336 ACMC Matrix 5. 3 Paragraph Comparison: Felipa Partner Postings (by Idea Unit) Felipa Postings (by Idea Unit) Felipa Paragraph (by Idea Unit) Message 1: From Partner The people of the U.S. IU1 ** Describing the pictures.IU2** Picture 1:IU3** Te rritory information describing the density of population distributed in all states of America. IU4** Besides the rapport density/territory seems to be different IU5** if we analyze the state of Alaska and the island of Hawaii. IU6** While the people of the biggest part of America are uniform distributed Message 1: From Felipa Hi Partner I am agree with you. IU1* I thing the topic is how live the people in the U.S. IU2* Description.IU3* Picture # 1 IU4* In the first picture I can see how the US although a very large country, IU5* one can see how people live throughout. IU6* The country equally; no matter if it to the South North east or West and even in I am going to talk about life and people in the US. IU1 IU2* [ Rephrased own syntactic changes succes sful grammar correction ] In some parts we can see very big and larges countries for example Miami, Los Angeles Dallas with skyscrapers with a dense population, IU2 [ Rephrased own combines IU9* and IU10*] also we can see IU3 IU29*[ Rephrased own use of non content IU for organization ] how the

PAGE 363

337 ACMC Matrix 5. 3 (Continued) Paragraph Comparison: Felipa Partner Postings (by Idea Unit) Felipa Postings (by Idea Unit) Felipa Paragraph (by Idea Unit) accordingly in Hawaii Island, IU7** in Alaska the majority density of the people is concentrated in the south. IU8** Picture 2: IU9** If we compare with European cities IU9** the structure of American cities has been as a result the opposite process IU10** to think about how to live individual life. IU11** While the business center of the American cities has a tendency to look similar every where, IU12** outside down town the rural parts of land keep the Alaska or the Hawaiian Islands IU7* Picture # 2 IU8* In the second pictur e I can see one of the US largest cities, like San Francisco Miami, Los Angeles Atlanta Dallas IU9* which clearly shows large skyscrapers as well as very dense population IU10* Picture 3 rd IU11* We can se a small town IU12* which are plentiful around the US; IU13* [ New, based on own extends IU14* 15* ] from to be mixture with people of ethnic diversity from all over the world IU5 IU18* [ Verbatim own, only clipping for cohesion ] marching for the same goal to be free IU 6 IU19* [ Rephrased own, synony ms, extension ] and have pursuit of happiness IU7 [ New development of own IU15* ] and to be equal in the life with their families IU8 [New ] and work together. IU9 IU35* [ Verbatim own ] How we can find dense population and big

PAGE 364

338 ACMC Matrix 5. 3 (Continue d) Paragraph Comparison: Felipa Partner Postings (by Idea Unit) Felipa Postings (by Idea Unit) Felipa Paragraph (by Idea Unit) peculiar characteristics of the American real dream. IU13** People live in an own houses with garden far from the ring of busi ness work. IU14** Picture 3: IU15** Most of American cities are crossed by river or sea IU16** therefore the bridge is a good idea to unify all part of the city IU17** and gives it sometimes an interesting engineering or architectonic characteristic. IU18** with lots of small buildings and houses where people live in a more peaceful way with less stress IU14* and perhaps happier. IU15* Picture 4 th IU16* In the 4th picture we can see IU17* what seems to be a mixture of the different from all over the world I U18* marching towards a common goal. IU19* Picture 5 th IU20* office buildings in those cities. IU10 IU10* [ Rephrased own, synonyms, syntax ] We can find also IU11 IU29* [ Rephrased own, synonyms, text organization IU, cohesion ] how the people enjoy in small towns and the peaceful way with less stress IU12 IU12* & IU14* [ Rephrased own, synonyms, combines 2 IU ] and take care of their family and their gardens IU13 [ New support own IU 12 IU14* ] and perhaps happier IU14 IU15* [ Verbatim own ]

PAGE 365

339 ACMC Mat rix 5. 3 (Continued) Paragraph Comparison: Felipa Partner Postings (by Idea Unit) Felipa Postings (by Idea Unit) Felipa Paragraph (by Idea Unit) Picture 4:IU19**This is the real fascinating skyline of the American cities. IU20** In a relative short space is concentrate in height most of the workers of the U.S. IU21** The skyscraper is not only a high building IU22* but became a symbol of the power of American business. IU23** Picture 5: IU24** The Unites states a nation of immigrants. IU25** As consequence of that from the beginning the U.S. has been seen a mix of culture IU26** On the 5th picture we can see IU21* what appears to be the golden gate bridge in NY City with the large skyscrapers in the background. IU22* You can also see some of the large boats IU23* which navigate the Hudson River IU24* as everybody know NY is one of the larger metropolis in the world for business and finances. IU25* Picture 6 th IUI26* In the last picture we can see the also I can see IU15 IU29* [ Rephrased own, synonyms,text organization IU, cohesion ] how is the US powerful with the Pacific Ocean and the West and the Atlantic Ocean in the East. IU16 IU30* [ Rephrased own extends ] And conclusion I think IU17 [ New text organization IU, cohesion ] that United States has an en ormous control for a good welfare of the people. IU18 [ new ]

PAGE 366

340 ACMC Matrix 5. 3 (Continued) Paragraph Comparison: Felipa Partner Postings (by Idea Unit) Partner Postings (by Idea Unit) Partner Postings (by Idea Unit) and the ethnic diversity is probably the most important thing. IU27** main land IU27* which stretches from Los Angeles to NY and from Picture 6: IU28** Two principal cost in exam: IU29** the Atlantic cost and the Pacific cost in relation with a principal most populated cities in the U.S. New York City is a symbol of the Atlantic Ocean. IU30** Moreover Los Angeles and San Diego are two important cities for different reason IU31** situated on the Pacific Ocean. IU32** Topic and controller idea: IU33** Chicago to Houston. IU28* We can also see IU2 9* that the US has the Pacific Ocean and the West and the Atlantic Ocean in the East. IU30* We can also observe the proximity of Cuba and Mexico to the south. IU31* Topic and controller idea.IU32* U.S is a large mass of land IU33* in which people from all over the world live IU34* and work together. IU35*

PAGE 367

341 ACMC Matrix 5. 3 (Continued) Paragraph Comparison: Felipa Partner Postings (by Idea Unit) Partner Postings (by Idea Unit) Partner Postings (by Idea Unit) Even though the United States has composed by large variety of enormous ethnic diversity it conserve own identity. IU34** Body:IU35** Supporting C.I. by example of what is recognizable nation identity of America. IU36** On the other hand what does make the real America today? IUI37** Example based fr om the description of the 6 pictures. IU38** In each one pictur has described America. IU39**

PAGE 368

342 ACMC Matrix 5. 3 (Continued) Paragraph Comparison: Felipa Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highli ght Color own idea unit 4. Blue Highlight Color

PAGE 369

343 In her post contribution. Her paragraph was based either on her own idea units or on new idea units that were compose d to support topics already discussed during the pre writing stage and borrowed word for word from her ACMC pre writing message; this constituted 16.67% of her paragra ph (IU5, IU9, and IU14). The highest number of idea units in her paragraph, nine (50%), were a paraphrasing of her own idea units from the pre writing message (IU1, IU2, IU3, IU6, IU10, IU11, IU12, IU15, and IU16). Thus, 66.66% of her post treatment writin g revealed high level intertextual connections when idea units were considered in the analysis. She also composed six new idea units. One of them (5.56%) showed medium level intertextual connection with her postings: she based one idea unit (IU4) on her idea unit from the discussion. Another of the new idea units she composed in her post treatment paragraph showed low level of intertextuality: she developed a new idea unit (IU7) based on a topic she proposed by her in the discussion. Thus, two of her new idea units revealed connections with her contribution to the pre writing discussion, one at a medium level and one at a low level of intertextuality when idea units were used as a unit of analysis. Further, she composed four paragraph idea units (IU8, IU1 3, IU17, and IU18) which did not have intertextual connections with the pre writing discussion; these new idea units constituted 22.22% of her paragraph. Table 5.4 presents the paragraph idea units composed by Felipa sorted by intertextuality type.

PAGE 370

344 Table 5.4 Paragraph idea units: Felipa Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 0 3 0 9 0 1 0 0 1 0 4 18 Paragraph % 0% 16.66% 0% 50% 0% 5.56% 0% 0% 5.56% 0% 22.22% 100% Criteria Coding Sche me: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. New based on own idea unit 7. 8. 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 371

345 Felipa used five distinct strategies when rephrasing her idea units from her ACMC pre writing message: (1) synonymous changes, (2) syntax changes, (3) idea unit extension, (4) combining two units from the discussion into one, (5) successful grammar correction. The most frequently used one was applying synonymic changes to the idea unit before including it in the text. She rephrased six of h er paragraph text idea units, which constitute 33.33% of the entire paragraph: IU1, IU6, IU10, IU11, IU12, and IU15. We can find also We can also see how the people enjoy in small towns and the peaceful way with less stress, changes were applied on a phrasal level. The id ea unit from the ACMC discussion is with lots of small buildings and houses where people live in a more peaceful way with less stress All idea units to which Felipa applied synonymous changes but one (IU16) were changed syntactically as well, which constituted 27.78% of the idea units in her post treatment paragraph writing. These changes varied from word order only (for example in We can find also We can also see restructuring of the idea unit as it I am going to talk about life and people in the US I thing the topic is how live the people in the U.S. There is one interesting aspect with regard to the idea units she either rephrased or used unchanged: three of them (IU3, IU11, and IU15) were used in the pre writing message and in the paragraph for cohesion purposes.

PAGE 372

346 In addition to syntactic and synonymous changes, another rephrasing strategy Felipa applie d was idea unit extension. She developed further two idea units adding details within the same idea unit (IU6 and IU16). The last of the four rephrasing strategies that Felipa applied was combining two idea units into one. She constructed idea In some parts we can see very big and larges countries for example Miami, Los Angeles Dallas with skyscrapers with a dense population In the second picture I can see one of the US largest cities, like San Francisco Miami, Los Ang eles Atlanta Dallas which clearly shows large skyscrapers as well as very dense population In her topic sentence (IU1), Felipa was able to correct the word order of idea unit 2* and adjust the sentence syntax to the synonymous org anizational phrase she used. Intertextual connections: conclusions This student participated in the pre writing task as if it were an individual brainstorming activity. This affected the manner in which she utilized the ideas from the pre writing discussi on contribution to the discussion and considered only hers to be used to support the composition of her paragraph The ratio of matching distinct lexical items of her paragraph and the pre writing discussion was low whe posting was considered, the score was higher than the average score for the overall group and close to the medium of the ACMC group of participa nts. The intertextuality on organizational level was weak. The strongest intertextuality was revealed on an idea unit level. Most of the intertextual connections identified between her writing and her

PAGE 373

347 discussion posting when analyzed using idea units wer e of a high level: these were mainly rephrased idea units. The strategies she used most frequently for rephrasing of her pre writing idea units were synonymous changes and syntax changes; she also applied idea unit combination and extension. In many of the rephrased idea units, more than one rephrasing strategy was applied. She used fewer idea units from her ACMC message without any changes. Felipa did not follow the organization of her message when composing her paragraph. The only clear intertextual con nection on the organizational level was traced to the topic sentence of the post treatment paragraph: she rephrased the topic sentence proposed in her pre writing posting. She also used in her paragraph either rephrased or unchanged cohesion idea units use d to organize her ACMC pre writing. Answering the Research Questions How do peers participate in asynchronous CMC interactions? Felipa used dominant patterns of interaction during the pre writing stage. She and her partner did not engage in a discussion; rather, they posted one message each; of her reading this posting. Felipa did not address, question or develop any of her er of content idea units Felipa composed at the pre writing stage were comparable in number to the number of content idea units posted by her partner. How do they use the specific ideas and language generated during these interactions in their writings? writing discussion and her post treatment paragraph revealed low

PAGE 374

348 considered; however, when only her posting was accounted for the intertextuality level was post treatment paragraph revealed strong intertextuality: 77.88% of her paragraph was based on her pre writing discussion posting. These connections were of a high intertex tuality level; in other words, she either borrowed or rephrased idea units from the pre writing discussion when composing her paragraph. The inclusion of new idea units was limited. She developed some idea units on topics that were not mentioned during the pre writing stage. In addition, she transferred cohesion devices from the pre writing discussion and showed some ability to correct the grammar of her own posts. Case Study Four: Isabella Isabella was a 28 year old English languag e learner from Italy. She graduated from a university in Italy with a degree in marine biology. The semester in which the study took place, Fall 2005, was her first semester at the ELI. She was chosen as an actor of this instrumental case study because, ba sed on the comparison of differences between the pulled z score of the post treatment and pre treatment writings, she obtained the second highest writing gains in the ACMC group. When answering my ACMC interview questions, Isabella was always on time pro viding sufficient and detailed answers. Based on this interview and on my observations, I could conclude that she is not only an experienced computer user but also I became to use the computer only 2 years ago when I st arted to write my thesi's degree and I have to say that now I can't stop to use In addition, in her interview Isabella explicitly stated her awareness about the use of

PAGE 375

349 communic ate with some researchers who live in the other part of the world text of the interview with Isabella is available in Appendix 5.8.) Isabella was a language learner who could articulate her language learning goals and identify her difficulties In the interview, she shared that the importance of developing good writing skills is related to the communicative nature of writing. The development of grammar was the most challenging aspect of language learning in general The most difficult I have when I write in English is to rende in corret grammar what I want to say Isabella was a part of a Level III section that was not collaborative as a whole; some of the students in this section had personality conflicts which affected the overall dynamics of the group. However, Isabella was not directly affected by this personality conflict and was able to create good connections with most of her classmates. She took an active role in all of the face to face class and computer laboratory small group activities that I observed; in all cases she was the member of the group that led the task completion. Isabella was also active during class discussion, frequently contributing her opinion and ideas and not hesitant to ask her teache r clarification questions. The treatment interaction During the treatment interaction, Isabella posted only one long message in which she addressed all parts of the task: she provided a description of the pictures in the context of the task topic and mad e suggestions about the paragraph organization. The total number of the idea units of her message was 38, and the content idea units, in other words, idea units directly related to the discussed topic, were 21. Her partner posted after her. His message was shorter because, as he stated, he agreed with

PAGE 376

350 most of her paragraph content suggestions. Mostly, in his message, he added his ideas about paragraph organization. His message had a total of 28 idea units, 13 of which were related directly to the content of the discussed topic. Although Isabella recognized her being late with her posting (line 2), she became disengaged with the task and did not ssage despite his explicit request (line 46). In her message, Isabella expressed her ideas in an organized and methodical way and in a dominant manner. The dominant trend was mostly realized through her early disengagement. In her message, she showed poten tial for developing a collaborative discussion. The use of the can could would view on the t ask. The full text of the pre writing discussion is presented below. Log 5.4 ACMC Pre writing Discussion: Isabella and Partner FROM: Isabella (10/17/05 9:36 AM GMT 06:00) 1 Hello Partner sorry for my delay. 2 These are my ideas: 3 1 picture: describe the den sity in the 50th states if America: people in the state of 4 America are uniform distribuited and so in Haway island. In Alaska the majority 5 density of the people is in the south part of island that is the farest from Pole. 6 2 picture: a lot of the people liv e in the big cities, in high and essential skyscrapers 7 with a lot of levels. 8 3 picture: other people lives in comfortable and with few levels houses in the 9

PAGE 377

351 suburbs far from the frenzied live of the big cities. 10 4 picture: in the US the people have a lot of origins. They came from England, 11 Ireland, Scotland, Germany, Spain, Mexico, Ireland, France, China, Sweden, 12 Vietnam, Canada and Korea. 13 5 picture: a lot of cities are built near the sea and often long bridges link 2 different 14 part of these cities. 15 6 pictur e: the most densely populated cities are New York on the Atlantic sea and 16 Los Angeles and San Diego on the Pacific coast. 17 ORGANIZATION: 18 These 6th photo can be correlated so: 19 First of all I would put the idea of the first picture: people in the state of.. ... 20 second I would put the idea inthe 6th picture: the most densely.... 21 third I'd put the idea of 4th picture: in the US the people.... 22 fourth I'd put the idea of 2nd picture: a lot of people live.... 23 fifth I'd put the idea of 3th picture: other people liv e in comfortable... 24 sixth I'd put the idea of 5th picture: a lot of cities are built.... 25 What do you think???? 26 The topic sentence could be: The people of the us can be described by the 27 geographical distribution and origins. 28 Concluding idea: Mixture of race in spread territory ( in order to enlarge!!!!). 29 FROM: Partner (10/19/05 11:07 AM GMT 06:00) 30 Hi, Izabella. You have done a tremendous work. However, your picture's ideas are 31 almost the same except some little change. so I don't need to rewrite them aga in. 32

PAGE 378

352 the only difference is the organization of the ideas. Here how I would organize 33 them : 34 1. I would put the idea of the 4th picture telling about the race and origin of U.S 35 2. I will put the idea of the 1st picture : about density and geographical distr ibution. 36 3. ''''''''' ' ' ''''''''''''''''''''''''''''''6th picture : the most populous cities 37 4. " " " " 5th picture : telling about the situation of the construction and so 38 on. 39 5. " : " " " " " "2th picture telling about High level people. 40 6. " " " " ?": " "" 3th picture, telling about the people living in the surburb 41 According to me ,the topic sentence should be ""The people of U.S. con be 42 distributed by diferrent ways. 43 topic sentence wiil be : people in U.S. 44 contr olling ideas : different was to describe it. 45 Then we will have : ST1 '1st way is by race and Origin 46 ST2: description by density and geographical distribution 47 ST3: Living situation of the people. 48 so you tell me about what do you think. 49 Although Isabella sho wed a dominant pattern of interaction, this pattern was predetermined by her early disengagement from the task rather than by the tone or the structure of her message. In her answers to the interview questions related to the task, she I feel good because problem about the choice of the organization or what put in the paragraph. We were suddenly agree about changes to do in our paragraph and how proceed.

PAGE 379

353 that she read and considered it was clear from this answer that she did not perceive to have any disagreement or communication problems with her partner and she recognized and appreciated his contribution. However, once she p erceived that they had reached an agreement, she abandoned the discussion. I observed similar disengagement with other ACMC dyads that participated in the study; in many cases, it was the factor determining the dominant pattern of the interaction. When answering my interview questions about the treatment task, Isabella she h ad a good partner: partner helped me changing the organization of the paragraph and it was really helpful. I compared my suggestions and I review these in order to impruve the writting. It was really simple because gave me really good suggetions stating that in a text based environment it might be sometimes difficult to understand learning process. The paragraph: intertextual connections The post treatment writing of Isabella showed improvement on seven out of the eight text criteria considered for the study: (1) the amount of information present in a single focus (measured by mean length of idea units), (2) the quantity of overall information present (measured by the number of idea units) (3) lexical information per clause (measured through lex ical density analysis), (4)

PAGE 380

354 vocabulary complexity (measured by analyzing the frequency of the unique words used), (5) rhetorical soundness, (6) presentation and development of main ideas, (7) and overall language use (the last three criteria were assessed using a multiple trait rubric). The only aspect of her writing that did not improve after the treatment was the syntactic complexity (measured by calculating the mean length of t units). Her overall gain as measured with the difference between the pulled z scores of post treatment and pre treatment writing was 1.31275934. than the average of the ACMC group: the ratio of the matching distinct lexical items of her post treat ment paragraph and her pre writing discussion was 44.68 and the ACMC group average ratio 41.77. This suggests that Isabella used the pre writing discussion text to a higher degree to support her own post treatment writing. However, based on the fact that she posted more as well as on the further intertextual analysis on idea unit level, I could conclude that this higher ratio of distinct lexical items was a result of her higher contribution to the discussion as measured by idea units rather than the inclu sion of her The organization of her paragraph suggested strong influence of the discussion. She adopted the idea organization sequence proposed by her partner (IU17** 20**). In many cases, she used her own rephrased idea units from the discussion. However, she reordered them to follow the paragraph organization her partner suggested. Further, she extended it with her own topic ( Cities on the coast, IU33 38). The full text of the are presented in a comparison matrix along with my analysis comments below ( ACMC Paragraph Comparison Matrix 4: Isabella)

PAGE 381

355 ACMC Matrix 5. 4 Paragraph Comparison: Isabella Partner Postings (by Idea Unit) ph (by Idea Unit) (2) Message 1 FROM: Partner Hi,Isabella. IU1** You have done a tremendous work. IU2** However, your picture's ideas are almost the same except some little change. IU3** so I don't need to rewrite them again. IU4** the only differenc e is the organization of the ideas. IU5** Here how I would organize them :IU6** (1) Message 1 FROM: Isabella Hello These are my ideas: IU1* 1 picture: IU 2* describe the density in the 50th states if America: IU3* p eople in the state of America are uniform distribuited and so in Haway island. IU4* In Alaska the majority density of the people is in the south part of island IU5* that is the farest from Pole. IU6* The people in the United States can be represented by t hree ways; IU1 [ New develops own ideas, text organization cohesion purposes ] one of these is by origin and race; IU2 IU25**[ extended and synonym changes ] the other is by geographical distribution, IU3 IU10**[ rephrased an d extended part of IU syntax changes ] and the last one is by the living status. IU4 [ New develops own ideas, summary ] The people of the United States represent a melting pot of different cultures. IU5 [ New develops own ideas, summary ] It is a mixture of black, white, and Asian people. IU6 IU12* and IU8**

PAGE 382

356 ACMC Matrix 5. 4 (Continued) Paragraph Comparison: Isabella Partner Postings (by Idea Unit) 1. I would put the idea of the 4 th picture IU7** telling about the race and origin of U.S IU8** 2. I will put the idea of the 1st picture : IU9**about density and geographical distribution. IU10** 3. ''''''''' ' ' ''''''''''''''''''''''''''''''6th picture IU11**: the most populous ci ties IU12** 4. " " " " 5th picture IU13**: telling about the situation of the construction and so on. IU14** 2 picture: IU7* a lot of the people live in the big cities, in high and essential skyscrapers with a lot of levels. IU8* 3 picture: IU9 other people lives in comfortable and with few levels houses in the suburbs far from the frenzied live of the big cities. IU10* 4 picture: IU11* in the US the people have a lot of origins. IU12* They came from England, Ireland, Scotland, Germany, Spain, Mexico, [ New ] Most of the people who live in the US come from other country IU7 IU13* [ Rephrased own, develops one IU into 2 ] like England, Scotland, Germany, Spain, Africa, Mexico, Ireland, France, China, Sweden, Canada, Vietnam, and Korea. IU8 IU13* [ Rephrased own, develops one IU into 2 ] Most of the people of the US live in the fourth eight states of America IU9 [ general suggestion in IU26**] and here they are uniformly distributed. IU10 IU4* [ Rephrased Own, splits one IU into 2, successful grammar correction ]

PAGE 383

357 ACMC Matrix 5. 4 (Continued) Paragraph Comparison: Isabella Partner Postings (by Idea Unit) 5. " : " " " " " "2th picture IU15** telling about High level people. IU16** 6. " " " " ?": " "" 3th picture, IU17** telling about the people living in the surburb IU18** According to me IU19, the topic sentence should be IU20** ""The peopl e of U.S. con be distributed by diferrent ways. IU21** topic sentence wiil be : people in U.S. IU22** Ireland, France, China, Sweden, Vietnam, Canada and Korea. IU13* 5 picture: IU14* a lot of cities are built near the sea IU15* and often long bridges link 2 different part of these cities. IU16* 6 picture: IU 17* the most densely populated cities are New York on the Atlantic sea and Los Angeles and San Diego on the Pacific coast. IU18* ORGANIZATION: IU19* These 6th photo can be correlated so: IU20* same in Hawaii islands IU11 IU4* [ Rephrased own, develops part of own IU into full IU ] while in Alaska the majority of the people lives in the southern part of this island. IU12 IU5* [ Rephrased own, coherence device added, synonyms used, successful g rammar correction ] Maybe, the reason for this heterogeneous distribution is IU13 [ New cohesion] that the southern part is farther from the Pole IU14 IU6* [ Rephrased own extended, successful grammar correction ] and offers better environmental conditi on for living. IU15 [ New ]

PAGE 384

358 ACMC Matrix 5. 4 (Continued) Paragraph Comparison: Isabella Partner Postings (by Idea Unit) controlling ideas : different was to describe it. IU23** Then we will have : IU24** ST1 '1st way is by race and Origin IU25** ST2: description by density and geographical distribution IU26** ST3: Living situation of the people. IU27** so you tell me about what do you think.IU28** First of all I would put the idea o f the first picture: people in the state of..... IU21* second I would put the idea inthe 6th picture: IU22* the most densely.... IU23* third I'd put the idea of 4th picture: IU24 in the US the people.... IU25* fourth I'd put the idea of 2nd picture: IU26* a l ot of people live.... IU27* fifth I'd put the idea of 3th picture: IU28* other people live in comfortable... IU29* The most populated cities in the US are Now York IU 16 IU18* [ Rephrased own, clipped, separated into 2 IU ] that is on the East coast. IU17 IU18* [ Rephrased own, clipped, separated into 2 IU ] This city is the core of American economy and the culture. IU18 [ New] Here the people use to meet in fashionable night club and resorts IU19 [ New or visit di fferent museums. IU20 [ New suggested topic, IU27*] On the opposite coast there are Los Angeles and San Diego, IU21 IU18* [ Own rephrased: synonyms, syntax]

PAGE 385

359 ACMC Matrix 5. 4 (Continued) Paragraph Comparison: Isabella Partner Postin gs (by Idea Unit) sixth I'd put the idea of 5th picture: IU31* a lot of cities are built.... IU32* What do you think???? IU33* The topic sentence could be: IU34* The people of the us c an be described by the geographical distribution and origins. IU35* Concluding idea: IU36* Mixture of race in spread territory IU37* ( in order to enlarge!!!!). IU38* two really densely cities too. IU22 IU18* [ Own clipped, coherence added ] These cities are known for the ocean and beaches, mountains and natural parks, IU23 [ New ]along with city life energy that is unsurpassable. IU24 [ New ] Many people of the United States live in the big cities, in high and essential skyscrapers with a lot of levels. IU25 IU8* [ Rephrased own synonyms] Houston and Chicago are in the middle of the US IU26 [ New ] and are really different from each. IU27 [ New ] For example Chicago is characterized from breezy winter IU28 [ New ]

PAGE 386

360 ACMC Matrix 5. 4 (Continued) Paragraph Compari son: Isabella Partner Postings (by Idea Unit) while Houston has a mite weather during the while year. IU29 [ New ] Even though a lot of people prefer to live in the big cities IU30 [ New ] where it is possible to find amusements and offices close at hand, IU31 [ New ]other people live in one story house in suburbs far from the exciting life of the big cities. IU32 IU10* [ Rephrased own rephrased, synonyms, successful grammar correcti on ] Many cities are built near the coast IU33 IU15* [ Rephrased own synonyms ] and often long and impressive bridges link two different parts of these cities; IU34 IU16* [ Rephrased own extended, successful grammar correction ]

PAGE 387

361 ACMC Matrix 5. 4 (Cont inued) Paragraph Comparison: Isabella Partner Postings (by Idea Unit) from those bridges people can enjoy a wonderful sight of the coast. IU35 [ New ] So we can speak about of Americ a IU36 [ New cohesion ]like a big country that offer various spaces in which live IU37 [ New ] and where you can meet a lot of people from other countries. IU38[ New ] Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color own idea unit 4. Blue Highlight Color

PAGE 388

362 In her paragraph, Isabella composed 38 idea units. Most of them (15 or 39.48%) were new idea units which presented topics not addressed in the pre writing discussion. In addition, Isabella composed seven idea units which revealed low level intertextual connection. Three of these idea units (IU9, IU19, and IU20) developed discussion topics suggested by her partner, which was 7.89% of her paragraph text. Further, in three id ea units (IU1, IU4, and IU5) she developed her own topics, which constituted 7.89% of her paragraph text. One of the paragraph idea units, idea unit 6 (2.64%), developed a topic discussion. Thus, most of the new idea units revealed no intertextual connections. Isbella also established intertextual connections between her paragraph and the pre writing postings at a high level. She extensively used her own idea units from the d iscussion after rephrasing them; the total of such idea units in her paragraph was 14 in tertextuality level was rather weak. No idea units revealed medium level of type.

PAGE 389

363 Table 5.5 Paragraph Idea Units: Isabella Intertextuality High level Medium l evel Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 0 2 0 14 0 0 0 3 3 1 15 38 Paragraph % 0% 5.26% 0% 36.84% 0% 0% 0% 7.89% 7.89% 2.64% 39.48% 100% Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rep hrased 5. 6. New based on own idea unit 7. 8. 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 390

364 writing discussion, rephrased and included into the paragraph, Isabella used seven rephrasing strategies: (1) syntax changes, (2) synonymous changes, (3) idea unit extension of part or full idea unit from the discussion, (4) extension of full idea unit development of one discussion idea unit into two, (5) clipping of a discussion idea unit, (6) successful grammar correction, (7) addition of a c oherence device. In many of the idea units more than one rephrasing strategy was applied. The strategy that Isabella used most frequently was developing two paragraph idea units from one discussion idea unit; she composed six idea units using this strate gy (IU7 8, IU10 11, IU16 The most populated cities in the US are Now York IU16 that is on the East coast idea unit 18* from the discussion ( ork on the Atlantic sea and Los Angeles and San Diego on the Pacific coast number of idea units with synonymous changes applied was 7 (IU2, IU3, IU12, IU21, IU25, IU32, and IU33). An example of this synonymous change strategy is the following idea unit used in the pre In Alaska the majority density of the people is in the south part of island while in Alaska the majority of the people lives in the southern part of this island to be live. also an example of multiple strategies application. When rephra sing it, Isabella added a while This strategy was used by her one more time in idea unit 22. Isabella also successfully

PAGE 391

365 corrected the erroneous noun phrase from the discussi the majority density of the people the majority of the people the rephrasing of five idea units (IU10, IU13, IU14, IU32, and IU34). Isabella extended some of her discussion idea units within the id ea unit itself using the regrouping of phrases in the discussion and paragraph idea units. For example, In Alaska the majority density of the people is in the south part of island (IU5*) that is the farest from Pole while in Alaska the majority of the people lives in the southern part of this island. (IU12 IU5*) Maybe, the reason for this heterogeneous distribution is (IU13) that the southern (IU14 IU6*). This regrouping triggered the extension of idea unit 14 in the paragraph. The last two rephrasing strategies used by Isabella were clipping and syntax changes. An example of both these strategies could be found On the opposite coast there are Los Angeles and San Diego writing discussion the most densely populated citi es are New York on the Atlantic sea and Los Angeles and San Diego on the Pacific coast. opposite, ea unit keeping only the second part of the discussion idea unit. She also made a syntactic change by adding the existential predicate there are. Isabella composed a high number of new idea units. The 11 idea units that were composed to support and devel op an already presented topic or to summarize several

PAGE 392

366 topics presented in the discussion into one idea unit in the paragraph text were IU1, IU4, IU5, IU6, IU9, IU15, IU18, IU19, IU20, IU35, and IU36. She added new idea units to support paragraph cohesion ( IU13 and IU35). In the second part of her paragraph, she composed ten idea units that were not a direct support to a pre writing discussion idea unit incorporated into the text (IU23, IU24, IU26, IU27, IU28, IU29, IU30, IU37, and IU38). Intertextual conne ctions: conclusions Although Isabella applied a dominant She followed his advice on paragraph organization, and there was a clear connection between two of his idea un paragraph and the discussion postings were intertextually connected with her own idea units. In fact, in her paragraph, she used all of the ideas which she shared during the discussion. This suggests that the connection of the pre writing discussion and her paragraph was rather strong, which is further supported by the high ratio of matching distinct lexical ite ms. However, this strong connection is mainly with her portion of her contribution to the pre writing discussion. writing posting was higher than the average for the ACMC group. Wit h regard to organization, level of intertextuality Isabella used complex and extensive rephrasing, applying various techniques including syntax changes, using synonyms, merging, splitting and clipping

PAGE 393

367 was limited. Isabella composed a significant number of new paragraph idea units based on discussion topics, the majority of them being of a low interte xtuality level. She also composed a high number of new idea units based on topics not addressed during the pre writing stage; some of which served paragraph cohesion purposes. She demonstrated ability for successful correction of erroneous structures from the discussion. Answering the Research Questions How do peers participate in asynchronous CMC interactions? Isabella contributed significantly to the pre writing interaction. However, she demonstrated a dominant pattern during the interaction. Although t here was a potential for the pre disengagement from the discussion prevented such development. How do they use the specific ideas and language generated during these interactions in their writings? Isabella approached her paragraph writing with a collaborative set of mind intertextuality connections on the lexical level were strong. The organizational level intertextuality was also evident. The intertextual connections at an idea unit level were mainly at a high level of intertextuality and with connection to her own idea units. She also composed a significant number of new idea units that did not show inter textual connections with the pre writing discussion.

PAGE 394

368 ACMC Actors, High Gains: Comparison some similarities and differences as well. They both were appro ximately the same age. They both were able to articulate their difficulties with writing in English and their goals in terms of writing development. In addition, both Isabella and Felipa were experienced computer users who enjoyed using computers in their everyday life, including computer mediated communication. However, in her answers to the interview questions, Isabella demonstrated an awareness of the importance of the computers for academic purposes; to the interview questions. Pre writing Interaction During the pre writing discussion, both Isabella and Felipa demonstrated dominant patterns of interaction. Although at certain points of the completion of the pre writing task, both of them demonstrated one of them engaged into a discussion with her partner. This dominance transferred to the way they approached their post treatment writing as well: Felipa disregarded completely The attitudes to the pre writing discussion of both actors were positive. Isabella expressed satisfaction of the fact that she and her partner could reach an agreement. Felipa shared that in the pr ocess of working on her pre writing discussion posting the task this respect.

PAGE 395

369 There was no strongly pronounced difference when comparing their postings on idea unit level. What unifies the pre writing task completion of the two dyads has the considerable contribution of both Isabella and Felipa: they posted a high number of idea units and most of them were content idea units. In other words, these idea units developed the topic under discussion. Post treatment Gains and Intertextual Connections In their post treatment paragraphs, both Isabella and Felipa showed gains on multiple criteria. They both showed improvement in their ability to include more information in a s ingle focus, as measured by mean length of idea units. They showed higher degree of lexical information per clause as measured by lexical density analysis as well. The rhetorical soundness, topic presentation and development, and overall language use, as m easured using a multiple trait rubric, also improved. Thus, their higher gain was not a result of improving on any single text analysis criterion but rather on multiple criteria. The intertextual connections of their post treatment paragraphs on lexical le vel were different. Isabella showed a higher ratio use of distinct lexical items from the pre writing discussion as compared to Felipa. However, these results should be interpreted keeping in mind that the postings of both the actors and their partners wer e considered might be higher due to the fact that her partner posted significantly less than Isabella did, han Felipa. When the matching close to the ACMC group medium and higher than the average for the whole group of

PAGE 396

370 students who participated in the study. Thus, the di contribution affected the DLI ratio score of both Isabella and Felipa due to the fact that ACMC discussion was different in length I woul d suggest that the intertextuality on the lexical level for these high gain participants was medium to high within the ACMC group, higher for the overall group of participants, and the highest when compared to the other actors in the case studies. The inte rtextuality on organizational level was also different. While Isabella the in tertextuality on organizational level was only evident with the topic sentence. The rest of the ideas proposed in the pre writing posting and related to paragraph organization were not included into the paragraph. When analyzing the intertextual connection writing discussion and post treatment paragraph texts, it appears that they have several commonalities. They both considered mainly their own contributions to the discussion: idea units after considerable rephrasing. They used similar paraphrasing strategies including: (1) the use of synonyms and synony mic phrases, (2) syntactic changes within an idea unit; (3) extending idea units, (4) adding organizational words and phrases, some of them borrowed from the discussion, for paragraph cohesion, and (5) successful grammar correction, which was demonstrated by a higher extent by Isabella.

PAGE 397

371 techniques as well. Felipa combined two idea units from the discussion into one. Isabella, on the other hand, used parts of her own idea unit s from the discussion and developed them into full idea units in her paragraph or developed one idea unit from the discussion into two idea units in the paragraph. However, applying these strategies, both of them demonstrated an ability to move beyond the single idea unit consideration when borrowing them from the pre writing discussion. (Refer to Appendix 5.5: Ordered Comparison Matrix for a table presentation of the case comparison.) Answering the Research Questions How do peers participate in asynchronou s CMC interactions? The two ACMC actors, Felipa and Isabella, who had high gains, demonstrated dominant patterns of interaction during the pre writing discussion. The participants in the discussion posted their opinions and did not revisit the discussion a fterwards. I noticed that in both cases, there was an early disengagement from the discussion demonstrated by one of the peers. The two case studies of high gain ACMC actors suggest that, when completing the discussion using ACMC mode of communication, the actors may fail to contributions. These high gain actors were able to contribute approximately the same amount of information to the discussion, as measured with idea units. Ho w do they use the specific ideas and language generated during these interactions in their writings? The intertextual connections of the pre writing discussions and the post treatment hing DLI score

PAGE 398

372 of the two high gain ACMC participants was different. In terms of intertextuality on graph showed weak connection with her own proposed organization, while Isabella used consistently her At idea unit level of intertextuality Felipa and Isabella showed some commonalities. They both were able to estab lish connection of a high intertextuality level by either borrowing verbatim or rephrasing idea units. They mainly considered their own idea units when establishing this intertextuality. The paraphrasing techniques applied were similar and included changes within idea unit boundaries as well as merging and/or splitting idea units from the discussion when paraphrasing them in their paragraphs. ACMC Actors: Comparison The comparison of the four ACMC case studies presented above revealed that there were both commonalities and differences between the actors who demonstrated high and low gains in their post treatment writings. Below, these differences and commonalities are presented in the context of the research questions which the case studies aimed to answer : How do peers participate in asynchronous CMC interactions? How they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implementation of these ideas? The following t rends considered in the data analysis were compared: (1) background, (2) the pre writing CMC interactions specifics, and (3) post treatment paragraph gains and intertextual connections.

PAGE 399

373 The background of the act ors who demonstrated high and low gains in their post be related to their computer related experiences in language learning settings and their willingness to use computers Shin, who was the oldest one of the case study actors and the actor with the least computer experience shared that she felt uncomfortable when working with computers and during the task completion. I also noted this trend during my computer class observa tions. These observations were confirmed also by the analysis of data outside of the case studies, presented and discussed in Chapter IV. The other two students who experienced some difficulties during the training period were, similarly to Shin, of an age significantly higher than the average age of the intermediate level students who participated in the study and had limited or no computer experience prior to their enrollment to the intensive English language program. Further, both experienced and non exp erienced computer users may fail to benefit from the pre writing discussion as assessed by using the criteria proposed in the study. Azad, the other low gain actor, reported feeling comfortable when using computers but also demonstrated lower gains in his post treatment writing. The Pre writing Interaction With regard to the dyadic patterns of interaction demonstrated during the pre writing discussion, the comparative case study analysis showed that dominant patterns were employed by all case study actors who participated in the ACMC pre writing task regardless of their gain. However, all of them attempted at a certain point in their ACMC interaction to engage their partner in discussion of the assigned topic. These attempts

PAGE 400

374 failed, and none of the ACMC ca se study participants were able to create a common discussion space. There was a prominent difference between the pre writing discussions of the low ACMC participants and the ACMC participants with high gains in terms of their contribution to the discussi on. The participants who had low gains contributed more turns (Azad, 6 and Shin, 4), than the participants with high gains (Felipa, 1 and Isabella, 1). However, the number of idea units that the low gain ACMC actors posted during the discussion was signifi cantly lower than the number of idea units posted by the high gain actors: Shin posted 15 idea units, eight of which pertained to content, and Azad posted 10 idea units, eight of which were also content, while Felipa posted 38 total idea units with 26 cont ent idea units, and Isabella posted 35, of which 21 were content. Thus, it appears that, although there was no difference in dyadic interaction pattern of the students who had high and low gains, the contribution to the pre writing discussion of the high g ain actors was more significant. Post treatment Paragraph Gains and Intertextual Connections In their post treatment paragraphs, the high and low gain ACMC participants improved on different criteria of their writing. The high gain participants showed ga ins on multiple criteria. They improved their ability to include more information in a single focus, as measured by mean length of idea units and they showed higher degree of lexical information per clause as measured by lexical density analysis. The rheto rical soundness, topic presentation and development, and overall language use, as measured using a multiple trait rubric, were also improved. Thus, the higher gain was not a result of the improvement of a single criterion but rather an improvement of multi ple criteria. On the

PAGE 401

375 other hand, the low gain participants improved on fewer aspects of their writing: Shin improved only the presentation and development of her paragraph, and Azad showed ability to use a higher amount of information in a single focus (as measured by mean length of idea units) and higher lexical information per clause (as measured through lexical density analysis). Table 5.6 presents the textual aspects improved by the ACMC actors. Table 5.6 ACMC Actors: textual aspects improved Tex Analys is Criteria Low ACMC High ACMC Azad Shin Felipa Isabella Syntactic complexity (T units mean length) X Information in a single focus (mean length of idea units) X X X Overall information present (number of idea units) X Lexical information per clause (lexical density) X X X Vocabulary complexity (frequency of unique words) X Rhetorical soundness X X Presentation and development of main ideas X X Overall language use X X

PAGE 402

376 With regards to intertextual connections of the pre writing discussion and the paragraphs written based on this discussion, it appeared that the actors who participated in a CMC pre writing discussion and had low gains, demonstrated mainly weak intertextual connections when these connections were measured on a lexi cal level. The two high gain actors had different results. One of the high gain participants, Isabella, wrote a paragraph that had high intertextual connections on a lexical level, while the other, Felipa, demonstrated low intertextuality on a lexical leve l in her paragraph when the whole ACMC interaction was accounted for; however, the lexical intertextuality was medium to high, with her own posting. Keeping in mind that both actors used in their paragraphs mainly their own topics and idea units from the d iscussion, this difference discussion. On the organizational level, there was no consistent difference between the high and the low gain actors: one member of each gain dyad did not have any intertextual connections on an organizational level and the other did. However, the intertextual connections on organizational level, when they existed, were rather loose. When idea units were considered as a unit of intertextuality a nalysis, it appeared that the ACMC actors established intertextual connections on all four levels. The idea unit tabulation by type and level of intertextuality for the four ACMC actors is presented in Table 5.7.

PAGE 403

377 Table 5.7 ACMC Actors: Paragraph Idea Unit s and the Level of Intertextual Connections Participant and Level Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Low ACMC Azad Number of IUs 1 0 2 5 0 0 0 0 3 0 0 11 Paragraph % 9.1 0 18.18 45.45 0 0 0 0 27.27% 0 0 100 Intertextuality High: 72.73% Medium: 0 % Low: 27.27% None: 0 100 Shin Number of IUs 1 0 0 5 5 1 0 0 0 0 8 20 Paragraph % 5 0 0 25 25 5 0 0 0 0 40 100 Intertextuality High: 30% Medium: 30% Low: 0 None: 40% 100

PAGE 404

378 Table 5.7 (Continued) ACMC Actors: Paragraph Idea Units and the Level of Int ertextual Connections Participant and Level Criteria 1 2 3 4 5 6 7 8 9 10 11 Total High ACMC Felipa Number of IUs 0 3 0 9 0 1 0 0 1 0 4 Paragraph % 0 16.66 0 50 0 5.56 0 0 5.56 0 22.22 100 Intertextuality High: 66.16% Medium: 5.56% Low: 5.56% None : 22.22% 100 Isabella Number of IUs 0 0 2 14 0 0 0 3 3 1 15 Paragraph % 0 0 5.26 36.84 0 0 0 7.89 7.89 2.64 39.48 100 Intertextuality High: 42.1% Medium: 0% Low: 18.42% None: 39.48% 100

PAGE 405

379 Table 5. 7 (Continued) SCMC Actors: Paragraph Idea Units an d the Level of Intertextual Connections Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. New based on own idea unit 7. 8. New dev 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 406

380 When approaching the intertextual connections through the comparison of th e idea units from the discussion and the following writing, it appeared that all of the actors considered their discussion when composing their paragraphs. When analyzing the data, I could not find consistent differences in this respect between the low gai n and the high the pre writing discussion (100%). In his paragraph, most of his idea units were either used verbatim or were paraphrased idea units from the discussi on; thus, most of the were either of high or medium level of intertextuality, and the rest were new. Felipa composed paragraph idea units that demonstrated intertextu al connections on all three levels with the majority of the idea units being of medium intertextuality with the pre intertextuality level (42.1%) and some were of low i ntertextuality level (18.42%). All of the participants considered mainly their own ideas when composing their post treatment unit word for word. All four participants employed multiple paraphrasing strategies when intergrating the low gain ACMC participants used when including discussion i dea units into their paragraphs were: (1) use of synonyms; (2) syntactic changes; (3) idea unit clipping, and (4) content changes. The high gain ACMC participants used the following paraphrasing strategies: (1) the use of synonyms and synonymic phrases, (2 ) syntactic changes within an idea unit, (3) altering the boundaries of the idea units through extending, clipping, or

PAGE 407

381 combining two idea units into on, (4) adding organizational words and phrases, some of them borrowed from the discussion, for paragraph c ohesion, and (5) successful grammar correction, which was demonstrated by a higher extent by Isabella. Thus, the difference between the high and the low gain participants were that the high gain ones showed the ability to move beyond the single idea unit consideration when incorporating the pre writing ideas into their writings, while this was not evident in the writings of the low gain actors. Further, the high gain participants were also able to transfer cohesion devices and to perform successful error c orrections to erroneous phrases from the discussion when including them into their paragraphs. The two high gain participants tended to include more new idea units in their post treatment writing. These new idea units were of several different types in t erms of their relation to the pre writing discussions: (1) further development of topic stated in their pre writing discussion and transferred into their writings, (2) new idea units added for text cohesion, (3) new idea units developing the topic of the p aragraph but not mentioned in the discussion, (4) new idea units summarizing several points made during the discussion (this was used only by Isabella), (5) the inclusion of new idea units was limited or non existent in the paragraphs of the low gain ACMC participants. Answering the Research Questions How do peers participate in asynchronous CMC interactions? I found that the ACMC participants did not differ in terms of their pre writing dyadic interaction. Regardless of their post treatment gains, they ap plied dominant patterns during the pre writing discussion. However, there was a difference in their

PAGE 408

382 contribution to the discussion. The high gain participants contributed more to the discussion than the low gain participants. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implementation of these ideas? paragraphs and pre wri ting discussions was of various levels which were not related to achievement level. There were some intertextual connections of the pre writing discussions and the paragraphs of two of the actors; however, these connections were rather loose. Nevertheless, the intertextual analysis on idea unit level revealed stronger intertextuality. Both high gain and low gain participants composed 60% or higher of their paragraphs considering specific idea units from the discussion or topics addressed during the discussi on. However, these were mainly their ideas units or topics, and a limited interaction affected their writing. The ACMC participants were limited in their consideration and All participants applied various paraphrasing techniques; however, the high gain participants were able to regroup idea units from their pre writing postings, perform successful grammar correction, and i ncorporate cohesive devises while the low gain participants were not able to do this. In addition, the high gain participants attempted to develop more new ideas in their post treatment writing. Lastly, the advance of the high s not determined by improving significantly on one of the

PAGE 409

383 considered text analysis criterion but rather on a group of criteria; the low gain participants improved their writing either on one or two criteria. Further in this chapter, four case studies of st udents who participated in the pre writing discussion in a synchronous environment are presented. Their pre writing interactions and the intertextual relations of their post treatment texts were described and compared based on the post treatment gains. The concluding section of the chapter presents the overall comparison of the eight instrumental case studies. (Refer to Appendix 5.5: Ordered Comparison Matrix for a table presentation of the case comparison.) SCMC Actors, Low Gains Case Studies Case Study Fi ve: Ajwad Ajwad was part of a large group of Saudi Arabian students who came to the ELI as part of a Saudi government program. He was one of the older students in the group. At the age of 27, he had completed a university program in Saud i interview I had with Ajwad, he also volunteered for a face to face interview answering some of the questions which he did not have the opportunity to answer during th e ACMC interview. I observed him in two computer laboratory sessions as well as in two face to observations and on his answers to my interview questions provided both in the ACMC and the face to face interviews. Based on my observations, Ajwad was a respected member of the Saudi Arabian student group. These students, especially in the beginning of the semester, stayed close

PAGE 410

384 together. In most of the instances that I observed him du ring small group activities, he was the leading group member that would often take charge of the task completion and would direct other members of the group during the discussions making sure that they did not deviate from the task. In some of the cases, h e would take over the discussion in order to clarify his view and convince the group members about his opinion. During class discussions, he preferred to interact directly with the teacher and to be actively involved in the interaction during class discus sions. When he was not personally involved in a task completion or discussion he would lose interest and get distracted. This usually happened during teacher presentations that were not interesting for him or during which he w as not directly addressed by the teacher. On several occasions, I noticed him to look over the shoulder of a student engaged in something not related to class discussion, to drawing figures in his textbook or to express his disinterest through body languag e looking around and tapping his fingers on the desk. During the face to face interview I had with Ajwad, he shared his perceptions of class organization at the ELI being too rigid leaving little space for creativity and personal involvement. Thus, he de monstrated a higher motivation to contribute to discussions when he was in the center of it and was able to share his personal opinion on the topic under discussion. This desire to contribute his personal opinion was evident in his view about the nature of writing which he shared in the interview with me. The full text of the ACMC interview along with my notes from the face to face interview is provided in Appendix 5.9. prominent: writin g as an important means of communication with other people. More

PAGE 411

385 specifically, Ajwad emphasized writing as an important means of communication of his the people I focus to them, so if I am a good writer and I have good idea the people will resave my idea repeated again in his answer about the importance developing writing skills in English: It is very imp ortant to me, because if I write in any subject even if I tack deferent position I will found many writer wrote in the same subject so I need in that time to be unique in that subject, In idea, in the steppes to present my subject, make my subject easy to When sharing the difficulties he experienced when writing in English, Ajwad again stressed communication. Along with spelling, he identified the formulation of his ideas in English as being a challenge for him. Communication of his ideas to other people when writing in English was also the theme which emerged in his answer about the enjoyable things related to writing in If I can delivery my idea to the people, and I can effect to the people so I make change. During the interview, Ajwad shared that he had a considerable experience with computers. He started using computers when he was 13 year old, when he enrolled in his first computer class. In his university course of studies, computers played an i mportant In the first class in my field, the chairmen of industrial engineering was the instructors, and I remember what he said about using ve computer in his home, he can transfers to another department, and we are welcome to help any student to bay computer Ajwad gave a general answer about his CMC experiences. He answered

PAGE 412

386 to conta ct or take appointment computer in an educational setting. I found further confirmation of this in his perception of the CMC pre writing treatment task. The treatment interaction Aj wad shared that the CMC interaction provided him in the static position It was something new for me, to see the word in the static positions, and some time I need to stand out of habit, just to cut the rotten. Also it give me good chance to listen to the people in deferent ways so at least I can look to the picture from deferent ways Ajwad did not answer some of the questions related to the treatment CMC task used in the study during the ACMC interview. However, he was willing to talk to me providing some of these answers in a face to face interview. He told me that he liked the chat better than the ACMC inter action training tasks. However, he did not like the interface of L.E.C.S. In his opinion, chat could support the development of better writing skills. His experience was that chat helped him to organize his ideas in a better way, and he confirmed again tha concerns with regards to the chat. When asked if typing hindered his discussion online, Absolutely yes, because when I speak with the people we have million and million idea to di scus, and we need to speak continuously and quickly to represent all these idea, but in the computer I thing it is limited he perceived his SCMC discussion to be less productive than a face to face discussion. Followin

PAGE 413

387 Log 5.5 SCMC Pre writing Discussion: Ajwad and Partner Partner: youhouuuuuuuuuuu 1 Partner: Is someone here? 2 Ajwad: hi 3 Ajwad: her i am 4 Partner: haaaaaaaaaa 5 Partner: we will discus s about ideas first 6 Ajwad: ok 7 Partner: I think that we have to talk about ethnical diversity, obesity, caractere, 8 way of life 9 Ajwad: O.K 10 Partner: do you have other ideas? 11 Ajwad: NO IT IS OK 12 Partner: ok.... so we can go on 13 Partner: what about t he paragraph organization? 14 Ajwad: defrent poeple in the same blase coming from all the wored 15 Ajwad: we will talk about Amireca in general 16 Partner: ???? 17 Ajwad: what is going on? 18 Partner: we have to choose a topic sentance 19 Ajwad: ok 20 Partner: for example: People in America are very different 21

PAGE 414

388 Ajwad: what is your sudgesting 22 Ajwad: i like this idia 23 Ajwad: ok then we write about the resons 24 Partner: let's choose a main idea 25 Partner: ok 26 Partner: immigration 27 Ajwad: yes 28 Ajwad: why the peop el come to usa 29 Partner: porety vs. rich (rich are often white and poors are often black or come 30 from the minorities) 31 Ajwad: what the advanteg they well get fron usa 32 Partner: yes 33 Ajwad: but i like to spike in general 34 Partner: to spike? what does i t mean? 35 Ajwad: the peopel coming to USA FROM ALL THE WORD, CUS THEY GET 36 GOOD OFFER IN WORK, STUDY, GO TO HOSPITEL 37 Ajwad: somthing like that 38 Partner: and for the american dream 39 Partner: we can atalk about what we see in our culture class 40 Ajwad: we can also start with the history of immegrint 41 Ajwad: AS WE STUDY IN BILL CLASS 42 Partner: YES!!!!! 43 Partner: :) 44

PAGE 415

389 Ajwad: SO OUR PRAGRAPH WELL BE : 45 Partner: hoooooooo it's just a brainstorming today! great! we don't have to write a 46 paragraph 47 Ajwad: 1. HISTORY OF IMMEGRINT 2. WHOME? 3. WHAT IS THE 48 ADVANTEG FRO IMMIGRINT 49 Partner: so actually.... we can talk about immigration. Ameri 50 Partner: why is the contry so diversified 51 Ajwad: diversified? 52 Partner: do you understand all teh pictures? 53 Ajwad: YES 54 Partner: one is about the number of people, another is about immigration, an other 55 one shows rich's and poor's houses 56 Ajwad: what are u doing? can u tell me 57 Partner: but I don't figure out the 3 others 58 Partner: I'm chatting! 59 Ajwad: i think # 3 ABOUT 60 Ajwad: THE OLD HUES 61 Partner: the old hues? 62 Ajwad: AND NUMBER 4 ABOUT UNITED OF PEOPEL IN USA, CUSE THEY 63 WAS FROM DEFRINT COUNTRY AND NOW THEY FROM USA 64 Ajwad: #5 ABUT THE BILDING TODAY 65 Ajwad: MAP FOR USA TODAY 66 Partner: the fifth one is N Y city 67

PAGE 416

390 Partner: yes I can see the pictures!thank you! 68 Partner: ;) 69 Partner: But I try to understand the rapport with the People in USA 70 Ajwad: YES BUT I THINK IT IS TO SHOW THE BULDING TODAY. AND # 3 71 FOR THE BELDING YASTERDAY 72 Partner: hoooh.... I w as thinking that the number 3 was showing bulding of poor 73 and rich people 74 Ajwad: HE U MAM WHERE ARE U 75 Ajwad: ????????????////// 76 Partner: anyway 77 Ajwad: WHAT? 78 Partner: We don't talk about architecture 79 Partner: so we don't really need this picture 80 Ajwad: WHY? 81 Partner: because our topic is about people 82 Ajwad: OK 83 Partner: Maybe we can say that a lot of foreign people who come in the USA live 84 in big city 85 Partner: I think that there is a few foreign people in Kansas 86 Partner: People prefer g oing in California, NY, Florida, Texas, Chicago 87 Ajwad: YES U GARED 88 Ajwad: BUT POEPEL ALSO PRERER TO TAMPA ALSO 89 Ajwad: DONT MISS THAT 90

PAGE 417

391 Partner: we finish 91 Partner: we prefer tampa 92 The overall patterns that I identified when analyzing the discussio n of Ajwad and and were able to move through the discussion resolving misunderstanding and constructing a common view on the topic using the pictures provided to sup port the process of interpreting the pictures and sharing ideas related to the task topic. The overall tone of their discussion was light and relaxed. However, I no ticed that this tone was set by the partner rather than by Ajwad. At certain moments during the discussion, especially when feeling insecure, Ajwad showed hints of irritability. In order ked questions such as what are u doing? Can u tell me HE U MAM WHERE ARE U posted multiple question marks and backward slashes (lines 75 and 76). At some point of the discussion (line 36), after his partner indicated misunderstan ding, he switched to caps locked font, which I interpreted to be a strategy to stress his opinion and attempt to make it clearer (similar to a louder, more articulated speech in a face to face interaction). He used caps lock in the beginning of the discuss ion as well but switched to a normal font after two short postings (lines 10 and 12). In contrast, his partner was able to preserve a calm and positive attitude during the discussion. She used emoticon of a smiley face to stress her agreement with Ajwad (l ine 44), implicitly invited the partner to interpret the pictures under discussion rather than simply state their topics and then supported her statement with an emoticon of a smiley face to indicate her friendly tone (lines 68 and

PAGE 418

392 69). In my opinion, her positive attitude helped the discussion to move on in a productive and collaborative manner. was the one who would perform more moves aiming to organize and direct the dis cussion. I found seven such postings (lines 6, 13, 19, 25, 46 47, 50, 79). She was also the one to initiate and conclude the task. Thus, although taking a collaborative stand uns uccessfully attempted to direct the discussion to a more general chat on the topic (lines 16 and 34); he successfully initiated the paragraph organization discussion which helped the dyad to specify the overall theme of their paragraph (lines 48 49). Des pite the leading role of his partner, Ajwad was able to make a significant contribution to the discussion. Both partners made a similar number of turns during the chat (Ajwad 39, partner 43). Further, Ajwad contributed a slightly higher number of content i dea units addressing specifically the task content: 20 out of the total 50 idea units that he composed during the discussion. His partner posted 19 content idea units out of 57 total idea units. The SCMC Paragraph Comparison Matrix 5.5 : Ajwad included in t he next section presents their chat with idea unit segmentation. The dynamics of this SCMC interaction was similar to the dynamics of the other SCMC interactions observed during the treatment and described in Chapter IV. Most of the postings of Ajawad an d his partner were short, containing mainly one or two idea units per posting; the highest number of idea units in one posting was four. The paragraph: intertextual connections In his post treatment paragraph, Ajwad showed negative gain as measured by th e overall difference of the pulled pre and post

PAGE 419

393 treatment z scores. He improved on two of the eight text analysis criteria used in the study: the amount of information present in a single focus (measured by mean length of idea units) and lexical informat ion per clause (measured through lexical density analysis). He had the lowest gain of the students who participated in the SCMC treatment group. writing chat on a lexical level as measured by the ratio of matching distinct lexical items of his paragraph and the pre writing discussion was 20.87. Comparing it to the mean ratio of the SCMC treatment group 26.31, it was low. Further, I present my analysis of the intertextual connections of s paragraph and his pre writing discussion on organizational and idea unit level. Ajwad proposed an organization sequence of discussed ideas during the pre writing chat: (1) history of immigration, (2) homes, and (3) advantages for immigrants (Pre writing Discussion SCMC Log 5.5: Ajwad and Partner, lines 46 47). He was less question mark. Further in the discussion, the partners agreed to abandon the idea of different types of houses and concentrate their paragraphs on people. This suggestion was topics of immigration history and the advantages of living in the United States. The the reasons for this diversity, which were suggested by his partner (IU13** and IU 31**). Thus, I would suggest that, in his post treatment writing, Ajwad followed the paragraph organizat ion suggestions shared during the discussion demonstrating a considerable intertextual connection of the pre writing task and the post treatment paragraph on

PAGE 420

394 writing d

PAGE 421

395 SCMC Matrix 5.5 Paragraph Comparison: Ajwad Partner: youhouuuuuuuuuuu IU1** Partner: Is someone he re? IU2** In 1815, the people in the United States was 8.4 millions, IU1 IU21* & IU33**[ New IU, ] after that many people from all of the word immigrants in the late of 1800s and 1900s, IU2 24* IU16** [ New IU topic ] they can found better life in USA, IU3 IU18* and IU19* [ New IU, summary of own topic ] more chance to get a job, better life and study. IU4 IU18* [ Own, rephrased takes part of Ajwad: hi IU1* Ajwad: h er i am IU2* Partner: haaaaaaaaaa IU3** Partner: we will discuss about ideas first IU4** Ajwad: ok IU3* Partner: I think IU5** that we have to talk about ethnical diversity, obesity, caractere, way of life IU6** Ajwad: O.K IU4* Partner: do you have other ideas? IU7** Ajwad: NO IT IS OK IU5*

PAGE 422

396 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Partner: ok.... IU8** so we can go on IU9** Partner: what about the paragraph organization? IU10** own IU and develops into a new, use of synonyms ] All of those ideas may be able to found in every where, IU5 [ New ] but the different is more chance and equal opportunity. IU6 [ New ] During t he last hundred years, the United States totally change, IU7 [ New, development of own topic ] the good system was billed, IU8 [ New ] the building improved, IU9 IU40* [ New, based on own topic ] the population increase more then 32 times, IU10 Ajwad: def rent poeple in the same blase coming from all the wored IU6* Ajwad: we will talk about Amireca in general IU7* Partner: ???? IU11** Ajwad: what is going on? IU8* Partner: we have to choose a topic sentance IU12** Ajwad: ok IU9*

PAGE 423

397 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Unit) Partner: for example: People in America are very different IU13** IU21* [ New, development of own top ic ] there are different type of people IU11 IU13** [ rephrased, syntactic changes ] but they dialing with them same. IU12 [ New ] The United states have around 280,000,000, IU13 IU33** [ New, ] and no way to be the people ha ve the same behaviors or same cutler, IU14 IU13** [ topic ] how come if all states have different whether environments, Ajwad: : what is your sudgesting IU10* Ajwad: i like this idia IU11* Ajwad: ok then we write about the res ons IU12* Partner: let's choose a main idea IU14** Partner: ok IU15** Partner: immigration IU16** Partner: porety vs. rich IU17** (rich are often white IU18** and poors are often black or come from the minorities) IU19** Ajwad: yes IU13* Ajwad: why the peopel come to usa IU14*

PAGE 424

398 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Ajwad: what the advanteg they well get fron usa IU15* IU15 [ New] and the distance from the south to the north or east to the west it take like more then 5 days by car. IU16 [ New ]The United state now is the force in this earth, IU17 [ New ] and it has very strong economy, IU18 [ New ] it has very huge building. IU19 IU 40* [ New, based on own topic ] There are many people in the word like in his dreams to be American citizen IU20 [ New ] to get some of life chance.IU21 [ New ] Partner: yes IU20** Ajwad: but i like to spike in general IU16* Partner: to spike? IU21** w hat does it mean? IU22** Ajwad: the peopel coming to USA FROM ALL THE WORD, IU17* CUS THEY GET GOOD OFFER IN WORK, STUDY, IU18* GO TO HOSPITEL IU19* Ajwad: somthing like that IU20*

PAGE 425

399 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Partner: and for the american dream IU23** Partner: we can atalk about what IU24** we see in our culture class IU25** Ajwad: we can also start with th e history of immegrint IU21* Ajwad: AS WE STUDY IN BILL CLASS IU22* Partner YES!!!!! IU26** Partner :) IU27** Ajwad: SO OUR PRAGRAPH WELL BE IU23*

PAGE 426

400 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Ajw Partner: hoooooooo it's just a brainstorming today! IU28** great! we don't have to write a paragraph IU29** Ajwad: 1. HISTORY OF IMMEGRINT IU24* 2. WHOME? IU25* 3. WHAT IS THE ADVANTEG FRO IMMIGRINT IU26* Partner: so actually.... we can talk about immigration. Ameri IU30** Partner: why is the contry so diversified IU31** Ajwad: diversified? IU27* Partner: do you understand all teh pictures? IU32

PAGE 427

401 SCMC Matrix 5.5 (Continue d) Paragraph Comparison: Ajwad Ajwad: YES IU28* Partner: one is about the number of people, IU33** another is about immigration, IU34** an other one shows rich's and poor's houses IU35** Ajwad: what are u doing? IU29* can u tell me IU30* Partner: but I don't figure out the 3 others IU36** Partner: I'm chatting! IU37** Ajwad: i think #3 ABOUT IU31* Ajwad: THE OLD HUES IU32*

PAGE 428

402 SCMC Matrix 5 .5 (Continued) Paragraph Comparison: Ajwad Idea Unit) Partner: the old hues? IU38** Partner: the fifth one is NY city IU39** Partner: yes I can see the pictures! IU 40** thank you! IU41** Partner ;) IU42** Partner: But I try to understand the rapport with the People in USA IU43** Ajwad: AND NUMBER 4 ABOUT UNITED OF PEOPEL IN USA, IU33* CUSE THEY WAS FROM DEFRINT COUNTRY IU34* AND NOW THEY FR OM USA IU35* Ajwad: #5 ABUT THE BILDING TODAY IU36* Ajwad: MAP FOR USA TODAY IU37*

PAGE 429

403 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Ajwad: YES IU38* BUT I THINK IU39* IT IS TO SHOW THE BULDING TODAY. IU40* AND # 3 FOR THE BELDING YASTERDAY IU41* Partner: hoooh.... IU44** I was thinking IU45** that the number 3 was showing bulding of poor and rich people IU46** Ajwad: HE U MAM IU42* W HERE ARE U IU43* Ajwad: ????????????////// IU44* Partner: anyway IU47 Ajwad: WHAT? IU45*

PAGE 430

404 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Unit) Partner: We don't talk about architecture IU48** Partner: so we don't really need this picture IU49** Ajwad: WHY? IU46* Partner: because our topic is about people IU50** Ajwad: OK IU47* Partner: Maybe we can say IU51** that a lot of foreign people who come in the USA live in big city IU52** Partner: I think IU53** that there is a few foreign people in Kansas IU54** Partner: People prefer going in California, NY, Florida, Texas, Chicago IU55**

PAGE 431

405 SCMC Matrix 5.5 (Continued) Paragraph Comparison: Ajwad Ajwad: YES U GARED IU48* Ajwad: BUT POEPEL ALSO PRERER TO TAMPA ALSO IU49* Ajwad: DONT MISS THAT IU50* Partner: we finish. IU56** Partner: we prefer tampa! IU57** Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color own idea unit 4. Blue Highlight Color

PAGE 432

406 In his post treatment paragraph, Ajwad compo sed a total of 21 idea units. He constituted 9.52% of his paragraph. The strategies he used to rephrase these idea units were borrowing an idea unit from the discussion a nd rephrasing it through the application more chance to get a job better life and study ). unit (IU13**) from the discussion, he to be there are because there are different type of people People in America are very different However, most of the idea units he composed in the paragraph were new. Out of a total of 21 idea units in his post treatment paragraph, 19 idea units were new. These new idea units were either based on topics d iscussed during the pre writing stage and showed a low intertextuality on an idea unit level, or they further developed the topic of the paragraph without a direct connection with the discussion; thus, demonstrating no intertextual connections with the pre writing discussion. He developed nine low intertextuality idea units: five based on topics suggested by him (IU3, IU9, and IU10), two on topics suggested by his partner (IU13 and IU14), and two developed based on a ics (IU1 and IU2). For example, in his paragraph During the last hundred years, the United States totally change composed a new idea unit developing the topic of history of immigration suggested in his idea unit 21 in the pre writing di we can also start with the history of immegrint

PAGE 433

407 the discussion, but rather they served to develop a topic initially specified in the discussion presenting them in t he overall context of his paragraph and adding facts such as statistics. Ajwad developed most of his paragraph idea units as a continuation of the overall topic of the paragraph. I did not find any direct intertextual connections between 10 of the paragr aph idea units and any of the specific idea units from the discussion (IU 5, IU6, IU8, IU12, IU15, IU16, IU17, IU18, IU20, and IU21). For example, in idea units 5 and 6 he compared the availability of chances for better life and opportunities in other coun tries with the United States, a topic related to the overall theme but not specifically addressed All of those ideas may be able to found in every where, [IU5] but the different is more chance and equal opportunity used these newly composed idea units to either connect idea units based on discussion and develop a topic related to them (IU5, IU6, IU8, and IU12) or to incorporate a topic related to the theme of the paragraph as he did with the idea units that concluded his paragraph (IU15, IU16, IU17, IU18, IU20, and IU21). (Table 5.8 presents the types and the number of idea units

PAGE 434

408 Table 5.8 Paragraph idea units: Ajwad Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 0 0 1 1 0 0 0 2 4 3 10 21 Paragraph % 0% 0% 4.76% 4.76% 0% 0% 0% 9.52% 19.05% 14.29% 47.62% 100% Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. N ew based on own idea unit 7. 8. 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 435

409 Intertextual connections: conclusions on a lexical level was low. This was influenced both by the nature of the language generated during the SCMC discussion as well as the high number of paragraph ide a units that had low level intertextual or no intertextual connections with the discussion. writing discussion with regards to intertextuality on organizational level, it s eems that he be developed in the paragraph. Thus, the organization of the paragraph was influenced rom the discussion was rather limited: Ajwad borrowed only two idea units from the discussion and rephrased those using synonyms and syntactic modifications. the idea units fr om the pre writing discussion. They were related to topics from the pre writing discussion, rather than to specific idea units. Most of the pre writing discussion topics that he developed were topics that he proposed; however, he also considered his partne during the SCMC pre writing discussion which contained incomplete sentences or sentences that aimed to address the partner in a discussion discourse while conveying a spec ific idea related to the discussion topic. In addition, the highest number of idea units topic under discussion. These newly composed idea units were used either as a con nection and a development of the idea units that were composed based on the discussion or as a conclusion of the theme of the paragraph.

PAGE 436

410 Answering the Research Questions How do peers participate in synchronous CMC interactions? Ajwad engaged in a collabor ative pre writing discussion with his partner. However, I found that the collaborative tone of the discussion as well as discussion topics to be discussed during the chat s lead, he performed as an equal peer during the discussion posting a similar number of turns, contributing a slightly higher number of content idea units, and successfully initiating the discussion of paragraph organization. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implementation of these ideas? Ajwad wrote his paragraph considering the contributions m ade by him and his partner during the pre paragraph on a lexical level was low. This could be attributed to the nature of the language generated during the SCMC discussion as well as the high numb er of idea units of low level or no intertextual connections with the discussion. The organization of the from the discussion was rather limited: most of the idea uni ts composed in the paragraph were of low level intertextuality or revealed no intertextual connections with the pre writing discussion.

PAGE 437

411 Case Study Six: Kamil Kamil was a 19 year old student from Kuwait. He started his college education in Kuwait majoring in finance. The Spring semester of 2006 was his first semester at the ELI. During the semester, I observed the Writing III face to face classes of the section in which Kamil was enrolled as well as the computer laboratory classes. Based on my class observation notes, during class interaction, Kamil was one of the passive students who would not participate in class discussion unless the teacher addressed him specifically. There were times when he was not able to answer the was more focused when he worked in a small group. Once, the triad in which Kamil participated volunteered to present to the class the outcomes of their collaborative work. However, I would say that the overall trend in his class interaction was rather passive. about class organization in his country. He said: [the classes in his country] was rea lly very good organized yeah some times my teacher lecture duing in my class, actually I From this answer it appeared that Kamil preferred listening to his teacher rather th an actively participating in class discussion. The full text of the interview is provided in Appendix 5.10. During the computer laboratory classes, I did not notice Kamil to have any felt very comfortable with computers and preferred completing tasks using computers

PAGE 438

412 the computer is really important for our life and it`s very easy to done the works with it that why i prefer to work by computer because i get used to work with it in my high school and in my life .Also the computer chek my spelling mistake finnaly i prefer the Thus, in addition to being an experienced computer user, Kamil appreciated the use of computers for learning pu rposes. When asked if he had used computer mediated communication in the past, Kamil provided an affirmative answer without adding details. Based on the SCMC training tasks as well as on his treatment interaction, I could conclude that he had some experien ce with chat: he used emoticon symbols in his discussion as well as common chat abbreviations. negative. When I asked him about the most enjoyable aspect of writing in English, he s tated: Nevertheless, he identified extensive practice as a strategy for improving his writing in English: ting In addition, in the interview, Kamil showed awareness of the connection of writing skills to reading. The treatment interaction During the treatment interaction, Kamil demonstrated passive dyadic patterns while his partner served as a domin ant participant. In fact, their discussion was rather short. It included 10 turns, five made by Kamil and five by his partner, while the average number of turns in the SCMC treatment group was 46.46. The total number of idea units used by Kamil in his disc ussion was eight. Only three of them were content idea units; in other words, they addressed the topic of the pre writing task.

PAGE 439

413 about the pictures and topic under di scussion in seven content idea units out of nine. The writing discussion is presented next in SCMC Log 5.6. Log 5.6 SCMC Pre writing Discussion: Kamil and Partner Partner: hi 1 Kamil: y whats up 2 Kamil: well. i think the first picuter tell us there are many people wanna moved 3 from them countrey or travel to somewhere 4 Partner: in the firs picture it's seem that the populatin of usa incease 5 Kamil: yeah maybe who`s know : ) 6 Partner: the secound picture it's show th e amazing desing for the buldings and how 7 it's organize 8 Partner: the 4th picture it shows that in USA there are diffirent nationalities which 9 cause diffirents culture. 10 Partner: the final picture shows the map of USA and the impotant city there. 11 Kamil: hey whats up 12 Kamil: first of all i`m disagree :P foe every thing 13 As it is evident from this log, Kamil initiated the task posting his interpretation of the first picture. His posting was followed by his partner; however, Kamil became disengaged from the tas k completion, and the rest of his postings were not related to the task and were not collaborative in nature. His partner, after completing the posting of his own ideas about the pictures related to the task, also disengaged by leaving the

PAGE 440

414 discussion. For the rest of the treatment session, they did not post anything more and attempted to talk to each other face to face. In the interview, Kamil stated that the task was generally fine but not interesting It [the task] was ok. Little boring and engagement in the task completion as low. When sharing his perceptions and postings in his paragraph only if he agreed with them. T he paragraph: inertextual connections treatment writing showed overall negative improvement as compared to his pre treatment writing. In his post treatment writing, he had a higher score only on two of the criteria used for text analysis: voc abulary complexity, measured by analyzing the frequency of the unique words used and ideas presentation and development. His overall gains, as measured by the difference between the pulled z scores of post treatment and pre treatment writing, were 0.63600 1647 which was the second lowest pulled score in the SCMC treatment group. pre writing discussion were low. The ratio of the matching distinct lexical items between the paragr aph and the discussion was 18.92, whereas the average for the SCMC treatment writing discussion of the dyad in which he participated on the idea unit level are discussed. Kam il and his partner addressed limited aspects of the task. His partner outlined briefly his opinion about the pictures which were part of the task, and Kamil posted only one message directly related to the task. They did not discuss the organization of the future paragraph nor did they indicate their intention of planning for writing the

PAGE 441

415 paragraph was different from the order of the ideas posted in the discussion. Following is writing interaction with his partner and his paragraph along with analysis notes (SCMC Paragraph Comparison Matrix 5.6 : Kamil).

PAGE 442

416 SCMC Matrix 5.6 Paragraph Comparison: Kamil eas (by ID or close) Partner : hi IU1** Actually the United States has high population and immigration. IU1 IU3*&4* [New develops own topic, combines two IU ] There are many reasons for this high immigration I U2[ New text organization ] First of all people move to The United States IU3 IU4*[ New based on own IU ] because they need good jobs, good salary IU4 [ New comfortable where they are. IU5 [ New ] Also the United States has great and bea utiful houses and wonderful places IU6 [ New development of like the high building in New York, Las Vegas, and Chicago. IU7[ New ] Kamil : y whats up IU1* Kamil : well. i think the first picuter tell us IU2* there are many people IU3* wann a moved from them countrey IU4* or travel to somewhere IU5*

PAGE 443

417 SCMC Matrix 5.6 (Continued) Paragraph Comparison: Kamil Partner : in the firs picture it's seem that the populatin of usa incease IU2** Partner : the secound picture it's show the amazing desing for the buldings IU3** and how it's organize IU4** Partner : the 4th picture it shows IU5** that in USA there are diffirent nationalities IU6** whi ch cause diffirents culture. IU7** Partner : the final picture shows the map of USA IU8** and the impotant city there. IU9** Kamil : yeah maybe who`s know IU6* In Fact The immigration in United States come from many countries IU8 IU6**[ New based on partner ] like Spain, France, Korea, Japan, And Medal East IU9 [ New ] Finally actually the people in The united states are so friendly and polite IU10 [ New ] that why many people moved to united states IU11 IU3* IU4*[ New based on own, extended r epetition of paragraph idea unit ] Also they like to live there IU12[ New ] to find a good life with this people IU13 [ New ]

PAGE 444

418 SCMC Matrix 5.6 (Continued) Paragraph Comparison: Kamil s Paragraph Ideas (by ID or close) Kamil : hey whats up IU7* Kamil : first of all i`m disagree :P foe every thing IU8* Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color own idea unit 4. Highlig ht Color

PAGE 445

419 As it is evident from the Matrix presented above, all of the idea units Kamil composed were new 13 out of a total of 13. Three of the new idea units (IU 1, IU3, and on idea unit (IU4*). Idea unit 1 and idea unit 3 First of all people move to The United States immigration expressed in idea unit 4* and were of medium lev el of intertextuality. Further, one more of the new idea units (IU11) was related to the same discussion idea unit but, unlike in the pre writing discussion, was used to show a result; thus, I considered this idea unit to be of a low level of intertextuali ty. In two of his paragraph idea units, Kamil developed a topic suggested by his Also the United States has great and beautiful houses and wonderful places the secound picture it's show the amazing desing for the buildings would suggest that this idea unit had a low level intertextual connection with the pre In Fact The immigration in United States come from many countries that in USA there are diffirent nationalities a medium level. Further, Kamil composed eight idea units that h ad no intertextual connections with the pre writing discussion. He added one new idea unit (IU2) for paragraph organization purposes. In four of the new idea units in the paragraph (IUI4, IU5, IU7 and IU9), he provided support to the previous idea units gi ving specific examples which were not discussed during the pre writing interaction. In three of the idea units (IU10, IU12,

PAGE 446

420 and IU13) he addressed topics that were not directly discussed during the pre writing stage. Table 5.9 presents the types and the n

PAGE 447

421 Table 5.9 Paragraph idea units: Kamil Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 0 0 0 0 1 2 0 1 1 0 8 13 Paragraph % 0% 0% 0% 0% 7.69% 15.38% 0% 7.69% 7.69% 0% 61.55% 100% Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. New based on own idea unit 7. 8. New development o 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 448

422 Intertextual connections: conclusions Based on the analysis of intertextual connectio writing discussion, I would suggest that these connections were very weak on the lexical level and non existent on the organizational level. These weak connections were further confirmed on an idea unit level. Kamil attempted to compose his paragraph based on his own ideas from the discussion, which were limited due to his superficial and inadequate participation in the discussion, and addressed only two of the topics proposed by the partner. The strategy he applied when using the pre writing discussion was to address a topic presented in one or two idea units in the discussion, and compose new idea units based on these topics; thus, the limited intertextual connections were of medium and low levels of intertextuality. Most of the idea units which Kamil composed in his paragraph were new idea units with no intertextual connections with the pre writing discussion. Answering the Research Questions How do peers participate in synchronous CMC interactions? During the pre writing di scussion, Kamil demonstrated a passive dyadic interaction trend. He expressed disagreement and disengaged early from the discussion. This manner of task completion might be related to his low motivation and lack of interest in writing, expressed in his int erview answers. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implementation of these ideas? s pre writing discussion was low on the lexical level and non existent in terms of organization. The intertextuality

PAGE 449

423 of interest into the pre writing task and his ear ly disengagement. Thus, he and his partner failed to develop a substantial amount of content idea units related to the discussed topic which could have been used during the composition of the post treatment paragraph. SCMC Actors, Low Gains: Comparison Par Ajwad and Kamil were the students who showed the least gain in their post treatment paragraphs. There was age difference between the two actors: Ajwad was older than Kamil. In addition, they had different educational background: Ajw ad had a university degree while Kamil just started his college studies. Their perception of writing was also different. Ajwad stressed that writing was important for him because it was an important means of communication of his own ideas to other people, while Kamil expressed his negative attitude to writing, qualifying writing as a boring task. Both were experienced computer users who valued and enjoyed using computers in their everyday life. However, there was no evidence that they had used computer medi ated communication for educational tasks prior to the semester in which the study was conducted. Pre writing Interaction The participation of each actor in the interaction task was not similar. While Ajwad was a collaborative participant in the chat disc ussion and engaged actively in the task, Kamil was a passive participant who contributed very little to the discussion and disengaged early from the task. Additionally, their attitude towards the task was different. Ajwad expressed a positive attitude, app reciating the opportunity to exchange ideas with

PAGE 450

424 his partner. Kamil, on the other hand, stated that the task was boring for him, an attitude which he also demonstrated in his limited engagement in the pre writing task. As a result, the pre writing interact ions of these two actors differed. Ajwad and his partner had an extensive discussion in which they addressed multiple aspects of the assigned topic as disengagement, the intera ction between him and his partner resulted in few ideas being shared, most of them by his partner. These differences were reflected in the number of idea units which were directly related to the task (content idea units): Ajwad composed 20 content idea uni ts, while Kamil composed only three. Post treatment Gains and Intertextual Connections In their post treatment paragraphs, Ajwad and Kamil showed gains on two of the eight criteria used for text analysis. Both improved the lexical information per clause (measured through lexical density analysis). In addition, Ajwad improved the quantity of overall information present (measured by the number of idea units), while Kamil showed gains on paragraph presentation and development However, the overall gain of th eir post treatment paragraphs was negative, meaning that they regressed as measured with the cumulative z scores. The number of the idea units composed in the paragraph was considerably higher The intertextual connections were limited on the lexical level as well as on the idea unit level but differed on the organizational level. The ratio of matching distinct lexical items in the paragraph and the discussion was lower than the average SCMC tr paragraph did not reveal any intertextual connections on the organizational level while

PAGE 451

425 Ajwad considered the pre writing discussion when organizing his post treatment paragraph. The intertextuality analysis on an idea unit lev el revealed some similarities between the two actors. Both Ajwad and Kamil attempted to incorporate rephrased writing discussion, but these idea units were limited in number and were mainly of medium or low level o f intertextuality. In addition, both of them composed a considerable number of idea units with no intertextual connection to the pre writing discussion In other words, they developed topics not addressed in the discussion. Thus, the majority of the idea un its they composed in their paragraphs were new. negative attitude to writing, the limited development of the discussion could be a factor that prevented him from using idea s discussed during the pre writing stage. His limited interaction could be attributed to his view of writing as a boring task as well as of learning as a passive experience. I would suggest that he failed to involve himself actively in the discussion in a text discussion was much more interactive and collaborative in nature. This was evident in the way he incorporated ideas units from the pre writing discussion which included his own units. Based on this comparison, I suggest that the two actors, who demonstrated the lowest gains in their post treatment paragraphs, had different levels of engagement in the pre writing task They also demonstrated different patterns of dyadic interaction. This

PAGE 452

426 might have affected the quantity of their writing. Ajwad, as compared to Kamil, composed a paragraph with a higher number of idea units. However, they both composed mainly new idea uni ts in their paragraph, part of which resulted in low intertextuality on the pre writing discussion on the organizational level. Answering the Research Questions How do peers participate in synchronous CMC interactions? The two low gain participants engaged in a different way in the pre writing interaction using SCMC as a mode of communication: one of the actors was a passive and the other was a collaborative peer. Al though a variety of factors might have influenced these different patterns of dyadic interaction, I would suggest that their attitude towards writing in general, and specifically towards the pre writing task, influenced these patterns. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implementation of these ideas? The manner in which Kamil and Ajwad used their pre writing SCMC exchange was different and corresponded with their patterns of interaction which they demonstrated during the completion of the pre writing task. In his post treatment other hand, Ajwad was abl writing discussion when

PAGE 453

427 intertextuality was approached on the lexical and idea unit levels. They composed a considerable number of new idea units by either developing topics from the pre writing discussion or by composing idea units which conveyed their opinion on the overall topic of the paragraph and not discus sed during the pre writing interaction. Ajwad was more able to build strong intertextual connections in terms of organization, while on this level, related to the manne r in which the two actors were involved in the pre writing discussion. Case Study Seven: Kang Kang was a 24 year old student from Korea. He was studying in a Korean university in order to obtain an undergarduate degree in mechanical engi neering and interrupted his studies in order to come to the United States. Kang was a part of a large group of Korean students who came to the ELI for six weeks to improve their English for academic purposes. This group was treated in the same way as the r egular students at the ELI; in other words, in the beginning of the semester, their language proficiency was tested for placement purposes. Based on the outcomes of the placement test, they were assigned to the appropriate level and participated in the sam e classes as the other students. interview, was to learn English and obtain acceptable Test of English for International Communication (TOEIC) scores. This he related to his f uture career as an engineer. The educational environment to which Kang was exposed in Korea, based on his own report provided in the ACMC interview, was mainly teacher In korea, the classes is usually lecture. It have a merit that a student can learn many content for a short time. But

PAGE 454

428 because of that many krean student have a passive trend. I think it is a problem text of the interview is available in Appendix 17. Kang expressed insecurity in his writing skills. In his answer to the wr iting I would like to learn the skill of writing and word orders and useful expression..etc but, My english is In addition, in his interview he shared that his most serious difficulty when writing in English was the attempt to translate from Korean. In the interview Kang also showed his awareness of the importance of practice for developing writing skills and that the re was a connection We should read many kinds of English book. And Trying to writting somthing english at all time his awareness of the importance of writing as a communication medium. When identifying the most enjoyable aspect of writing in English for him, he stated that it was when people understood his writing. Unfortunately, due to the early departure to Korea, Kang was unable to provide me with more information with regar d to his views on the CMC pre writing discussion task and his computer experience. However, despite the lack of this information, I considered Kang as an actor in the instrumental case study. This decision was based on r the completion of the treatment task: he obtained the highest gain in the SCMC treatment group on his post treatment writing as measured by the pulled z scores difference between the post and pre treatment paragraphs. The analysis of his pre writing int eractions as well as his writing, along with the other

PAGE 455

429 information sources (e.g. his training pre treatment CMC interactions) would provide sufficient information for the instrumental case study. mputer experience, I considered the SCMC training interview task which the students completed before the treatment. In this task, the students worked in dyads to interview each other about their computer experiences and opinions about the importance of com puters in their everyday life. From the interview postings, it appeared that Kang was an experienced computer user. In the past, he had accessed information on the Internet, had used online banking, and played games. Based on this peer interview it appeare d that Kang was also an avid computers] just 10hr in a day. computer game, browing information and so on. addition, I noticed that in his interactions he used the Asia n versions of smiley face U you plz please The treatment interaction Kang and his partner engaged in a collaborati ve pre writing interaction in which they shared their views on the pre writing task and brainstormed ideas related to the paragraph topic. During the interaction, they were able to create a common discussion space sharing their own understanding of the pic tures and writing interaction.

PAGE 456

430 Log 5.7 SCMC Pre writing Discussion: Kang and Partner 1 PARTNER: HI 2 PARTNER: WE HAAVE TO DO BRAIN STORM ABOUT THE PO EPLE OF 3 usa 4 Kang: ok~ I see 5 PARTNER: FIRST OF ALL THERE IS ALOT OF POEPLE IN USA,and alot of 6 immegrants ull finde Spanish,Latins,Arabs,Indeians 7 Kang: usa is very wide continent. and have a many population. differnt races. 8 Kang: right.. 9 PARTNER: and from m y experiance i see that the Americans so pround of thier 10 freedom 11 Kang: mm .. excellent 12 Kang: and each state has a big city.. newyork, LA, chicago... 13 PARTNER: and u know what thats freedom cause this big cities they know how to 14 respect time ,they built thi er cuontry.As you can see large streets,huge 15 buldings,bridges,resorts alot of aspects 16 Kang: what is mean that freedom cause a big city" ..? 17 Kang: If I see a third picture, we can know that many people live in big city, and 18 the others live in small town.. 19 PARTNER: And they do alot of festivals to celebrate their countriy 20 PARTNER: ya and alot of of the small towns poeple go to big cities to have jops 21

PAGE 457

431 and have agood career 22 Kang: ok, so I think that the main idea is "Many people in the US lived different 23 way ." different from.. different race, different district.... 24 Kang: and they like freedom.. individualism.. 25 PARTNER: and dont forget something very imporatnt that all these poeple come to 26 USA because the oprotinity is available here in an aother way u can ma ke good 27 money here and own house and good jop 28 Kang: and then.. first.. People came to US because they have opportunity. 29 freedom..etc 30 and there are many population. 31 Kang: some people live in bic city.. and others live .. 32 Kang: big 33 PARTNER: so tell me have u ever been in any big city such as Sanfransisco or L.A 34 or Chicago? 35 Kang: not yet.. 36 Kang: I'm plannig to NewYork tomorrow.. 37 Kang: go to.. 38 Kang: Do you know about big city of USA?? 39 PARTNER: i went to Sanfransisco and Chikago its amazing poeple are so polite 40 and they treat u so goog and u know what attract me poeple work very hard really 41 Kang: ok.. good point.. actually I don't know exactly.. 42 PARTNER: so we r done we dont have any more idea 43 PARTNER: bye see u soon 44

PAGE 458

432 Kang: ok bye~ ^^ 45 In the discussion, it seems initiation (line 3) and closure (lines 41 42). However, Kang was an equal peer during the example, in line 12 he propo sed a new topic for discussion related to the overall topic of the task. Further, he summarized the ideas of the discussion and proposed a main idea for their paragraph (lines 22 contribution. This cons ideration was demonstrated in several ways: (1) he expressed a part of his part corrected a misspelling to make his posting more clear (line 31), and (5) in his summary (lines 22 24, and lines 28 the discussion. Kang and his partner did not address directly paragraph organization. Kang made an attempt to summarize and start ordering the outcomes of their discussion (lines 28 30), but it was not completed due to a topic change initiated by his partner. Towards the end of the discussion (lines 32 experiences related to the task, and Kang followed this shift. Thus, the end of their discussion was not directly related to the paragraph brainstorming. (initiating the chat, shifting the topic of the discussion away from the task, and proposing

PAGE 459

433 pa rticipants contributed almost equally to the task: Kang made 19 postings while his partner made 12. The idea units they composed during the interaction were also close in number: Kang composed 31 total idea units, while his partner composed 33. Their contr ibution to the discussion in terms of content idea units (in other words, idea units directly related to the topic of the pre writing task) was also similar in number: 14 content idea units were posted by Kang and 16 by his partner. This idea unit segmenta tion is presented in the SCMC Paragraph Comparison Matrix 5.7 : Kang, and it is further discussed in the following section in the context of the intertextual connections of the pre writing chat and the treatment paragraph. The treatment paragraph: intertext ual connections In his post treatment paragraph, Kang showed improvement on five out of the eight text criteria considered in the study: (1) lexical information per clause (measured through lexical density analysis), (2) vocabulary complexity (measured by analyzing the frequency of the unique words used), (3) rhetorical soundness, (4) presentation and development of main ideas, and (5) overall language use (the last three criteria were assessed using a multiple trait rubric). He demonstrated negative gain o n the other three criteria: (1) the syntactic complexity (measured by calculating the mean length of t units), (2) the amount of information present in a single focus (measured by mean length of idea units), (3) the quantity of overall information present (measured by the number of idea units) The overall gain as measured by the difference between the pulled z scores of post treatment and the pre treatment writing was 0.64492224. This was the highest gain in the SCMC treatment group. (Refer to Appendix 5.1 by measure.)

PAGE 460

434 treatment paragraph w ith the pre writing discussion on a lexical level was lower than the average for the SCMC treatment group: the ratio of matching distinct lexical items of his paragraph and the pre writing discussion was 21.43, while for the SCMC treatment group the average ratio was 26.31. pre wri ting discussion on the organizational and idea unit levels. Kang and his partner did not discuss explicitly the organization of the paragraphs they were preparing to write. Kang attempted to start this discussion; however, this attempt was underdeveloped. In his paragraph, he did not follow the order in which the ideas were shared. For example, he proposed to start the paragraph with the topic of population; however, his paragraph topic sentence developed the idea of immigration, and population came as his first supporting idea. The topics of opportunity and freedom were addressed in different postings in the pre writing discussion, but were merged in the paragraph in two consecutive idea units. Thus, on the organizational level, no intertextual connections of his paragraph and the pre writing discussion were detectable. The full text matrix that incorporates notes on my analysis on idea unit connections.

PAGE 461

435 SCMC Matrix 5. 7 Para graph Comparison: Kang Partner: HI IU1** Many people have come to the United State of America for 250 years IU1 [ New based on details ] including Europe, North America, South America, etc. IU2 [ New The population of America were getting more and more. IU3 IU4*, IU21* [ Rephrased own, synonyms and syntax changes ] It was a good chance for immig rant to immigrate, IU4 [ New, IU18**] Kang: ok~ I see IU2* Kang: usa is very wide continent. IU3* and have a many population. differnt races. IU4* Partner: WE HAAVE TO DO BRAIN S TORM ABOUT THE POEPLE OF usa IU2** Partner: FIRST OF ALL THERE IS ALOT OF POEPLE IN USA, and alot of immegrants IU3** ull finde Spanish, Latins, Arabs, Indeians IU4**

PAGE 462

436 SCMC Matrix 5. 7 (Continued) Paragraph Comparison: Kang dea Unit) Kang: right.. IU5* because of freedom and quality of opportunity. IU5 [ New freedom IU6** and IU8**] They had to mix diverse cultures. IU6 [ New based on own topic from IU16* and IU17*] So, they have had a strong individualism IU7 [ New develops own topic, IU18*] based on self reliance. IU8 [ New ] Many of the American live big city, IU9 IU12*[ Own rephrased, syno nyms ] such as Newyork, Chicago, Partner: and from my experiance i see IU5** that the Americans so pround of thier freedom IU6** Kang: mm .. excellent IU6* Kang: and each state has a big city.. newyork, LA, chicago... IU7* Partner: and u know IU7 ** what thats freedom IU8** cause this big cities they know how to respect time IU9** ,they built thier cuontry. IU10**

PAGE 463

437 SCMC Matrix 5. 7 (Continued) Paragraph Comparison: Kang ragraph (by Idea Unit) As you can see large streets,huge buldings,bridges,resorts alot of aspects I U11** Los Angeles, Sandiego. IU10 IU7* [ Own part of IU7*] There are many tall building in big city. IU11 [ New topic, part of IU11** ] The others live small town. IU12 13* [ Own, verbatim, only drops ] Kang: what is mean IU8* that freedom cause a big city" ..? IU9* Kang: If I see a third picture, IU10* we can know IU11* that many people live in big city, IU12* and the others l ive in small town.. IU13*

PAGE 464

438 SCMC Matrix 5. 7 (Continued) Paragraph Comparison: Kang Unit) Kang: ok, so I think IU14* that the main idea is IU15* "Many peop le in the US lived different way" IU16* different from.. different race, different district.... IU17* Kang: and they like freedom.. individualism.. IU18* Partner: and dont forget something very imporatnt IU17** that all these poeple come to USA IU18* because the oprotinity is available here IU19** in an aother way u can make good money here IU20** and own house and good jop IU21**

PAGE 465

439 SCMC Matrix 5. 7 (Continued) Paragraph Comparison: Kang Idea Unit) Kang: and then.. first.. People came to IU19* because they have opportunity. freedom..etc IU20* and there are many population. IU21* Kang: some people live in bic city.. IU22* and others live .. IU23* Postings (by Idea Unit) Partner: so tell me IU22** have u ever been in any big city such as Sanfransisco or L.A or Chicago? IU23** Kang: big IU24* Kang: not yet..IU25*

PAGE 466

440 SCMC Matrix 5. 7 (C ontinued) Paragraph Comparison: Kang (by Idea Unit) Kang: I'm plannig to NewYork tomorrow.. IU26* Kang: go to.. IU27* Kang: Do you know about big city of USA?? IU28* Partner: i went to Sanfransisco and Chikago IU24** its amazing IU25** poeple are so polite IU26** and they treat u so goog IU27** and u know IU28** what attract me IU29** poeple work very hard really IU30**

PAGE 467

441 SCMC Matrix 5. 7 (Continued) Paragraph Comparison: Kang agraph (by Idea Unit) Kang: ok.. good point.. IU29* actually I don't know exactly..IU30* Partner: so we r done IU31** we dont have any more idea IU32** Partner: bye see u soon IU33** Kang: ok bye~ ^^ IU31* Coding scheme: 1. IU underlined con tent idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color own idea unit 4. Blue Highlight Color

PAGE 468

442 In his paragraph, Kang composed a total of 12 idea units. Eight of these idea units were new. However, there was a significant relation of idea units to the text of the pre writing discussion. All of the paragraph text idea units but one showed a clear relation with the text. I identified five types of new idea units composed by Kang as based on their relation with the new idea units based on own idea unit or topic, (3) new idea unit developing or combining topics discussed by partner, (4) new idea unit developing his own idea unit, and (5) new idea uni t not traceable in the pre writing interaction. Table 5.10 presents the

PAGE 469

443 Table 5.10 Paragraph idea units: Kang Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 To tal Number of IUs 0 1 0 3 2 1 0 3 1 0 1 12 Paragraph % 0% 8.33% 0% 25% 16.68% 8.33% 0% 25% 8.33% 0% 8.33% 100% Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. New based on own id ea unit 7. 8. 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 470

444 Mo st of the idea units composed by Kang in his paragraph showed connection with the pre writing discussion. In 33.33% of the paragraph idea units, this connection was of a high intertextuality level being directly traceable to specific idea units which were either borrowed verbatim or rephrased. He used one of his own idea units verbatim, dropping only a conjunction (IU12 IU13*). He also rephrased or clipped three of his own idea units from the discussion and incorporated them into his paragraph (IU3 IU4* and IU21*, IU9 IU12*, and IU10 In population of America were getting more and more and have many population. different races and there are many population of synonyms and syntactic changes. units in the discussion; thus, they revealed a me dium level of intertextuality between the discussion and the paragraph. Kang based these idea units in his paragraph on specific idea units from the discussion (IU1 IU18**, IU2 IU4**, IUI6 IU16* and IU17*). For Many people have come to the United State of America for 250 years th at all these people come to USA writing discussion. One related to the pre writing discussion; thus, I identified their level of intertextual

PAGE 471

44 5 connection as low. Four idea units developed topics which had been discussed during the pre writing stage but were not directly traceable to a specific idea unit (IU4, IU5, IU11, and IU7). For exa because of freedom and quality of opportunity partner several times during the discussion and then addressed by Kang himself. However, his paragraph id ea unit was not directly related through verbatim borrowing or rephrasing of any of the several idea units from the discussion in which the topic was connection with the discussion (IU 12). Intertextual connections: conclusions When analyzing the intertextual writing discussion, it appeared that they had rather weak intertextuality on the lexical level. The ratio o f matching distinct lexical items of his post treatment paragraph and the pre writing discussion was rather low. The discussion of paragraph organization was underdeveloped; thus, the intertextual relations of the discussion and the paragraph were not pres ent. However, when I analyzed his paragraph on an idea unit level examining the connections between the discussion and paragraph idea units, it became evident that the connections between the pre writing discussion and the post treatment paragraph were muc h stronger revealing intertextual connection at a high medium and low levels of intertextuality. Kang demonstrated a collaborative dyadic interaction pattern in his pre writing discussion. This was further transferred to his paragraph writing: he was able to consider He applied multiple strategies when using the pre writing discussion. Some of these

PAGE 472

446 strategies allowed him to use language strings from the discussion: he borrowed verbatim one idea unit and rephrased three. Other strategies were directed towards the use of topics discussed in specific idea units and developing them in newly composed idea units. He used this strategy for three idea units. He developed fo ur new idea units based on topics addressed during the discussion but with connections not directly traceable on an idea unit level. Only one idea unit was not related directly to the discussion. Thus, most of the treatment para graph (91.77%) were related to the pre writing discussion, which suggested a strong intertextual connection between the two texts on the idea unit level. Answering the Research Questions How do peers participate in synchronous CMC interactions? lvement in the pre writing discussion showed his ability to collaborate with his partner and develop the assigned discussion topic. Although his partner had a leading role at certain moments of the discussion with regards to initiation, topic shift, and di scussion conclusion, Kang was able to contribute to the discussion discussion exchange. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implementation of these ideas? writing discussion in which he participated was weak on the lexical level and non existent on the organiz ational level. writing discussion, which was evident from

PAGE 473

447 the paragraph multiple idea units of high and medium level intertextuality. He applied various strategies when using the pre writing discussion such as ve rbatim borrowing, rephrasing, using topics discussed in specific idea units and developing them in newly composing his paragraph. Case Study Eight: Sun Sun w as a 27 year old student from Taiwan. She had received her undergraduate degree in the field of management information systems and worked as and becoming a more proficient writer in English were related both to personal and I hope to learn how to writing mails and reports for business and personal purposes. I woul d like to improve my speaking and listening abilities In the ACMC interview when answering my questions about writing, Sun addressed the importance of having extensive vocabulary. She incorporated this opinion in her answer about the importance of being have a lot of vocabulary to describe what he/she want to say vocabulary are not enough vident again when she shared her most I can learn more and more vocabulary and let the other know my opinion important communication medium. The full te Appendix 5.12.

PAGE 474

448 Based on her interview answers, her academic and professional background, and my observations of Sun in the computer laboratory writing sessions, I concluded that she was an experienced computer user wh o used computers for various tasks. In the interview I am pretty comfortable when working withe computers. Actually, my job is computer software designer and I work use computers everyday. I will uncomfortable if no computers can be used anymo re We used computer to design software, do homework, look for information for presentation. Prior to her enrolment at the ELI, Sun had use d CMC extensively for various purposes that included personal communication as well as study and work related communication. In her interview, she posted a detailed message answering the question yes, I used computers to communi cate with other people, 1) chat with my classmates, friends, customers and my team members 2) email to my classmates, friends, customers and my team members. For business purposes and personal communication. 3) Using the online banking services, communicat e with banks 4) using the online stores to buy something, like books had used CMC for personal communication and problem solving tasks prior to the semester in which data were collected. The treatment interaction In her in terview, Sun shared that she liked both the ACMC and SCMC training tasks. However, she did not like the design of L.E.C.S., the software used for SCMC communication. This dislike of L.E.C.S. was also shared by several other students in their interviews. Su n felt comfortable while completing the SCMC treatment pre I found it

PAGE 475

449 was easy to discussion of the pictures. Nothing [was difficult] expect I must increase my vocabulary y to be helpful for her because it allowed her to However, she did not perceive CMC discussion activities to be helpful for the Discuss ion activities are like chats. I can get for information but it's not useful for learning how to write better typing skills were well developed; therefore, they did not impede her online pre writing discussion. She also shared th at she used online and an electronic dictionary during the SCMC Log 5.8 Pre writing Discussion: Sun and Partner Partner: hola 1 Partner: are you there ? 2 Sun: hello,nice to cu .. 3 Sun: do u have a page with six pitures?? 4 Partner: nice to chat with you 5 Partner: let me read the topic 6 Partner: yes i have it 7 Sun: the pictures presenting about "the people of the US. 8 Sun: me too.. 9 Partner: so 10 Partner: what we have to talk about ...j ust about the people ? 11 Sun: the topice is "the people of the US. yes.. 12

PAGE 476

450 Partner: i think ya 13 Sun: i will write something about my pictures.. 14 Partner: ok 15 Partner: but i have the same pic 16 Sun: the first one... is the areas of the US inclueds.. there are a lot of people live 17 there.. 18 Sun: yes...i don't know why .. but they all have the same picture.. 19 Partner: i think they mean that .. in U.S there is a lot of people not just from U.S 20 there are people from evey where in the world 21 Sun: maybe chat with the detail of the picture.. 22 Partner: not the pic ..we have to write about the people in U.S 23 Sun: and do u know .. which three areas the us has ?? 24 Sun: not in the pic?? 25 Partner: no 26 Sun: so what should we do..chat about american people?? 27 Partner: yes 28 Sun: maybe you first ...and i will follow u. 29 Partner: ok 30 Sun: sure..let's start 31 Partner: the U.S is a counry there are people from every where in the world .. so it's 32 a counry has a lot of cultures and different people and may be language 33 Sun: is true.. 34 Partner: the 2nd pic a bout the building in the U.S ..so it showes there are a hight 35

PAGE 477

451 building in some where in U.S like New your city or in some city ..so it's different 36 from each city ..every city has a different things 37 Sun: the 3nd pic also about the house in us..people maybe will not only live in the 38 city but also live in country.. 39 Sun: so 2nd 3nd show myabe different cluture has different housing sytels.. 40 Partner: ot i think it talk about the people who moved from the village to the city .. 41 sotge U.S had a village in the pa st but now everything changed ..so they have a 42 cities 43 Partner: what about the 3rd pic 44 Partner: the 4th 45 Sun: 3nd show myabe different cluture has different housing sytels.. 46 Sun: maybe.. 47 Partner: ya 48 Partner: the 4th pic 49 Sun: you will say something about 4n d one?? 50 Partner: it may be mean the freedom in the U.S ..so here in people can say what 51 ever they want and they can walk in the street and judg the goverment 52 Sun: 5nd about Los Angeles is the place where has most different culture in the 53 city. 54 Partner: it' s in los angelos or in New your city 55 Sun: maybe.. 56 Partner: this brige in New york 57 Partner: i'm sure 58

PAGE 478

452 Partner: the last pic 59 Sun: so always has a lot of different culture in us city. 60 Sun: ok,thanks 61 Partner: it's apears how bit is the U.S 62 Sun: yes 63 Sun: ok go od job.. 64 Sun: we have done 65 Sun: ok let's go home now... 66 Partner: she said keep chating 67 Partner: we can chat about something different 68 Sun: ok sure.. 69 Sun: i like it 70 Sun: will u study in USF ? 71 Partner: i think yes 72 Partner: but after 10000000000 year 73 Partner: because i need the toefl 74 Partner: it's too diffecult 75 Sun: wow ... so long time 76 Sun: rite..it's really very different.. 77 Partner: i guese 78 Sun: but i think i can do it very well 79 Sun: what major will u attach? 80 Partner: for me it's too diffecult 81

PAGE 479

453 Partner: they need 550 82 Partner: buisness 83 Sun: i believe that you can do well 84 Sun: business ... why? 85 Partner: i hope 86 Partner: i dont know 87 Partner: i dont have another choise 88 Sun: :) 89 Sun: good luck..it time for going home... 90 Sun: cu tomorrow 91 Sun: have a nice day 92 Partner : ok see you 93 In their chat, Sun and her partner engaged in a collaborative online discussion with equal contribution to the task completion, Sun posting 41 messages, and her partner posting 42. Sun was the one to initiate and conclude the task discussion. In addition, when Sun and her partner perceived that they had completed the tasks, they remained online and continued chatting about personal plans and experiences not related to the task. This continuation of the chat was initiated by her partner. During the pre writing discussion, Sun initiated the discussion of the assigned topic (line 4) and guided the process of task completion in several of her postings (lines 8 and 21). There was a short confusion about what exactly the task entailed which was resolv ed during the SCMC interaction (lines 16 Sun contributed to the discussion sharing her opinion about the topic and the images they

PAGE 480

454 had to use in order to develop the discussion and prepare for the paragraph writing task (lines 37 38, 44, 51, 57). She also invited her partner to share his opinion about the pictures (lines 26, 48) and indicated recognition of his contribution (lines 33, 58, 60, 61). Sun was the one to indicate her perception of the task being completed (line 62) and to initiate the closure of the discussion (line 63, 87). However, when she proposed that they with the chat even though they finished the pre writing discussion. Sun followed her 90). Although Sun and her partner indicated that they perceived the discussion task to be completed, in their interaction they did not disc uss the organization of the paragraph and did not address explicitly every picture from the task. Due to the reason that the chat incorporated greetings, discussion of the task requirements, as well as non task related chat, the content idea units compose d during the pre writing discussion were fewer than the overall number of idea units. The overall number of idea units was 112, with 54 idea units posted by Sun and 65 by her partner. However, the content idea units (in other words, the idea units directly related to the topic was partially a result of her more active involvement in chat organization and directing the task completion. In the following section, the chat and paragraph idea unit segmentation is presented and discussed in order to establish their intertextual connections. The paragraph: intertextual connections The post treatment paragraph written by Sun showed a positive gain as compared with her pre treatment paragraph. Sun improved

PAGE 481

455 her writing on five of the proposed eight text analysis criteria: (1) the syntactic complexity (measured by calculating the mean length of t units), (2) the amount of information present in a single focus (measured by mean length o f idea units), (3) the quantity of overall information present (measured by the number of idea units), (4) vocabulary complexity (measured by analyzing the frequency of the unique words used), and (5) rhetorical soundness (the last criterion was assessed u sing a multiple trait rubric). The overall gain as measured by the difference between the pulled z scores of post treatment and the pre treatment writing was 0.460250215. This was the second to the highest gain in the SCMC treatment group. (Refer to Appen dix 5.1 for the complete list scores by measure.) treatment paragraph showed low intertextual connections with the pre writing discussion on the lexical level. It was lower than the average for the SCMC treatme nt group: the ratio of matching distinct lexical items of her paragraph and the pre writing discussion was 16.51, while for the SCMC treatment group the average ratio was 26.31. Paragraph organization was not discussed by Sun and her partner during the pre shared during the discussion. Thus, I could not establish an intertextual connection on the organizational level between the organization of the paragraph and the discus sion. Below, writing discussion in which she participated on the organizational and idea unit levels. The full text of the pre presented in the following comparison matrix along with my idea unit analysis (SCMC Paragraph Comparison Matrix 5.8 : Sun)

PAGE 482

456 SCMC Matrix 5.8 Paragraph Comparison: Sun Pa rtner: hola IU1** Partner: are you there ?IU2** There are a lot of people in United States, IU1 IU8* [ Own rephrased synonyms and syntactic changes ] including United States continent, Alaska and the other insular area. IU2 [ New ] The capital of United States is Washington, D.C. IU3 [ New ] People who live in USA always treated the government with respect. IU4 [ New ] People in United States live in cities or towns IU5 IU22* and 23* [ Own rephrased merges two IUs, sy ntactic and synonym changes ] where the housings Sun: hello,nice to cu .. IU1* Sun: do u have a page with six pitures?? IU2* Partner: nice to chat with you IU3** Partner: let me read the topic IU4** Partner: yes i have it IU5** Sun: the pictur es presenting about "the people of the US. IU3* Sun: me too.. IU4*

PAGE 483

457 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Partner: so IU6** Partner: what w e have to talk about IU7** ...just about the people? IU8** have several different styles. IU6 IU25* [ Own rephrased, IU clipped, syntactic change ] Because of the culture, climate and location are different from each other. IU7 [ New t opic of diversity IU21 25**] In Big city, the buildings are tall and fashion, IU8 [ New, ] however the buildings in the towns or countries are small. IU9 [ New ] There are a lot of people IU10 Sun: the topice is "the people of the US. yes.. IU5* Partner: i think ya IU9** Sun: i will write something about my pictures.. IU6* Partner: ok IU10** Partner: but i have the same pic IU11**

PAGE 484

458 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun ea Unit) Sun: the first one... is the areas of the US IU7* inclueds.. there are a lot of people live there.. IU8* Sun: yes...i don't know why .. IU9* but they all have the same picture.. IU10* IU15* [ clipped and NP extended with a modifier ] who came from different cultures and different countries, IU11 [ New 25** ] and they live peace with each other. IU12 [ New ]They all follow the rules which e stablished by the government with their will. IU13 [ New IU45**] So, the people of the United States are brothers and sisters IU14 [ New ] who live in a big family. IU15 [ New ] The New York is the one of the representatio n of cultural city. IU16 Partner: i think IU12** they mean that .. IU13** in U.S there is a lot of people not just from U.S IU14** there are people from evey where in the world IU15** Sun: maybe chat with the detail of the picture.. IU11*

PAGE 485

459 SCMC Matrix 5.8 ( C ontinued) Paragraph Comparison: Sun Partner: not the pic ..IU16** we have to write about the people in U.S IU17** [ New based on a combination of ow IU30* IU31*, and IU47** ] There are numerous people IU17 IU22** [ develops ] who came from different countries and areas IU18 IU22** [ s clipped, part extended, splits ] get together IU19 [New ] and work for themselves. IU20 [ New ] There are several big cities in United States, IU21 [ New ] not only New York City IU22 Sun: and do u know IU12* . which three areas the us has ?? IU13* Sun: not in the pic?? IU14* Partner: no IU18** Sun: so what should we do IU15* ..chat about american people?? IU16* Partner: yes IU19**

PAGE 486

460 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Partn Sun: maybe you first ...IU17* and i will follow u IU18* [ New ] but only Los Angeles, Sam Diego, Houston and Chicago, IU23 [ New ] in where many people live. IU24 [ New ] The United States continent is around by Pacific Ocean, Atlantic Ocean and Gulf of Mexico IU25 [ New ] and the continent is big, IU26 IU51** [ rephrased, syntactic and synonym changes ] so there are a lot of natural resources IU27 [ New ] which people can use.IU28 [ New ] By the way, people in United States are racial, ethnic, cultural diversity. IU29 IU22**, Partner: ok IU20** Sun: sure..let's start IU19* Partner: the U.S is a counry IU21** there are people from every where in the world .. IU22** so it's a counry IU23** has a lot of cultures IU24** and different people and may be language IU25**

PAGE 487

461 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Su n: is true.. IU20* IU23**[ New IU24**, IU25**] Partner: the 2nd pic about the building in the U.S .. IU26** so it showes IU27** there are a hight building in some where in U.S like New your city or in some city IU28** ..so it' s different from each city .. IU29** every city has a different things IU30**

PAGE 488

462 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Unit) Sun: the 3nd p ic also about the house in us.. IU21* people maybe will not only live in the city IU22* but also live in country.. IU23* Sun: so 2nd 3nd show myabe IU24* different cluture has different housing sytels.. IU25* Partner: ot i think IU31** it talk about t he people IU32** who moved from the village to the city .. IU33** sotge U.S had a village in the past IU34** but now everything changed .. IU35** so they have a cities IU36**

PAGE 489

463 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Idea Unit) Partner: what about the 3rd pic IU37** Partner: the 4 th IU38** Sun: 3nd show IU26* myabe different cluture has different housing sytels.. IU27* Sun: maybe.. IU28* Partner: ya IU3 9** Partner: the 4th pic IU40** Sun: you will say something about 4nd one?? IU29* Partner: it may be mean the freedom in the U.S .. IU41** so here in people

PAGE 490

464 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Unit) can say IU42** what ever they want IU43** and they can walk in the street IU44** and judg the government IU45** Sun: 5nd about Los Angeles is the place IU30* where has most different cultur e in the city. IU31* Partner: it's in los angelos or in New your city IU46** Sun: maybe.. IU32*

PAGE 491

465 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Pa rtner: this brige in New York IU47** Partner: i'm sure IU48** Partner: the last pic IU49** Sun: so always has a lot of different culture in us city. IU33* Sun: ok,thanks IU34* Partner: it's apears IU50** how bit is the U.S IU51** Sun: yes IU35* Sun: ok good job..IU36*

PAGE 492

466 SCMC Matrix 5. 8 (C ontinued) Paragraph Comparison: Sun Sun: we have done IU37* Sun: ok let's go home now... IU38* Partner: she said keep chating IU52** Partner: we can chat about something different IU53** Sun: ok sure.. IU39* Sun: i like it IU40*

PAGE 493

467 SCMC Matrix 5.8 (C ontinued) Paragraph Comparison: Sun Sun: will u study in USF ? IU41* Partner: i think yes IU54** Partner: but after 10000000000 year IU55** Partner: because i need the toefl IU56** Partner: it's too difficult IU57** Sun: wow ... so long time IU42* Sun: rite..it's really very different.. IU43* Partner: i guese IU58**

PAGE 494

468 SCMC Matrix 5.8 (Continued) Paragraph Comparison: Sun Sun Sun: but i think IU44* i can do it very well IU45* Sun: what major will u attach? IU46* Partner: for me it's too difficult U59** Partner: they need 550 IU60** Partner: business IU61** Sun: i believe that you can do well IU47* Sun: business IU48*... why? IU49* Partner: i hope IU62** Partner: i dont know IU63** Partner: i dont have another choise IU64**

PAGE 495

469 SCMC Matrix 5.8 (C ontinued) Paragraph Comparison: Sun ea Unit) ShuFen, :) IU50* Sun: good luck..IU51* it time for going home... IU52* Sun: cu tomorrow IU53* Sun: have a nice day IU54* Partner: ok see you IU65** Coding scheme: 1. IU underlined content idea unit 2. Yellow Highlight Color new idea unit 3. Grey Highlight Color own idea unit 4. Blue Highlight Color

PAGE 496

470 Sun composed 29 idea units in her post treatment paragraph. Out of these 29 idea units, I was able to find connection between 15 paragraph idea units and idea units from the discussion. The degree and the nature of these connections were different. Sun used sion or developed idea units based on topics discussed in the pre writing discussion. Thus, in addition to the rephrased idea units from the paragraph, I identified five types of new idea units composed by Sun based on their relation with the text: (1) new own idea units, (3) new, developing and interpreting a topic suggested by the partner, (4) new, developing of own topic, and (5) new, not traceable in the pre writing interaction. Table 5.11 presents the

PAGE 497

471 Table 5.11 Paragraph idea units: Sun Intertextuality High level Medium level Low level No Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Number of IUs 0 0 3 4 5 1 0 1 0 0 15 29 Paragraph % 0% 0% 10.34% 13 .79% 17.24% 3.45% 0% 3.45% 0% 0% 51.73% 100% Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. 6. New based on own idea unit 7. 8. Ne 9. New development of own topic 10. 11. New no traceable connections with pre writing discussion

PAGE 498

472 Seven out of the 29 idea units of her paragraph were rephrased idea uni ts from the pre writing discussion; thus, they revealed high level intertextual connection with the pre There are a lot of people ) and idea unit 11 ( who came from different cultures and different co untries there are people from evey where in the world. idea units. Four idea units were paraphrased own idea uni ts (IU10, IU17, IU18, and strategies she used were applying (1) syntax changes, (2) replacing words and phrases with synonymous words and phrases, (3) merging two discussion id ea units into one, (4) shortening idea units, (5) extending idea units, and (6) separating one idea unit from the discussion in to two paragraph idea units. These strategies were not used alone but were always combined. Sun made syntax changes and used syn onyms in two idea units. She combined the merging strategy with syntax changes in one idea unit. One idea unit was shortened with an application of syntax changes. In one paragraph idea unit Sun used part of the discussion idea unit and extended it by appl ying syntax changes. Finally, she separated one idea unit from the discussion into two in the paragraph, extending them and applying syntax changes. Sun also included seven new idea units in her writing which showed connections with the pre writing discus sion. Six of them (IU7, IU8, IU11, IU13, IU16, and IU29) were developed based on several idea units from the discussion; thus, I identified the intertextual connection of these idea units with the pre writing discussion as medium level. One was more genera lly related to topics proposed by her partner; thus, it was of a

PAGE 499

473 Because of the culture, climate and location are different from each other units shared b has a lot of cultures (IU24**) and different people and may be language climate location discussion idea units is too distant for it to qualify as paraphrasing. However, it was traceable; thus, I interpreted it rather as a development of discussion idea units towards a new idea unit. Another example of a medium level of intertextual connection between a parag raph idea unit and discussion idea units is illustrated with paragraph idea unit 16 (IU16). This 5nd about Los Angeles is the place and idea unit 31* ( where has most different culture in the city the discussion. In addition, in idea unit 16, Sun showed consideration of he this brige in New York changed the name of the city in her paragraph idea unit. Only one idea unit from the post treatment paragraph showed low level intertextuality as compared to the pre writing discussion. In this idea unit Sun interpreted and judg the government and developed it in i They all follow the rules which established by the government with their will The highest number of idea units composed by Sun were new idea units that,

PAGE 500

474 although related to the topic of the paragraph, we re not directly related to the pre writing discussion, and thus had no intertextual connections with the discussion. These were idea units IU3, IU9, IU12, IU14, IU15, IU19, IU20, IU21, IU22, IU23, IU24, IU25, IU27, and IU28. Most of them were used in the s econd part of the paragraph after the idea units from the discussion were developed. Sun succeeded in establishing connections between her new idea units that were not related intertextually to the pre writing discussion and the ones that were related. It ertextual connections: conclusions paragraph and the pre writing discussion on a lexical level were rather weak. The ratio of matching distinct lexical items of her post treatment paragraph and the pre writing discussi on was lower than the average for the SCMC treatment group. I could not detect intertextual connections on the organizational level. When I analyzed the intertextual connections on the idea unit level, it appeared that Sun considered the pre writing inter action when composing her paragraph. Sun based nearly half of her idea units (48.28%) on idea units from the discussion by (1) establishing high level intertextual connections through the inclusion of rephrased own ion into her paragraph, (2) establishing medium level intertextual connections through the use of idea units from the discussion as a basis for development of new idea units, and (3) establishing medium level intertextual connections through the developmen t of a new idea unit based on a topic shared during the discussion into a new idea unit. In her paragraph writing, Sun showed a strong connected to the discussion were her

PAGE 501

475 However, most of her idea units (51.72%) were not related to the discussion. Sun composed 15 new idea units which were not directly related to idea units or topics from the discussion. Thus, I would qualify the intertextual connection of her paragraph as analyzed on the idea unit level as medium. Answering the Research Questions How do peers participate in synchronous CMC interactions? In their chat, Sun and her partner engaged in a collaborative online discussion with equal contribut ion to the task completion. Sun showed her ability to collaborate with her partner and develop the assigned discussion topic. Although her partner had a leading role at certain moments of the discussion, Sun was able to contribute to the discussion equally discussion exchange. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and similarities in the implem entation of these ideas? writing discussion was weak on the lexical level. On the organizational level, there was no intertextuality between the pre writing discussion and the post treatment paragr aph texts. In her paragraph, Sun showed strong consideration of the pre writing discussion. This consideration was evident from the multiple idea units which showed either high or medium level of intertextuality. In addition, in her writing Sun was able to consider both idea units that were not related to the pre writing discussion.

PAGE 502

476 SCMC Actors, High Gains: Comparison Sun and Kang were the two case study actors from the SCMC treatment group who showed the highest degree of gain in their post treatment writing. Both actors were of approximately the same age: Sun was 27 years old and Kang 24. Although there was a difference in th eir educational background (Kang was still in the process of obtaining his university degree, and Sun already started her career after graduating from university), their goals for studying English were career related. They expressed similar attitudes towar ds writing in their interview answers, suggesting that they were motivated to learn how to write better. They were able to articulate their difficulties and positive aspects as related to writing in English. In his note to the writing teacher, Kang recogni zed the addition, he stated that translation from Korean was a specific difficulty that he experienced in the process of writing. On the other hand, in her interview w ith me, Sun stressed her lack of extensive vocabulary in English as one of the factors preventing her from being a good writer in English. They both pointed out that the enjoyable aspect related to writing in English was the ability to communicate their id eas. Both Sun and Kang were experienced computer users. They expressed their appreciation of the possibilities with which the computer provides them. In addition to being an experienced computer user, Sun was also involved in professional tasks related t o designing computer software. While Kang did not explicitly specify his CMC experience, Sun shared the various experiences she had with CMC prior to her enrollment at the ELI which included using CMC during team work. Based on specific features of

PAGE 503

477 CMC interactions, such as the use of emoticons, I assume that he had some experience with CMC as well. The Pre writing Interaction The involvement in the pre writing task of these two actors was similar: they were able to collaborate with their partners i n the process of the SCMC discussion. They demonstrated abilities to share their ideas about the topic. Both dyads, that of Sun and Kang, stayed online after they perceived their discussion to be completed and engaged in a chat not directly related to the topic. This suggested to me that they felt comfortable in the text on to the discussion in terms of number of postings were different: Kang posted 19 chat message out of 31 (61.29%), while Sun posted 41 times out of total 83 postings (49.40%). However, the content idea units related to the pre writing task were similar in number: Kang composed 14, while Sun composed 13 content idea units. Post treatment Paragraph Gains and Intertextual Connections The high gains demonstrated by Kang and Sun were a result of improving multiple text aspects in their post treatment writings However, these aspects were mainly different. Out of the five criteria on which they improved, only two overlapped, namely rhetorical soundness of the text and the text vocabulary complexity. Kang also improved lexical information per clause (measured by lexical density analysis), presentation and development of main ideas, and overall language use (the last three criteria were assessed using a multiple trait rubric). Sun, on the other hand, improved the syntactic complexity

PAGE 504

478 (measured by calculating the m ean length of t units), the amount of information present in a single focus (measured by mean length of idea units), and the quantity of overall information present (measured by the number of idea units). as different (Sun composed 29 idea units in her paragraph while Kang composed 12), the intertextual connections of their writing and the pre writing SCMC discussions shared common characteristics. Their paragraphs showed a low degree of intertextuality on a lexical level: the ratios of writing discussions were lower than the average SCMC treatment group ratio. I did not identify any intertextual connections between their pre writin g interactions and paragraphs on an organizational level either. Neither Kang nor Sun discussed paragraph organization with their partners. In addition, in their paragraphs they did not follow the order in which the ideas were posted during the pre writing discussion. With both actors, the use of idea units borrowed verbatim from the text of the pre writing discussion was limited or non existent. Kang borrowed one of his own idea units from the discussion, which constituted 8.33% of the overall paragraph text, while Sun did not use any of the pre writing discussion idea units verbatim. Further, they both rephrased idea units from the discussion. In his paragraph, Kang used four rephrased idea units, which constituted 33.33% of his writing. One of these rep hrased idea units was contributed by him and three by his partner. Sun used seven rephrased discussion idea on her own. Thus, both of them were able not only to collab orate during the discussion

PAGE 505

479 pre writing ideas into their writings. The strategies they used for rephrasing included the use of synonyms or synonymic phrases and syntac tic changes. In addition, Sun reshaped idea units from the discussion when rephrasing them in her writings. She extended and further developed idea units, as well as shortened and merged two idea units from the discussion into one in her writing. This abil further demonstrated in the manner in which they composed new idea units in the paragraph. Part of these new idea units were neither borrowed from the discussion nor rephrased but the actors developed topics that were stated in specific idea units in the discussion. Both Kang and Sun composed idea units based on chat topics which were stated in specific idea units. In five of her new idea units in the paragraph (17.24%), Sun developed topics proposed by her partner in the discussion. Kang composed three such Further, in their paragraphs, both actors developed idea units on topics addressed during the dis cussion but not directly related to a specific idea unit in the discussion. With both of the actors these were topics proposed by their partners. In her paragraph, Sun composed one new idea unit (3.45%) with more general topic connection with the topic be ing proposed by her partner, while Kang composed three such idea units (25%). Thus, based on these detected intertextual connections, I determined that on the idea unit level, the actors demonstrated a strong consideration and use of the pre writing discus sions and

PAGE 506

480 There was one considerable difference between the two actors in relation to the new paragraph idea units which showed no connection with pre writing discussions. such idea units, which constituted 51.72% of her paragraph, while Kang composed only one (8.33%) such idea unit. However, these new idea units were included mostly towards the end of the paragraph after Sun composed a considerable number of idea units rela ted to the discussion. This difference might be attributed not so much to a different level of discussion use during the paragraph stated in her interview that her t yping skills were very strong. While I do not have interview), based on the fact that the pre writing discussion in which he participated was much shorter, (although produc ed with the same time constraints), I would assume that his typing skills might have been weaker. This, in addition to a possible difference in composition skills, might be the reason for their different inclusion of idea units that were not related to the discussion. Sun might have had more time to think about and incorporate new ideas in order to further extend her paragraph. Answering the Research Questions How do peers participate in synchronous CMC interactions? I concluded that these two actors, who showed improvement on their post treatment writings, shared several common characteristics. During their pre writing interactions, they were able to engage in collaborative discussions showing consideration This could be related to a combination of factors: they were skilled computer users with some experience in CMC interactions,

PAGE 507

481 they also expressed somewhat positive attitudes towards writing and were able to define their goals, difficulties, and positive a spects related to writing in English. Further, they perceived writing as being an important medium of communication. How do they use the specific ideas and language generated during these interactions in their writings? What are the differences and simila rities in the implementation of these ideas? Both Kang and Sun considered the ideas that were generated during the although the intertextuality of their paragraphs with their pre writing discussions was low on the lexical level and non existent on the organizational level, the analysis on the idea unit level revealed medium to high intertextuality. They were able to consider and ideas into their post treatment writing. The differences between Sun and Kang in terms of the number of the new paragraph idea units not related to the discussion could be attributed to the different level of typing and composition skills of the two parti cipants. However, in my opinion in these two cases it was important to view this difference taking into account the fact that Kang and Sun used in their paragraphs the majority of the ideas shared during the discussion. SCMC Actors: Comparison The comparis on of the four SCMC case studies presented above revealed that there were both commonalities and differences between the actors who demonstrated high and low gains in their post treatment writings. In the current section of Chapter V, these differences and commonalities are presented in the context of the research questions which the case studies aimed to answer: How do peers participate in synchronous CMC

PAGE 508

482 interactions? How do they use the specific ideas and language generated during these interactions in t heir writings? What are the differences and similarities in the implementation of these ideas? The following trends considered in the data analysis were compared: (1) background, (2) the specifics of pre writing CMC interactions, and (3) post treatment par agraph gains and intertextual connections. The four participants who performed their pre writing discussion in a n SCMC environment were of somewhat different backgrounds. Three of the participants (Ajwad, Kang, and Sun) were in their mid twenties, close to the mean age of the students enrolled in the writing class during the two semesters in which the data were collected, while the fourth actor (Kamil) was much younger: 19 years old. This difference was also refl ected in their educational level. Ajwad and Sun had already obtained their university degrees. Kang, although still a university student, was studying English with specific career related goals in mind. On the other hand, Kamil, who had graduated from high school, was just starting his university studies and did not share any specific career related goals with regards to his English language studies. The older actors were able to define their neral negative attitude towards writing. In terms of their computer skills, all four were experienced computer users who had participated in computer mediated communication before the study. Sun indicated explicitly that she had used CMC for personal comm unication as well as study and job related tasks, while the other actors were not specific about the type of CMC experiences they had had before the semester in which data were collected.

PAGE 509

483 The Pre writing Interaction The pre writing interactions of Sun and Kang, the actors who demonstrated the highest positive gains in their post treatment writing, revealed the ability to collaborate with their partners during the task completion. Their pre writing interactions were extensive and addressed multiple aspects of the task. Their discussions involved numerous turns: Kang and his partner posted 31 messages, and Sun and her partner posted 83. Although their pre writing discussions differed in terms of total number of postings, both Sun and Kang contributed signific antly to their dyadic discussions in terms of number of postings: Kang posted 19 messages and Sun 41. Their contributions as measured based on content idea units were similar: Kang posted 14 content idea units and Sun 13. Although, in their discussions, bo th dyads explicitly indicated that they perceived the task to be completed, they failed to address paragraph organization during the discussion. In addition, both Sun and Kang remained online after they perceived the task to be completed and chatted about their personal matters for the remaining time allocated for the SCMC pre writing discussion. treatment writing revealed the lowest gain, I found similar trends in his pre writing SCMC discussion and the SCMC discussion of the two ac tors who had the highest gain. He was able to engage into a collaborative discussion with his partner, contributing to the development of ideas during the discussion. The discussion was extensive, involving 82 turns posted by Ajwad and his partner. Ajwad p osted 39 times during the discussion, contributing 20 content idea units. In addition, Ajwad and his writing discussion and the manner in which he participated in it shared similar characteristics with the pre writing

PAGE 510

484 discussions of the actors who demonstrated the highest gains in their post treatment writings. In their interviews, Ajwad and Sun shared that the task was helpful. Sun pointed out that although chat, in her opinion, would not support the development of writing skills, the opportunity to share opinions with her partner during the task helped her. Ajwad also pointed out that the task was helpful for him, allowing him to organize better his ideas for the paragraph. On the other hand, K amil, who was the second low gain SCMC actor, participated in his pre writing discussion differently. His contribution was limited and rather passive. The discussion itself was short: a total of 10 turns was posted by Kamil and his partner. In addition, Ka mil posted only three content idea units. He disengaged from the discussion early, showing disagreement with his partner and lack of motivation to complete the discussion. Later he shared in his interview that the task was boring, which I found to be relat ed to his overall view of writing as a boring task. From this comparison, it appears that collaborative dyadic interaction patterns were used both by the actors who gained the most on the post treatment writing as well as by one of the actors who gained t he least on the post treatment writing. However, passive patterns and early disengagement were demonstrated by the second actor who showed low gains. Thus, the contribution to the discussion was linked to the patterns of interaction rather than to the post treatment writing gains: the collaborative actors were the ones to contribute significantly to the discussion, while the actor who demonstrated a passive trend failed to contribute a significant amount of information to the discussion.

PAGE 511

485 Post treatment Pa ragraph Gains and Intertextual Connections In their post treatment paragraphs, the SCMC actors who demonstrated high and low gains improved different aspects of their writing. The low gain actors, Kamil and Ajwad, improved on two out of the eight aspects considered in the text analysis in this study, while the high gain actors, Sun and Kang, improved on five. Both Kamil and Ajwad improved the lexical information per clause of their post treatment paragraphs measured by lexical density analysis. In additio n, Kamil improved the presentation and development of his paragraph, while Ajwad improved the amount of information present in a single focus (measured by mean length of idea units). Sun and Kang improved their post treatment writing on two common text mea sures: vocabulary complexity and rhetorical soundness. Sun also improved on the syntactic complexity (measured by calculating the mean length of t units), the amount of information present in a single focus (measured by mean length of idea units), and the quantity of overall information treatment writing improved on lexical density, presentation and development and overall language use. st treatment paragraphs were a result of improving multiple aspects of their writing rather than improving significantly on one or two specific aspects, while the low gains of Ajwad and Kamil were a result of failing to improve on multiple aspects of their writing. Table 5.12 presents the textual aspects on which the actors who participated in the SCMC pre writing discussions improved in their post treatment paragraphs.

PAGE 512

486 Table 5.12 SCMC Actors: textual aspects improved Tex Analysis Criteria Low SCMC High SC MC Ajwad Kamil Kang Sun Syntactic complexity (T units mean length) X Information in a single focus (mean length of idea units) X X Overall information present (number of idea units) X Lexical information per clause (lexical density) X X X V ocabulary complexity (frequency of unique words) X X Rhetorical soundness X X Presentation and development of main ideas X X Overall language use X The intertextual connection analysis of all four SCMC case study actors revealed that these co nnections were weak when their pre writing discussion and the post treatment paragraphs were compared on a lexical level. These connections were also weak in terms of paragraph organization as related to the sequencing of discussion ideas and the organizat ion of discussed paragraph. However, there was a considerable difference in the intertextual connections of the high

PAGE 513

487 the four SCMC actors is presented in Table 5.13.

PAGE 514

488 Table 5.13 SCMC Actors: Paragraph Idea Units and the Level of Intertextual Connections Participant and Level Criteria 1 2 3 4 5 6 7 8 9 10 11 Total Low SCMC Ajwad Number of IUs 0 0 1 1 0 0 0 2 4 3 10 21 Paragraph % 0 0 4.76 4.76 0 0 0 9.52 19.05 14.29 47.62 100 Intertextuality High: 9.52% Medium: 0% Low: 42.86% None: 47.62% 100 Kamil Number of IUs 0 0 0 0 1 2 0 1 1 0 8 13 Paragraph % 0 0 0 0 7.69 15.38 0 7.69 7.69 0 61.55 100 Intertextuality High: 0% Medium: 30.76% Low: 15.38% None: 61.55% 100

PAGE 515

489 Table 5.13 (Continued) SCMC Actors: Paragraph Idea Units and the Level of Intertextual Connections Participant and Level Criteria 1 2 3 4 5 6 7 8 9 1 0 11 Total High SCMC Kang Number of IUs 0 1 0 3 2 1 0 3 1 0 1 12 Paragraph % 0 8.33 0 25 16.68 8.33 0 25 8.33 0 8.33 100 Intertextuality High: 33.33% Medium: 25% Low: 33.33% None: 8.33% 100 Sun Number of IUs 0 0 3 4 5 1 0 1 0 0 15 29 Paragraph % 0 0 10.34 13.79 17.24 3.45 0% 3.45 0 0 51.73 100 Intertextual ity High: 24.13% Medium: 20.69% Low: 3.45% None: 51.73% 100

PAGE 516

490 Table 5.13 (Continued) SCMC Actors: Paragraph Idea Units and the Level of Intertextual Connections Criteria Coding Scheme: 1. 2. Own verbatim 3. 4. Own rephrased 5. New 6. New based on own idea unit 7. 8. 9. New development of own topic 10. 11. New no traceab le connections with pre writing discussion

PAGE 517

491 As it is evident from Table 5.13, the two low gain actors, Ajwad and Kamil, composed a high number of new idea units not related to the discussion: Ajwad composed 10 (47.62%) new idea units out of 21 total para graph idea units, and Kamil composed 8 (61.55%) new idea units out of 13 total paragraph idea units. The idea units composed by Ajwad and Kamil, for which I identified intertextual connections, were mostly of either medium or low level of intertextual conn ection with the discussion. In level intertextual connections with the discussion, none of the idea units were of medium level, and 42.86% of the idea units had low intertextual connection with the d iscussion. Kamil composed three idea units (30.76%) with medium intertextual connections with the pre writing discussion, and two idea units (15.38%) with low connection. Thus, their paragraph writings and pre writing discussions showed medium to low inter textuality. In addition to the medium to low level of intertextuality, I also found that their one had a low intertextuality level. Ajwad showed a higher level of consideration of his constituted 14.29% of his paragraph. However, he also composed three idea units discussion at a low level of intertextuality. Thus, I would suggest that in addition to the med ium to low intertextual connections between the pre writing discussion and the post treatment paragraphs, the low gain students also showed lower consideration of their

PAGE 518

492 might collaborative mode of the pre writing discussion of his dyad. Sun and Kang, the two actors who had positive gains on their post treatment paragraphs, demonstrated higher ability to use ideas generated during the pre writing related to the dis paragraph idea units (33.33%), I found high level intertextual connections with the pre writing discussion: Kang used one of his own idea units verbatim and rephrased three. Further, three paragraph idea units (25%) showed a medium level of intertextual connection. Finally, in four paragraph idea units (33.33%), he developed topics mentioned in the discussion but these paragraph idea units were not related to specific idea units from the di scussion which showed intertextual connections of low level. Only discussion. Sun, also showed high consideration of the ideas produced during the discussion and included bo her idea units were new and not traceable to the pre writing discussion. Although, the writing she incorporated most of the ideas shared during the discussion. I found that units; thus, they revealed high level intertextual connections with the pre writing

PAGE 519

493 discuss ion. In her paragraph, Sun composed 20.69% idea units with medium level intertextual connections with the pre writing discussion. Only 3.45% were idea units of low intertextuality. Thus, taking into consideration that the new idea units which were not trac speculate that she considered and exhausted the ideas shared during the discussion which led her to the development of new ideas on the assigned topic. Three of the participant s paraphrased discussion idea units when incorporating them into their post treatment paragraphs. The paraphrasing techniques that they used were of various kinds. Ajwad and Kang mostly applied the use of synonyms and synonymic phrases as well as syntax ch anges, while Sun used a variety of techniques including development of idea units, as well as shortening and merging idea units from the discussion. Kamil did not paraphrase any of the discussion idea units. Answering the Research Questions How do peers pa rticipate in synchronous CMC interactions? Based on this between group comparison, I would suggest that the high and low gain actors demonstrated a number of differences and similarities in terms of their participation in the SCMC discussion. The actors w ho were motivated and were able to define a specific goal as well as specific difficulties and enjoyable aspect of writing in general and writing in English tended to engage in a collaborative discussion. Kamil, who expressed a negative attitude towards wr iting transferred this attitude towards the task: he was passive during the discussion and disengaged early from the task. Thus, the patterns of interaction and contribution to the pre writing task for these actors were not apparently

PAGE 520

494 related to their gain level but possibly to the way they perceived the task and the acquisition of writing skills in English and the task specifically. How do they use the specific ideas and language generated during these interactions in their writings? What are the differen ces and similarities in the implementation of these ideas? The intertextual connections on the lexical level for all SCMC actors regardless their gains were weak. The intertextuality on the organizational level was also weak for three of the actors; only A jwad (the low gain collaborative actor) showed a considerable intertextuality of his paragraph and the pre writing discussion on this level. The other actors neglected to discuss this aspect of the pre writing task. In terms of intertextuality on an idea unit level, the low and the high gain SCMC actors differed. Sun and Kang were able to establish higher level intertextual connections between their post treatment writing and the pre writing discussion; as a result, they may have benefited more from their pre writing discussion as compared to the low gain actors. While composing their paragraphs, they were also somewhat more considerate of gain participants. This was demonstrated by the paragraphs. However, Ajwad took a somewhat middle position in terms of considering attributed to his collaborat ive interaction patterns. All of the actors, except for one of the high gain actors, Kang, composed a considerable number of new idea units not directly traceable to the pre writing discussion. However, the low gain actors, Ajwad and Kamil, did this whil e neglecting

PAGE 521

495 some of the ideas shared in the discussion, while Sun, in addition to the new idea units, composed idea units based on almost all of the ideas shared during the discussion. ACMC and SCMC Actors: Comparison In the current section, the actors i n the eight case studies that were presented in this chapter are compared. They were selected based on their gains when their pre and the post treatment writings were compared. Four of the actors Felipa, Isabella, Azad, and Shin participated in the ACMC treatment group. Felipa and Isabella showed the highest gain among all Level III students (n=34) who participated in the pre writing discussion using ACMC mode of communication, while Azad and Shin showed the lowest gain. The four actors who participated in the SCMC pre writing discussion were Ajwad, Kamil, Kang, and Sun. Kang and Sun achieved the highest gain among all Level III students who participated in the pre writing task using SCMC mode of communication (n=26), while Ajwad and Kamil had the lowest gains in this group. their CMC interactions and the intertextual connections of their paragraphs and their pre writing discussion in order to answer the research question s proposed for the qualitative instrumental case study part of current research project: Research Question 1: How do peers participate in asynchronous and synchronous CMC interactions? Research Question 2: (a) How do they use the specific ideas and languag e generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas?

PAGE 522

496 The eight participants came from seven different countries: Felipa was from Colombia, Is abella from Italy, Kang and Shin were from Korea, Kamil was from Kuwait, Azad from Qatar, Ajwad from Saudi Arabia, and Sun was from Taiwan. Five of the actors, Isabella, Felipa, Ajwad, Kang, and Sun were in their mid to late twenties. Their age was close to the average age of the Level III students who took classes during the data collection period (mean age=25.5 years). Three of the actors were at different ages: Kamil and Azad were 19 years old, and Shin was in her late 30 s. In terms of their education al background, only Kamil had recently graduated from high school, while the other seven actors were either university students in their countries who interrupted their studies in order to improve their English language skills (Felipa, Azad, and Kang), or had already graduated with a university degree (Shin, Ajwad, Isabella, and Sun). Everyone, except for Shin, was an experienced computer user who had performed various tasks related to their everyday life using computers. Shin had limited computer skills. A lthough this was not her first semester at the ELI, and she had participated in the computer skills workshops offered at the ELI and had used computers during the computer laboratory sessions incorporated into the ELI curriculum, she still felt uncomfortab le and reluctant using computers. She also shared with me that computer mediated interaction made her uncomfortable, because of her limited computer use expertise. In addition, the CMC treatment interaction task made her feel uncomfortable because she was expected to post while in class she could remain silent. Further, all of the actors had used CMC prior to the period in which data were collected. However, based on my interviews with the actors, only two of them, Isabella

PAGE 523

497 and Sun, had previous experience s of using CMC for task solving purposes. Isabella had used computers, mainly email, to communicate with researchers and peers in her professional field in the process of writing her thesis. Sun had participated in team work related tasks using both synchr onous and asynchronous communication. These two actors were among the students who benefited from the pre writing CMC discussion improving various aspects of their writing. In their interviews with me, all actors, except for Kang, shared their educational backgrounds and their perceptions of writing. While all of the case study participants who showed positive gains were willing and able to define their perceptions of writing and were able to articulate the enjoyable and difficult aspects of writing in Eng lish, Azad and Kamil were reluctant to provide this information. Azad avoided answering the questions, admitting in an informal conversation with me that it was hard for him to formulate his opinions about writing, while Kamil provided a short answer point ing out that writing reluctance to articulate his writing experiences and Kamil boring task suggested rather negative attitude towards writing. Based on the background information, I would suggest that for three of the actors (Kamil, Shin, and Azad) the pre writing discussion followed by the paragraph wr iting was a difficult task, and they appeared not to have benefited from it. Their previous experiences were related to this difficulty. For Azad and Kamil these were the negative attitudes towards writing. In addition, for Shin and Kamil, the task involve d stepping out of their zone of comfort when required to participate actively in the dyadic pre writing

PAGE 524

498 interactions. Two of the students who benefitted from the task Shin and Isabella, seemed to be somewhat better prepared for the CMC pre writing discussi on due to their previous experiences with CMC that involved task solving interactions. The Pre writing Interaction The comparison of the pre writing dyadic interactions revealed that the actors applied different dyadic interaction patterns. This difference was not related to their gains; rather, they were based on the mode of interaction. All four actors, who participated in ACMC pre writing discussion, demonstrated dominant patterns, and the same patterns were demonstrated by their partners. These actors a nd their partners approached the pre writing task completion separately without engaging in a discussion of the topic. They simply posted their views about the topic as based on the pictures provided for discussion. Thus, there was no evidence of an attemp t to create a joint the exception of Kamil who was rather passive and disengaged from the task early without completing the discussion of the topic. In addition to the d yadic pattern differences, the ACMC and the SCMC interactions differed in terms of the number of the postings. The ACMC discussions tended to involve longer postings that resembled in their format and organization a paragraph or paragraph segments. Both hi gh and low gain ACMC actors posted much fewer times as compared to their SCMC counterparts; they mostly used complete sentences, providing lists of ideas related to the pictures they discussed. Most of the idea units the ACMC actors posted were directly r elated to the task topic. There were fewer non content idea units in which the actors would greet and address their partners, direct

PAGE 525

499 turn taking, and/or request more information or clarification. On the other hand, all of the SCMC interactions were more dy namic with a higher number of non content idea units exchanged between the dyad members. The collaborative SCMC dyads also posted a much higher number of turns; only the passive/dominant dyad in which Kamil participated had a number of turns comparable wit h the ACMC dyad that had the highest number of turns, that of Azad. However, I found that this similarity was triggered by the way in which Azad and his partner engaged in the ACMC interaction task: they used the ACMC application for synchronous exchange o f messages. Another aspect of the CMC interaction that was different when comparing the SCMC actors and the ACMC actors was the contribution to the task. For the ACMC actors, the post treatment gain was directly related to the amount of contributed content idea units. Thus, Isabella and Felipa, who posted significantly higher number of content idea units, had high gains, as compared to Azad and Shin, the low gain actors. In fact, the number of the content idea units of the high gain ACMC actors was comparab le to the number of idea units of the collaborative SCMC actors. For the SCMC participants, the contribution to the task was only partially related to their gains; it, rather, reflected the mode of interaction. As compared to the low gain ACMC actors and t he SCMC actor who was passive during the pre writing task completion, the collaborative SCMC actors and the ACMC actors who had high gains contributed significantly to the discussion in terms of idea units. Post treatment Paragraph Gains and Intertextual Connections The analysis of the intertextual connections of the pre writing CMC interactions and the post treatment paragraph aimed to reveal how the participants used their pre

PAGE 526

500 writing discussions when composing their paragraphs. I considered three diffe rent aspects of intertextuality: (1) lexical level intertextuality, (2) intertextuality on paragraph organization level, and (3) intertextuality on an idea unit level. The analysis of the lexical level intertextuality was performed by matching distinct lex ical items. Using the software designed for the current study, the lexical items in each CMC interaction were identified and matched against the post treatment paragraphs and the ratio of matching distinct lexical items to the overall distinct lexical item s from the CMC interaction was calculated which allowed obtaining the matching distinct lexical items (DLI) ratio score. Further, this score allowed me to suggest the degree to which the post treatment paragraphs were related to the pre writing discussions on a lexical level. The detailed descriptions of the procedures related to the score calculation are presented in Chapter IV. The comparison of the matching DLI scores of the actors in the instrumental case studies to the average scores of the ACMC and th e SCMC treatment groups that participated in the quantitative stage of the study revealed that all actors had lower than average intertextual connection on a lexical level with the exceptions of the two high gain ACMC actors; however, the high lexical inte rtextuality writing discussion was evident only when just her posting was considered. On the organizational level, I found that the consideration of the organization proposed during the discussion was not necessarily followe d by the actors in their paragraphs. Four actors mentioned paragraph organization during their pre writing interaction: Shin (a low gain ACMC actor), Isabella and Felipa (high gain ACMC actors), and Ajwad (a low gain SCMC actor). Isabella followed her part

PAGE 527

501 organization suggestions, while the other three actors used loosely the organization suggestions shared during the pre writing discussion. The rest of the actors did not discuss the organization of their paragraphs, and there was no evidence that the y followed in their paragraphs the sequence in which the ideas were posted during their pre writing discussion. Thus, I could not identify a trend between the participants in terms of the differences in intertextuality on organizational level when the CMC mode and gains were considered; however, I could suggest that, in general, the intertextual connections on this level are rather weak. The high level intertextual connections, which involved verbatim use of idea units from the discussion, were more evident in the paragraphs composed by the ACMC actors. Each of the two low gain actors used word for units. Felipa, one of the high gain ACMC actors used verbatim three of her own idea units, while the other high gain actor, Isabe lla, did not use any. Only one SCMC actor, Kang, used one of his own idea units from the pre writing discussion without incorporating any changes. Thus, I could conclude that the ACMC actors tended to use idea units borrowed verbatim from their discussions I also considered paraphrased idea units to show a high level of intertextual connection. The comparison of the paraphrased idea units of the ACMC participants revealed that both high gain participants and low gain participants used in their paragraphs e writing discussion after rephrasing them. The low gain actors applied the use of synonyms, syntactic changes, clipping idea units, and content changes within an idea unit. In addition to the rephra sing strategies used by the low gain actors, the high gain actors extended idea units

PAGE 528

502 borrowed from the discussion, added organizational words and phrases, developed two idea units based on one, or combined two idea units into one. Thus, I would suggest th at the low gain ACMC actors tended to apply the changes within the boundaries of the idea units borrowed from the pre writing discussion, while the high gain ACMC actors were able to move beyond the single idea unit boundaries. The paragraphs of the SCMC actors showed more limited high level intertextual connections that involved idea unit rephrasing, as compared to the paragraphs of the ACMC actors. In addition, the low gain SCMC actors used paraphrased discussion idea units less than the high gain ones. They mostly applied the use of synonyms and synonymic phrases. The use of paraphrased idea units from the discussion by the SCMC high gain actors was comparable to the low gain ACMC actors and much lower as compared to the ACMC high gain actors. Only Sun a pplied rephrasing techniques similar to those of the high gain ACMC group techniques which allowed her to move beyond the single idea unit boundaries. Thus, the ACMC actors showed overall stronger intertextual connections of a high intertextuality level wh en considering the use of both verbatim and rephrased idea units. In addition, the high gain actors, regardless the mode of CMC pre writing communication, were more likely to move beyond the single idea unit boundaries when rephrasing idea units borrowed f rom the discussion. Finally, the high gain ACMC actors were also able to implement grammar correction to the idea units they were rephrasing. When analyzing the intertextual connections on a medium level, I considered paragraph idea units whose topic could be traced to a specific idea unit or units from the discussion but the changes made were beyond paraphrasing. It seems that the ACMC

PAGE 529

503 actors were less likely to establish intertextual connections on this level. Only Shin composed several paragraph idea uni ts at this level of intertextual connection, while Felipa had only one such idea unit in her paragraph, and Azad and Isabella had none. On the other hand, the SCMC group participants were more likely to establish intertextual connection on a medium level b etween their pre writing discussions and their paragraphs. Only Ajwad did not have any idea units in his paragraph that would reveal this level of intertextuality; the rest of the SCMC participants based a considerable number of their paragraph idea units on topics traceable to idea units from the discussion. The low level of intertextuality showed consideration of a topic mentioned during the discussion with no explicit relation between specific paragraph and discussion idea units. At this level of intert extuality, I could not trace any consistent patterns among the actors based on the mode of their interaction and their gain level. In each of the mode/gain pairs, one of the actors revealed stronger intertextual connections, and one showed weak or no low l evel intertextual connections. With regards to the new idea units composed in the paragraph which do not reveal intertextual connections with the discussions, I also could not find a trend. Most of the actors composed a significant number of idea units tha t developed the topic of the paragraph but were not related to the discussion. One low gain ACMC actor, Azad, did not compose any paragraph idea units not related to the discussion and one of the high gain SCMC actors, Kang, composed one such idea unit. Th us, I would suggest that most of the actors, regardless to their gains and mode of pre writing interaction, tended to develop further their paragraphs including idea units that presented topics related to the overall paragraph topic but not mentioned in th e discussion.

PAGE 530

504 shared during the discussion. Based on the comparison across the groups, I could not find a consistent pattern. All of the actors, with the exception of Felipa (a high gain ACMC patterns of interaction were taken into account for the comparison rather than the gain level, it appeared that the collaborative actors tended to includ idea units or topics than the passive actor. Answering the Research Questions Research Question 1: How do peers participate in synchronous and asynchrounous CMC interactions? nteractions, I would propose that the CMC mode of interaction affected strongly the nature of dyadic patterns of interaction. The ACMC actors, regardless of their post treatment gains, were mostly concerned with their own postings, demonstrating dominant p atterns of interaction. On the other hand, the SCMC actors were more likely to participate in a collaborative mode of interaction. The collaborative dyadic patterns they demonstrated during the interactions were not necessarily related to their post treatm ent gains; however, the passive SCMC actor had a low treatment gain. The mode of interaction may have influenced the number of postings as well. However, when the postings were analyzed on an idea unit level, considering content idea units, it appeared tha t the high gain ACMC actors and the collaborative SCMC actors, regardless of their gain, contributed significantly to the discussion, while the low

PAGE 531

505 gain ACMC actors and the passive SCMC actor failed to contribute significantly to the interaction on the ass igned topic. Research Question 2: (a) How do they use the specific ideas and language generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas? In order to answer this resear ch question, I considered three levels of intertextuality: lexical, organizational, and idea unit levels. The high gain ACMC actors demonstrated higher level of lexical intertextuality as compared with the rest of the actors. On organizational level, most of the actors, regardless their mode and gain level showed no intertextual connections. Only two ACMC actors considered partially the paragraph organization proposed during the discussion. In addition, organization was more likely to be discussed by ACMC a ctors. With regard to the intertextual connections between the pre writing discussion and the post treatment paragraph, when idea units were used as a unit of analysis, I found several different trends related to the level of intertextual connections and consideration high level intertextual connections by borrowing verbatim or rephrasing idea units from the discussion; (2) Medium level intertextual connections were more li kely to be gain SCMC participants were more likely to establish high and medium levels of intertextual connections, as compared to the low gain participants who participated in the same CMC mode; (4) Hig h gain actors, regardless to the CMC mode, were more likely to move beyond the single idea unit boundaries when paraphrasing; (5) High gain ACMC actors were able to implement

PAGE 532

506 some grammar correction into the idea units when rephrasing them; (6) Collaborati ve actors, who were all SCMC actors, were able to consider and include in their paragraphs cussion to a higher extent, as compared to the low gain collaborative SCMC actor; and (8) Most of the actors developed their paragraphs beyond the topics discussed during the pre writing interactions. Figure 5.1 presents the above findings in a graphic for mat.

PAGE 533

507 Figure 5.1 Instrumental Case Study Findings

PAGE 534

508 Summary of Findings The communication environment as well as the patterns of dyadic interaction influenced the communication that took place during the pre writing stage and this influence could be traced to the first draft writings of these eight extreme case participants in the study. The immediacy of peer responses during the pre writing discussion in SCMC environment allowed for creating joint problem solving space, while the time and place ind ependence of the ACMC interaction isolated the dyad members from one another. Further, the findings with regards to the pre writing communication and first draft outcomes are summarized. First, there was a difference in the nature of the postings produced in the two communication environments. The ACMC mode prompted the participants to post longer discussions were completed in short interactive postings which contained mainly phrases and syntactically incomplete utterances. Second, the degree of involvement in the pre writing task in terms of information contributed was related to the communication environment, the gain level, and the dyadic pattern of interaction. The higher i nvolvement and the collaboration during the task completion might have supported the first draft writing outcomes. In addition, the task involvement might have been influenced negatively due to the challenges of the text based communication environment. Third, the intertextual connections between the pre writing postings and first drafts were evident for both the ACMC and the SCMC actors. However, the degree of intertextuality on the lexical, idea unit, and organizational levels differed. Some of the

PAGE 535

509 aspe cts of the intertextual connections were influenced by the environment in which the pre writing task completion took place, other by the gain level of the participants. On a lexical level, the intertextual connections were strongest with the writings of th e actors who showed improvement of their writing after completing the pre writing task in an ACMC mode. The CMC mode also influenced the intertextuality on an idea unit level. When the pre writing task was completed in an ACMC mode, the participants tended to apply less rephrasing to the idea units borrowed from the discussion as compared to the SCMC participants. However, the mode of communication did not affect more complex rephrasing when idea units from the discussion were merged or developed into more than one idea units; these changes were more likely to be applied by the participants who had high writing gains. The collaborative dyadic pattern of the discussion (evident only with to the pre writing discussion when the first draft was composed. The intertextuality on the organizational level was weak or non existent which suggests that the complexity of the pre writing task prevented most of the actors to attend to this aspect of the task as well as that they may benefit from multiple cycles of the pre writing task. The specifics of the organizational intertextuality also suggest that, in the examined cases, the text organization was a fluid aspect of the writing process and writer s changed planned text organization in their first drafts. Finally, the results of the analysis suggest that regardless to the mode of communication in which the pre writing task was completed and the gain level of the actors, new ideas were incorporated into the first drafts. Thus, during the pre writing stage, only a general blueprint of ideas was created. All actors further refined and

PAGE 536

510 extended the outcomes of the pre writing discussion through the addition of more details or new interpretation of the p re wring ideas. Chapter Summary This chapter presented eight case studies which actors were selected using Extreme/Deviant sampling technique (Kemper et al., 2003). The presented analysis and the comparison of these case studies aimed to reveal the dynami cs of two specific stages of the writing process accounting for the social context within which they took place. The writing interaction, dyadic pattern of interaction, and gain level. The analysis part of the chapter was concluded with a section in which a comparison across gains and CMC groups was performed. Further in the chapter, the outcomes of the study were outlined answering the two proposed research questions. The chapte r was concluded by a section in which the findings were summarized and discussed. In the following chapter, the summary of the overall study findings and their discussion will be addressed, along with implications for further research and pedagogical recom mendations.

PAGE 537

511 CHAPTER VI: SUMMARY, DISCUSSION, RECOMMENDATIONS, AND LIMITATIONS Introduction The theoretical framework of the current study is the Writing as a Process Approach (Grabe & Kaplan, 1996). The study addressed two specific stages of the wri ting process, accounting for the social contexts within which they took place. Thus, writing and first draft stage (Grabe & Kaplan, 1996), were observed and analyzed. The focu s of analysis was the social environment, including the learning task, peer interaction, mode of communication, and the intertextual connections between pre writing discussions of the participants and their first drafts. Tables summarizing the qualitative and quantitative findings along with the research questions are provided in Appendices 20 22. Summary of Qualitative and Quantitative Stage Findings The data analysis and findings presented in Chapter IV: Pre Case Study Analyses and Results aimed to ans wer a set of qualitative and quantitative research questions. The following qualitative research questions were addressed: (1) What are the skills? (2) What patterns of d yadic interaction do participants manifest during the asynchronous and synchronous CMC interaction process? (3) What are the factors that

PAGE 538

512 influence the CMC pre writing interaction process? How do these factors influence the interaction process? With regar ds to Research Question 1 the data analysis suggested that the participants held various views about the role of the CMC pre writing task: these views encompassed both positive and negative perceptions of the role of CMC and are summarized and discussed f urther in this chapter. Answering Research Question 2 it could be stated that the participants manifested patterns of interaction similar to those that were observed in the face to face environment and also reported by Storch (2002). However, emergent pat terns of interaction were identified for the CMC environments; more specifically these were mutuality mismatch and task disengagement. Further, answering Research Question 3, the data analysis suggested that these patterns of interaction were strongly infl uenced by the CMC environment in which the prewriting task was completed. The prior task solving experiences in CMC environments might have influenced the way students perceive and complete discussion tasks in such environments. Finally, the participants who had limited computer skills performed successfully in CMC tasks when computers were used on a regular basis in the classroom and additional support was provided. The quantitative stage of the study aimed to answer the following research questions: (1 ) What is the difference in the syntactic complexity present in the post treatment paragraphs of the students who participated in the synchronous versus asynchronous CMC prewriting discussion? (2) What is the difference in the amount of information present within a single focus of the post treatment paragraphs of students who participated in synchronous versus asynchronous CMC prewriting discussion? (3) What

PAGE 539

513 is the difference in the quantity of the overall information present in the post treatment paragraph s of the students who participated in synchronous versus asynchronous CMC prewriting discussion? (4) What is the difference in the lexical information present in the post treatment paragraphs of the students who participated in synchronous versus asynchron ous CMC prewriting discussion? (5) What is the difference in the vocabulary complexity present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC prewriting discussion? (6) What is the difference in the rhetorical soundness present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC prewriting discussion (7) What is the difference in the presentation and development of the main point and its support pr esent in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC prewriting discussion? (8) What is the difference in the overall language use present in the post treatment paragraphs of the students who partic ipated in synchronous versus asynchronous CMC prewriting discussion? The data analysis of the proposed eight textual features revealed that not a single one emerged as statistically significant draft writings of the ACMC with the SCMC groups. In other words, no differences could be found for any of the proposed textual measures when the writing outcomes of the ACMC and SCMC participants were compared. Finally, an Additional Research Question was proposed: Do the ACMC and the SCMC modes of the pre writing discussion influence to a different degree the texts produced after those discussions as measured by matching distinct lexical items? This research question was p rompted by the differences in the language production in the two

PAGE 540

514 CMC environments which were revealed by the qualitative data analysis. The researcher aimed to investigate further the influences of the interactions performed during the pre writing stage on the first drafts. The results of the data analysis revealed that there was a statistically significant ritings and their pre writing discussions was stronger for the post treatment (first draft) writings of the ACMC group participants as compared to the writings of the participants in the SCMC group. In the following sections, the above findings are present ed in detail and discussed. The Qualitative Stage Findings: Discussion writing Task In the interviews conducted during the study, most of the students shared a perception of the pre writing task preformed i n a CMC environment as being supportive for their first draft composition. They specified that the task provided them with the opportunity to share ideas, understand better the topic of the written assignment, refer to the discussion during the composition of their first drafts, notice their mistakes when reviewing the pre draft. These findings add to previous research of online discussion which specifies that such discussion provide s a non threatening space in which social interaction could support language learning and effective literacy skills practices (Ware & Warschauer, 2006). However, some participants explicitly stated that, in their writings produced after the discussion, the y used only their own ideas; interestingly, the majority of these students completed their pre wring discussions in an ACMC environment and

PAGE 541

515 demonstrated a dominant pattern of interaction. Thus, it appears that the specific characteristics of the CMC envir onment in which the interaction takes place, namely the mode of the CMC interaction along with the dyadic pattern of interaction in which the writing task. views on the writing process itself may have influenced the way they engaged in the pre writing task. Some of the participants shared negative views of the pre writing task that were not related to the CMC environment in which the pre writing task was per formed but were rather related to peer collaboration promoted in the pre as more valuable than peer opinion during the pre writing stage. Secondly, some of the participan ts viewed writing as an individual activity and did not appreciate the requirement of the pre writing task that they share their brainstorming and consider the ideas of their peers during the pre writing stage. These findings could inform the pedagogical p ractices that involve the completion of collaborative writing tasks in a CMC environment. The pedagogical implications are discussed further in this chapter. As stated above, not all students shared positive opinions about the pre writing task, and all of these participants demonstrated a dominant pattern of dyadic interaction; in other words, they failed to create a common discussion space with their partners. writing interactions was the o nly factor influencing these interactions, further in this section the additional factors that influenced the CMC interactions are presented and discussed. Although the number of students who shared such concerns and attitudes towards the task was relative ly low, it is important to outline the reasons which led them to their unfavorable

PAGE 542

516 as an individual process, there were concerns related to the CMC environment in whi ch the pre writing task was performed. Most of these participants felt uncomfortable during the discussion due to its text based format. These difficulties were related to both developing language and computer skills. The following specific difficulties re lated to the text based format of the communication were revealed: (a) the absence of non verbal cues to stimulate the formulation of ideas, (b) the great amount of time needed to express ideas in writing as compared to oral communication, (c) the limitati ons of text based discussions posed by underdeveloped typing skills, and, finally, (d) the preference of using hand research which interprets new literacies as a complex collection of skills that shape contemporary literacy practices (Simpson, 2005; Warschauer, 1999) and call for careful planning of CMC use in learning environments while supporting the development of new literacy skills. CMC Mode Influences on Pre writing Interacti ons Language use The pre writing interaction data revealed different language production in the two modes of communication. The ACMC interactions resulted in fewer postings (M=5), but longer strings of words between end of sentence punctuation marks (M=12 .66), which were often organized in complete sentences. The SCMC discussions, on the other hand, consisted of dynamic and multiple postings (M=46.46), much higher in number than the postings of the ACMC group. The number of words used in the two modes of C MC also differed; the SCMC group participants produced a higher mean number of words (M=410.46) and a higher mean number of distinct lexical items

PAGE 543

517 (M=92.15) as compared to the ACMC participants with means respectively 272.76 and 66.23. (The descriptive sta tistics of the pre writing discussions are presented in Table 4.1). It could be argued that because the ACMC participants were more likely to post messages which tended to emulate a paragraph on the assigned topic, they might have used vocabulary strictly postings which resembled a conversation, may have resulted in multiple turns related to the task completion process rather than to the specific topic assigned for discussion. Although in the cur rent research project, the nature of the lexis used in the interactions was not further analyzed in terms of its relations to the discussion topic, these findings could be interpreted in light of the findings reported by Smith (2003). He found that about o ne third of the entire SCMC task based interaction of language learners was devoted to negotiation directly related to the task while the rest of the interaction exchange was directed towards collaborative progression through the communicative task. This s tandpoint was further supported by the findings of the quantitative stage of the study (discussed later in this chapter). Future research that compares task based interactions performed in the two CMC environments would shed more light on the particular i ssue of the relation of lexical connections established in ACMC and SCMC environments with the task topic. The language use differences presented above suggest that the participants engaged in the task differently. It could be argued that the participants were completing the task in two different ways: through collaborative idea generation (promoted to a stronger extent by the SCMC environment) and through individual idea organization and

PAGE 544

518 refining (presented more visibly in the ACMC environment), possibly, but not necessarily, followed by peer feedback. Thus, two sub stages of the pre writing stage could be proposed: an initial sub stage of peer idea generation followed by a more advanced sub stage of individual idea refining and organization. Patterns of dyadic interaction The patterns of dyadic interaction were influenced by the CMC environment as well. When the pre writing was completed in an SCMC mode, the interactions were more collaborative in nature, as compared to the ACMC interactions. When perform ing the pre writing task in an SCMC environment, the discussion and attempted to interpret together the images related to the task. They were more likely to demonstrate high degree of mutuality and equality. There were more instances of dominant patterns when the pre writing discussion took place in an ACMC environment; thus, the mutuality of the ACMC dyadic interactions was lower (see Table 4.2 for the distribution of the dya dic interaction patterns in the two groups). In addition, the CMC environment prompted a dyadic pattern mismatch when one of the participants would engage in the discussion assuming a high level of mutuality, while the other would assume a low level of mu tuality. In other words, the latter would be concerned mainly with his or her own ideas while the former would show stronger intention to share and discuss ideas related to the task. Based on the data, it seems that the learners were more likely to demonst rate such a mismatch in an ACMC environment. Finally, some of the ACMC participants preliminar il y disengaged from the task after completing their posting. They did not revisit the discussion space to post additional messages or reply to the ones posted by their partners. This suggest s that they were

PAGE 545

519 focusing on task completion through posting and answering the task requirements rather than through the creation of collaborative discussion on the topic. Interestingly, no such disengagement was observed in the SCMC dyads. The differences in the dyadic patterns of interaction, the existence of pattern mutuality mismatch, and the instances of task disengagement during CMC interactions reveal additional dimensions of the interactions performed in ACMC and SCMC en vironments. It could be argued that the dominance of the interactions, the mismatch, and task disengagement observed mainly in the ACMC interaction might have been caused by the text based nature of the task and further enforced by the space and time inde pendence of ACMC. In this environment, the participants engaged in a time delayed expense of task completion. Thus, as Sotillo (2000) suggests, the ACMC mode prompted in teractions that resembled more teacher request and student response type of task, the ACMC dynamics shifted to higher consideration of task requirements rather than peer response. Further, it could be argued that the immediacy of the SCMC interactions and their resemblance to face to face conversations pointed out in previous research (Jepson, 2005; Smith, 2003; Sotillo, 2000; Warschauer, 1996) may have promoted the highe r collaboration and consistency of mutuality of the SCMC interactions. research of face to face interactions (Storch, 2002). However, such mismatch was identified in seven of the dyads, two SCMC and five ACMC. This mismatch could be

PAGE 546

520 promotes less negotiation of meaning and repair moves during the language learning task as compared to voice c hat. It seems that in ACMC the level of negotiation is even lower which resulted in more instances of mutuality mismatch. However, considering that the research comparing the dyadic patterns of interaction in ACMC and SCMC environments is limited, it coul d be expected that future research would shed more light on pattern specifics in such environments and would provide further interpretation on the reported findings. CMC Task oriented Experiences The interviews with the participants revealed that the m ajority of them had some, in many cases extensive, experience with CMC. These findings were not surprising. Luke (2006, cited in Miller, 2007) noted that today CMC tools are well known by younger generation learners. Considering that the mean age of the st udents who participated in this study was 25.55, it could be concluded that they belong to this younger generation of learners. This was also confirmed by the fact that only three out of the 60 participants, all three of whom in their thirties and forties, had limited exposure to computers prior to enrollment in the program and thus entered the task training period of the study with limited computer literacy skills. However, based on the interview data, it was concluded that most of the participants in the study used CMC exclusively for recreational and personal purposes. Only three out of the 60 participants shared in their interviews that they had engaged in CMC communications related to problem solving professional or academic tasks prior to their enrollm ent in the intensive English language program. Similar findings were reported by Jin (2007).

PAGE 547

521 The data collected in this study were not enough to make any conclusions about how this lack of problem solving CMC experience may have affected the CMC pre writin g interactions. However, the predominant lack of such experience in the targeted population of ESL learners and its effect on CMC language learning task attitudes and participation calls for future research. Taking into account that one of the main goals o f intensive ESL programs is to prepare international students for study in U.S. universities in which computers are an integral part of the curriculum, targeting the development of problem solving skills in text based CMC environments should be seriously c onsidered in curriculum development. Thus, findings of such research would inform the curriculum design and pedagogical practices of these programs. Limited Computer Skills Participants As stated in the previous section of this chapter, it could be sugge sted that the population of learners who enroll in intensive language programs in general may have acquired computer literacy prior to their enrollment in such programs. Most of the participants in this study who were perceived to be a representative sampl e of intensive and used computers on a daily basis for various purposes, including communication. However, as the qualitative data analysis showed, in this population, there might be learners who have limited or no computer skills. In the current study, three of the participants had considerably more limited computer skills as compared to the majority of the students who participated in the study. All of them were older than the majority of the participants, being in their late thirties or early to mid forties. They benefited significantly from the consistent use of computers during the language learning process

PAGE 548

522 and the extracurricular support offered for their computer skill development in the form of workshops. All three expressed overall positive attitudes towards the task and two of the participants demonstrated ability to engage in a collaborative ACMC discussion during the pre writing task completion. It could be ar gued that the systematic exposure to various computer applications including the CMC applications used in the study prior to the pre writing allowed them to gain skills that supported their successful performance in a CMC environment. Thus, as stated in th e previous section, computer skills development calls for careful consideration in the intensive ESL programs curriculum design and pedagogical practices. The Quantitative Stage Findings: Discussion ACMC and SCMC Groups: Comparison of Textual Features The following textual features were considered for statistical analysis when the first drafts of the participants who performed the pre writing tasks in ACMC and SCMC mode were compared: (1) the syntactic complexity (measured by calculating the mean length of t units), (2) the amount of information present in a single focus (measured by mean length of idea units), (3) the quantity of overall information present (measured by the number of idea units) (4) lexical information per clause (measured through lexical density analysis), (5) vocabulary complexity (measured by analyzing the frequency of the unique words used), (6) rhetorical soundness, (7) presentation and development of main ideas, and (8) overall language use (the last three criteria were assessed usin g a multiple trait rubric). These features addressed textual aspects of surface and deep level (Grabe & Kaplan, 1996). The ANCOVA analysis of these textual features, with alpha = .05, showed that there was no difference between the two groups. It is possib le that the

PAGE 549

523 differences in the proposed textual features that could be contributed to the dissimilarity of the pre writing mode of communication were too subtle to be detected considering the small sample size of the study. In addition, there is a possibil ity that the students used their pre writing discussion logs when composing their first drafts to a different extent and in a different manner. Possibly a study that involves a higher number of participants and/or addresses the strategies related to the us e of their pre writing discussions during the first draft composition could reveal more insight into the matter. However, the differences between the pre writing discussions in the two CMC environments call for further examination of the influence of the p re writing task on the first drafts. Because of the lack of possibility to obtain data from a larger sample or collect post treatment writings. Thus, the intertext ual relations between the pre writing discussion and the first draft were addressed. This examination was performed through a quantitative analysis of the intertextual connections on a lexical level as well as through conducting eight Extreme/Deviant instr umental case studies aimed at providing insight into the intertextual relationship on various levels. Comparison of Lexical Intertextual Connections The influence of the pre wring discussion on the first paragraphs was statistically analyzed applying a t test. This analysis revealed that there was a statistically significant writing discussion and first draft lexical intertextual connections. The participants who conducted their pre writing discussions in an ACMC mode were more likely to use lexical items from the discussion as compared to the SCMC participants.

PAGE 550

524 These results could be interpreted in the light of the Qualitative stage findings reported earlier. The extent to which the pre drafts seem to be grounded in the specifics of the language produced during the pre writing stage in both CMC environments. The ACMC interactions were task rather than interaction oriented; they resembled more advanced steps of the writing process such as outlining and first draft production. This resemblance could have made it easier for the participants to incorporate the vocabulary of their ACMC discussions into their first drafts. Further taking into account the high interactivity of the SCMC postings revealed in the qualitative stage and the findings reported by Smith (2003) that the high interactivity may pose limitations on task focused SCMC discussions, it could be argued that this hi gher interactivity of the SCMC discussions influenced negatively the intertextual connections on a lexical level. These connections were restricted by the production of fewer vocabulary items that were inherent to the task and thus could be readily used in the first drafts. However, the researcher realizes that such influences may not affect all participants and might vary with different levels on intertextual connections. Thus, in order to reveal how the CMC discussions conducted during the pre writing st age influence the first draft writing, the intertextual relationships were examined further in a qualitative manner via eight instrumental case studies. In these case studies, a line by line comparison of the CMC pre writing interactions and post treatment paragraphs of selected participants was performed. The summary and the discussion of the findings of these case studies are presented in the following section of the current chapter.

PAGE 551

525 Instrumental Case Study Findings: Summary and Discussion The eight parti cipants in the instrumental case study were selected from the Qualitative/Quantitative stage participants using purposive Extreme/Deviant sampling technique (Kemper et al., 2003). Thus, participants who showed the highest and the lowest gains on the first drafts they composed after the CMC discussions were selected. The specific research questions addressed in the case studies were as follows: Research Question 1: How do peers participate in asynchronous and synchronous pre writing CMC interactions? Researc h Question 2: (a) How do they use the specific ideas and language generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas? Participation in Pre writing Discussion The analy sis of the eight case studies presented revealed that the learning context within which the assigned task was situated and the environment in which the pre writing interactions took place influenced multiple aspects of the pre writing process. The particip ants were influenced by the mode of communication. The actors who participated in the pre writing interaction in an ACMC environment, regardless of their gains, failed to create a common discussion space with their partners. Their participation in the task was somewhat one sided as they appeared to be more concerned with their own postings and failed to respond to each other in a collaborative manner or failed to respond at all. On the other hand, the postings of the actors who participated in the pre writ ing task using SCMC were more interactive. This higher interactive mode of discussion was evident with all four of the SCMC actors, even with the one who demonstrated a dominant dyadic pattern of interaction and preliminarily disengaged from the discussion

PAGE 552

526 These observations of the influence of the environment within which the communication ons of face to face language learner interactions (Storch 2002), namely that the dyadic patterns of interaction were predominantly collaborative, were not transferable to an ACMC environment in these particular cases; however, the tendency for collaborativ e patterns of interaction was observable in the cases in which the interaction took place in an SCMC environment. This influence of the mode of communication could be attributed to the immediacy of the SCMC interaction and time and space distance of the A CMC interaction. All actors who engaged in the pre writing task in an SCMC environment, even the one who disengaged early from the task, posted multiple short messages and participated in the discussion, interactively recognizing the presence of their par tners. On the other hand, the high gain ACMC actors addressed the task in one long message after which they and their partners did not resume their postings. In addition, the low gain ACMC actors exchanged several shorter messages (however, still longer an d fewer in number than the messages of the SCMC actors), but, despite the exchange of multiple messages, The communication environment was n ot the only aspect of the pre writing interaction that influenced the discussion. The amount of information that the actors contributed to the discussion was connected to the dyadic interaction pattern for the SCMC actors and to the writing gain for the AC MC actors. The high gain ACMC and

PAGE 553

527 SCMC actors were the ones who contributed more, as compared to the low gain actors with the exception of the collaborative low gain SCMC actor, who contributed a comparable amount of information to the high gain actors. Th us, the high involvement in the pre writing task in terms of information contributed during task completion might have supported the first draft writing outcomes of the high gain participants. On the other hand, the SCMC mode supported the creation of comm on discussion space and provided opportunities for more active and collaborative discussion exchange of both high and low gains actors. It seems that in the SCMC environment, once the actors were involved in (as opposed to disengaged from) the task, the immediacy of the communication supported and answered. It also allowed for resolving misunderstandings and problems that arose during their communication. The students we re able to control the task and to collaborate guiding each other through the task completion. Similar findings regarding the dynamics of SCMC were reported by Sotillo (2000) and Shin (2006). On the other hand, the time and place independence of the ACMC exchange may have impeded the interactive, and thus collaborative, manner of task completion. In addition, the time and place independence of the ACMC interaction allowed the high gain ACMC actors to attain in more detail to the task; they were able to ad dress multiple aspects of the task and reach for additional support during the task completion such as dictionaries and help from other individuals. However, once the high gain actors sent their messages, they seemed to perceive the pre writing task to be completed. Wang (1996) reports similar findings in a study of e mail journaling in an ESL classroom.

PAGE 554

528 The nature of the pre writing task called for collaborative brainstorming and and their partners to engage into a discussion failed. This failure might be attributed to neglect of challenges of the text based communication environment, such as slower reading and writing (including typing) in the target language and lack of non verbal communication cues combined with the time delay of the interaction, might have influenced negatively the ACMC pre writing discussions. The Influence of the Pre writing D iscussion on the First Drafts: Differences and Similarities The intertextuality approach adopted in this study was based on the view that all that addressing inte process could support the understanding of the text composition development of language learners Further, with regard to collaboration, whose role is perceived to be significant in the wr iting process (Ferris, 2003; Liu & Sadler, 2003; Abbott, 1989; Ferris & Hedgcock, 1998), the results of intertextual analysis might reveal more specifics of the intertextual connections established during the writing process. Thus, the analysis of the inte rtextuality between the pre writing discussions and first drafts would shed light on the progression of the writing process from the pre writing stage to the first draft. The current study disclosed that participants differed with respect to the degree to which they manifested intertextual connections on the lexical, idea unit, and organizational levels.

PAGE 555

529 The ACMC high gain actors demonstrated stronger lexical intertextuality, as compared to the low gain ACMC actors and the SCMC actors, regardless of their gains. The stronger intertextual connections that the high gain actors established in their writings could be attributed to the detailed nature of their pre writing postings and to the fact that the asynchronous nature of the interaction allowed them enou gh time to reference dictionaries and use outside help. In these postings they were able to find or recall the appropriate vocabulary and use it to articulate the main points related to their writing and follow those in their first drafts. Therefore, for t hese actors the ACMC mode of communicati on provided more opportunities for the construction of detailed postings using words which would not necessarily be part of their immediate active vocabulary that they would use in a fast paced interaction (e.g. face to face or SCMC). The intertextuality on the idea unit level showed differences that were related to the mode of communication. Based on the comparison of the idea units of the pre writing interactions and the post treatment paragraphs, the ACMC actors, r egardless of their gains, were more likely to establish high level intertextual connection. In other words, there was a stronger tendency to borrow verbatim pre writing discussion idea units or compose draft idea units that were closely related to those fr om the discussion. On the other hand, the SCMC actors were more likely to establish medium and low level intertextual connections they applied more extensive rephrasing techniques or composed new idea units based on information shared during the discuss ion. This difference between the ACMC and the SCMC actors could be attributed to the nature of the language produced during the pre writing interactions in the two communication environments. While the ACMC actors posted longer and mostly complete sentence s

PAGE 556

530 during their pre utterances and phrases. Thus, it could be implied that the ACMC actors were prompted by the asynchronous nature of the communication to compose messages that were closer to their actual first draft writings. It seems that, when composing their first drafts, they relied both on information and specific language structures shared during the pre writing exchange. On the other hand, the SCMC pre writing discussion, due to its dynamic and fast paced nature, provided the means for the collaborative generation of ideas but allowed for less time to compose complete chunks of language that could be readily used in a subsequent draft. Consequently, the environment in which the pre w riting task completion took place seemed to stretch its influence to the first draft stage. The rephrasing strategies the actors applied when transferring idea units from the pre writing discussion to their post treatment paragraphs differed based on the g ain level rather than CMC mode. The high gain actors demonstrated stronger ability to move beyond the single idea unit when rephrasing, either splitting one idea unit from the discussion by developing two or more idea units or merging two idea units into o ne. On the other hand, the low gain actors stayed within the boundaries of the pre writing idea unit which they attempted to rephrase. Thus, the high gain actors demonstrated a stronger ability for further manipulation and polishing of the language produce d during the discussion. The actors who used a collaborative pattern of pre writing interaction, all of In addition, the two high gain SCMC actors (both of them engaged in collaborative

PAGE 557

531 contributions when composing their first drafts. This may suggest that in these cases, the dyadic pattern of pre writing interaction which resulted from the exchange of ideas in a mutual problem solving space was traceable in the first draft writings of the SCMC actors and might have provided stronger support to the composition process. The third level of intertextual connections examined in the case st udies presented in this chapter was the organizational level of intertextuality. At this level, the intertextual connections were the weakest. For three of the ACMC actors and one SCMC actor (regardless of gain level) they were somewhat traceable. The rest of the actors did not show any intertextual connection at this level, and the paragraph organization was either not addressed in the pre writing discussion or the suggestions were neglected in the e been too early in the writing process for the participants to decide on future paragraph organization during the pre writing stage and follow on these decisions without implementing any changes in their drafts. However, the data suggest one more reason f or the weak intertextuality on the organizational level. Four out of the eight participants entirely neglected this aspect of the task in their pre writing discussions. This suggests that it might have been challenging for these actors and their partners t o address the paragraph organization along with the picture interpretation and topic discussion within the same CMC pre writing session. Thus, the pre writing stage appears to be a complex phase of the writing process This phase require s multiple steps in order to address the various aspects of the text to be composed afterwards. These findings confirmed the pedagogical suggestions of Grabe and Kaplan (1996), such as brainstorming, class and group discussion, and peer response.

PAGE 558

532 Finally, all of the actors, with the exception of one low gain ACMC actor, continued the generation of ideas past the pre writing stage. In their first drafts, they composed new idea units that presented new topics not addressed in the discussion. I would argue that the pre writing discussions set general blueprints for the future first drafts which were further developed and modified. Implications for Further Research The understanding of writing as a process rather than a product has reshaped the way researchers and practitioners in the field of SLA approach writing (Grabe & Kaplan,1996; Hyland & Hyland, 2006a). In addition, the recognition of the strong influences of social environments on this process offered new venues for research that informed the pedagogical practices in th e second language classroom. Peer and student/teacher interactions remain in the focus of writing research for more than a decade now (Ferris, 2003; Grabe & Kaplan, 1996; Hyland & Hyland, 2006a,b). However, while the stages of the writing process were reco gnized and described (Grabe & Kaplan, 1996), most of the research has been focused on revising and editing which are the later stages of the writing process. In the field of SLA, there is a limited body of research on the first two stages of the process: p re writing and drafting. Thus, more research that would address the specifics and dynamics of these stages of the writing process is warranted. Based on the outcomes of the exploratory research project presented in this volume, a number of topics that wou ld shed more light on the pre writing and drafting stages and their relations could be suggested. Multiple questions await to be addressed in future research. How do students progress through these first stages of the process? What

PAGE 559

533 is the role of peer coll aboration and of teacher involvement during the pre writing and drafting stages? How do students perceive and participate in peer interactions during the pre the student s benefit from different pre writing tasks such as individual and group brainstorming, mapping, and outlining? Are these tasks interchangeable or could they support a specific sub stage of pre writing? How could language acquisition in terms of vocabulary and specific grammar forms be supported in the pre writing processes and further transferred to drafting? How could writing skills development in terms of text organization, rhetorical soundness, and reader consideration be supported at these stages of the writing process? In addition, the active inclusion of new technologies into the language learning process poses additional issues for future research in the context of the writing process. The role of computer mediated interactions and the specifics of th ese interactions in different modes during the first stages of the writing process could be addressed. The current study showed that there were significant differences in the dyadic patterns of interactions in the two modes. Further research is needed in o rder to attend to the specifics of these differences and their relation to the mode of the computer mediated problem solving tasks. Accounting for the text based nature of the computer mediated communication, this study addressed the intertextual connections between the pre writing discussions and the first drafts of language learners; in future research, it could be instrumental to attend in more detail to the degree and the specifics of language transfer that takes place in the early pre writing stages.

PAGE 560

534 Taking into account the exploratory nature of this study, the results do not allow to establish any definite conclusions about the collaborative and composition practice s that take place during the early stages of the writing process. However, it addressed multiple aspects of these stages such as the dynamics of peer interactions in CMC environments and intertextual connections between pre writing discussions produced in such environment and first drafts. These aspects could be re addressed in future replications of the study or its parts. Pedagogical Implications Although pre writing and drafting stages of the writing process have not been addressed specifically in the S LA writing research or in ESL writing text books, these stages are presented and activities to be completed are suggested. The outcomes of the current research provide some insights on the learning processes which take place in CMC environments during thes e first stages. These outcomes could further inform the pedagogical practices of writing teachers. Based on student opinions shared in the interviews, it could be suggested that writing teachers should not assume that learners come ready to engage in wr iting as a process that entails planning, multiple revisions, peer collaboration, and reader consideration. Thus, the teachers should be ready not only to guide their students through the writing process but also to explicitly address and discuss with the students the process nature of writing and the role of peer collaboration. Similar to the recommendations concerning peer feedback training during the more advanced stages of the writing process (Berg, 1999), the researcher would suggest that ESL learners need training that would support their collaborative participation in pre writing activities. Further, the results of

PAGE 561

535 in a pre writing task that is teacher controlled t o a lesser degree, such as the CMC pre writing task in this study, learners may lose focus, disengage early from the task without combination of approaches could be applied in the pre writing stage students could brainstorm and discuss the topic assigned in small groups or dyads and when the discussion advances, the teacher could provide additional guidance and feedback. Thus, the pre writing stage tasks should not be addresse d in a one step manner but rather should incorporate multiple sub steps such as : (1) task design (2) introduction of the collaborative pre writing task performed by the teacher (3) student interaction during the task completion, and (4) teacher guidance once the discussion advances. It should be noted that the prewriting stage is not a homogeneous stage. It has its own sub stages that should be considered as well when the pre writing tasks are designed and applied in the writing classroom. The existence of such sub stages was evident by the way the participants in this study approached the pre writing task in different CMC environments. The use of computer mediated discussions can support the collaborative planning and idea generation during the pre writ ing stage through its text based nature and the possibility to accomplish parts of this stage outside the classroom as a homework assignment. However, the writing teacher should carefully consider the mode of the CMC interaction in the design of the pre wr iting task. The study revealed that the students participate in the same task completion in a different manner depending on the mode of communication. The interactive nature of the SCMC discussion might be more

PAGE 562

536 appropriate when ideas related to the future writing project are generated, while later, when these ideas need to be refined, ordered, and organized, the ACMC mode of peer collaboration should be considered. Finally, taking into account the fact that computer mediated communication has become an int egral part of the educational process in U.S. universities and that a major part of writing in academic settings is performed using computers, it is crucial to address the development of new literacy skills of ESL learners in intensive English language pro grams. An early diagnosis of computers skills might provide valuable information placement purposes during the first days of the semester, the level of computer skills development should be also assessed. This would allow administrators and/or teachers to identify students that might need additional help in developing their computer skills. Such students can be provided systematic support through extracurricular workshop s and peer or teacher support when computer based activities are assigned. When designing the CALL curriculum of intensive English language programs, it should be taken into account that although the majority of the students that enroll are younger and hav e been exposed to various uses of computers in their every day life, many of them have not used computers for academic purposes. Limitations of the Study Limitations of the Quantitative Stage The main limitation of the statistical tests reported in the quantitative part of this chapter was the sample size the results of only 30 pairs were examined. Considering that the development of writing skills detectable on the surface (vocabulary, syntactical

PAGE 563

537 complexity) and on deep textual levels (presentation and development, overall language use, rhetorical soundness) is a long process, it could be inferred that the progress is pronounced over a longer period of time and statistically significant results could not be detected when examining a small sample. Wit h a larger sample size, a difference between the groups could become more visible. Further, although all efforts were made to control the conditions under which the treatment was administered, the participants had been exposed to different teaching styles during the pre treatment period, and this could have affected their performance. Another limitation is that some of the participants perceived writing as an individual activity (as shared in the interview) perhaps leading them towards intentional neglect o f the outcomes of the pre writing discussion when composing their first drafts; however, due to the small sample size it was impossible to control this variable. It is important to state some of the limitations of the additional analysis as well. The main one is that single lexical items (rather than idea units) were considered: the assumption that they present identical or similar ideas may not always be true. However, despite this limitation, the researcher decided to conduct the t test on the matching DL I ratio scores and to report the findings in order to consider them when discussing the results of the qualitative analysis of the CMC pre writing discussion and post treatment writing intertextual relationship. The second limitation of this analysis is th e fact that it was based solely on the matching DLI ratio scores. This limitation was imposed by the small sample size (n=30); if a larger group of participants were involved in the study, more powerful inferential statistical tests could be used to possib ly reveal additional dimensions of the intertextual relationships considering more dependent variables such as

PAGE 564

538 total number of unique words in each discussion and paragraph, DLIF of CMC discussions and paragraphs, lexical density, etc. Limitations of the Qualitative Stage The study had several limitations identified by the researcher during the design and implementation stages. These limitations concern study design as well as the instruments used in the study. Each of the limitations identified by the re searcher is discussed below. The first limitation refers to the observation frequency; a closer systematic observation of all Level III Writing sections participating in the study would provide the researcher with richer and more detailed data about proce sses of computer and writing skills interaction. More extensive observations of both face to face classes with recorded dyad face to face interactions would also allow the researcher to infer with more confidence how CMC dyadic interactions are related to the face to face patterns of interaction. The interview mode used in the current study was CMC. This mode of interaction provided the researcher and the participants with a higher degree of flexibility and allowed reaching more participants for interview s. However, the researcher found that this interview mode actually impeded the application of follow up and clarification questions. Most of the attempts to ask such questions were ignored by the participants. In addition, the interview could include quest ions related to the trends identified in the CMC interactions. This was not achieved in the current study: due to the large body of qualitative data, the researcher could not complete the data analysis before the end of the semester. In addition, this part icular study focused on one CMC pre writing activity; a longitudinal study with more than one round of CMC interactions performed at different

PAGE 565

539 points during the semester could provide more insights about how students develop their writing skills. Finally, the researcher would consider the use of a different SCMC application. Although L.E.C.S. the application used for this particular study had the advantage of being fully web based and allowed the teachers and the researcher direct access to each interac tion during and after the discussion, it had certain weaknesses in terms of interface design and transmission of postings. These differences affected the display of the chat and limited the analysis of the SCMC chat data. Limitations of the Instrumental Case Studies The main limitation of the instrumental case studies presented in this research project was that the participants were chosen after the data collection was completed. This posed certain restraints on the richness of the qualitative data. More specifically, the researcher did not have access to the participants after the data were analyzed in order to obtain clarification and further explanation about trends revealed in the data. The late selection of the participants also imposed limitations i n terms of face to face and computer class observations, which could be instrumental in identifying additional trends or addressing in more depth trends that were already observed. However, considering the instrumental nature of the case studies, the resea rcher perceived that, despite of the limitations presented above, the findings reported as a result of the data analysis provided valuable insights related to the writing processes under investigation. These findings could be further addressed through in d epth longitudinal case studies.

PAGE 566

540 Conclusions This study aimed to address a gap in the SLA writing research: it examined processes occurring during pre writing activities were involved. More specifically, in the quantitative stage, the completed after a CMC pre writing discussion task. While no differences were found with regard to those features when the ACMC and the SCMC groups were compared, the combination of quantitative and qualitative analysis allowed the researcher to focus on mode in which the pre writing discussion was first drafts on the lexical level. The completion of a pre writing task in a CMC environment and the intertextual connections existing between the pre writing dyad discussions were in the focus of the group case study and the subsequent collection of eight instrumental case studies. Several aspects related to the pre writing and drafting stages of the writing process were stude writing tasks completed in CMC environments, (3) patterns of interaction used during the CMC pre writing task completion, and (4) intertextual connections between the pre writing dyad brainstorming and planning and first drafts. The resul ts of this study confirmed previous research findings related to CMC discourse produced in second language educational environments. In the analysis, the researcher an d textual specifics of language produced in ACMC and SCMC modes of

PAGE 567

541 communication. In addition, in this research project, the interactions and intertextual connections of high and low gainers were compared and similarities and differences were highlighted. Finally, new patterns of interaction, which differed from the patterns of face to face interaction reported in previous research, were identified for CMC environments. The researcher believes that this study provides insights that would broaden the underst inform computer assisted second language writing instruction.

PAGE 568

542 REFERENCES Abbott, S. (1989). Talking it out: A prewriting tool. The English Journal 78 (4), 49 50. Al len, G. (2000). Intertextuality. Routledge. Abras, C. (2002). The principle of relevance and metamessages in online discourse: Electronic exchanges in a graduate course. Language, Literacy and Culture Review 1 (2), 39 53. Aronowitz, S. (1992). Looking out : The impact of computers on the lives of professionals. In M. C. Tuman (Ed.), Literacy Online: The promise and peril of reading and writing with computers (pp. 119 137). University of Pittsburgh Press. Astika, G. G. (1993). Analytical assessment of forei RELC Journal, 24, 61 72. Bamberg, B. (1983). What makes a text coherent? College composition and communication, 34 417 429. Bardovi Harlig, K. (1992). A second look at T unit analysis: Reconsidering the sentence. TESOL Quarterly, 26 ( 2), 390 395. Barker, T., & Kemp, F. (1990). Network theory: A postmodern pedagogy for the written classroom. In C. Handa (Ed.), Computers and community: Teaching composition in the twenty first century (pp. 1 27). Portsmouth, NH: Heinemann. Bartholomae, D. & Petrosky, A. (Eds.). (1986). Facts, artifacts and counterfacts: Theory and method for a reading and writing course Upper Montclair, NJ: Boynton/Cook.

PAGE 569

543 Beauvois, M. H. (1994). E talk: Attitudes and motivation in computer assisted classroom discussion. C omputers and the Humanities, 28 (3), 177 190. Beauvois, M. H. (1997). High tech, high touch: From discussion to composition in the networked classroom. Computer Assisted Language Learning, 10 (1), 57 69. Beauvois, M. H., & Eledge, J. (1996). Personality ty pes and megabytes: Student attitudes toward computer mediated communication (CMC) in the language classroom. CALICO Journal, 13 (2 & 3), 27 45. Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition Hillsdale, NJ: L. Erlbaum. Berg, and writing quality. Journal of Second Language Writing, 8 215 241. Berger, V. B. (1990). The Effects of Peer and Self Feedback on ESL Students' Revisions. CATESOL, 3 (1), 2 1 36. Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spoken and written English Longman Publication Group. Bizzell, P. (1986). Composing processes: An overview. In A. Petrosky & D. Bartolomae (Eds.), The teachin g of writing (pp. 49 70). Chicago: University of Chicago Press. Blake, R. (2000). Computer mediated communication: a window on L2 Spanish interlanguage. Language Learning & Technology, 4 (1), 120 136. Blanton, L. L. (1999). Classroom instruction and languag e minority students: On teaching to "smarter" readers and writers. In L. Harklau, K. M. Losey, & M. Siegal (Eds.), Generation 1.5 meets college composition: Issues in the teaching of

PAGE 570

544 writing to US educated learners of ESL (pp. 119 142). Mahwah, NJ: Lawrenc e Erlbaum Associates. Boardman C. A., & Frydenberg, J. (2002). Writing to communicate: Paragraphs and Essays New York, NY: Longman. Brown, G., & Yule, G. (1983). Discourse analysis Cambridge University Press, Cambridge. Caracelli, V. J., & Greene, J. C. (1993). Data analysis strategies for mixed method evaluation designs. Educational Evaluation and Policy Analysis, 15 195 207. CARLA (n.d.). Virtual Assessment Center Retrieved September 18, 2004, from: http://www.carla.umn.edu/assessment/VAC/Evaluation/P _1.html. Carrell, P. L. (1987). Content and formal schemata in ESL reading. TESOL Quarterly 21 (3), 461 481. Carter, R., & McCarthy, M. (1988). Vocabulary and language teaching London: Longman. Canadian Test Centre Inc. (2003). Comprehensive English Langu age Test (CELT). Retrieved September 15, 2003, from: http://www.canadiantestcentre.com/Products/CELT.asp. Chafe, W. L. (1980). The deployment of consciousness in the production of narrative. In W. L. Chafe (Ed.), The pear stories: Cognitive, cultural and l inguistic aspects of narrative production (pp. 9 50). Norwood, NJ: Ablex. Chafe, W. (1982). Integration and involvement in speaking, writing, and oral literature. In D. Tannen (Ed.), Spoken and written language (pp. 35 54). Norwood, NJ: Ablex.

PAGE 571

545 Chafe, W. L. (1985). Differences between speaking and writing. In D. R. Olson, N. Torrance, & A. Hildyard (Eds.), Literacy, language, and learning: The nature and consequences of reading and writing (pp. 105 123). Cambridge: Cambridge University Press. Chapelle, C. A. (2000). Is network based learning CALL? In M. Warschauer & R. Kern (Eds.) Network based language teaching: Concepts and Practice, (pp. 204 228). Cambridge: Cambridge University Press. Chomsky, N. (1995). The minimalist program Cambridge, MA: MIT Press. C hun, D. (1994). Using computer networking to facilitate the acquisition of interactive competence. System, 22 (1), 17 31. Cody, R. P., & Smith, J. K. (2006). Applied statistics and the SAS programming language (5th ed.). Upper Saddle River, NJ: Prentice Ha ll. Cohen, A. (1994). Assessing language ability in the classroom (2nd ed.). Boston: Heinle & Heinle. Connor, U., & Asenavage, K. (1994). Peer response groups in ESL writng classes: How much impact on revision? Journal of second language writing, 3 (3), 257 276. Cook, V.J. (1998). Relating SLA research to language teaching materials. Canadian Journal of Applied Linguistics, 1 9 27. Cook, V. J. (2001). The neglected role of written language in language teaching Unpublished article. Retrieved on December 15 2004, from: http://homepage.ntlworld.com/vivian.c/Writings/Papers/RoleOfWriting.htm. Cooper, M. (1986). The ecology of writing. College English, 48 364 75.

PAGE 572

546 Cooper, M. (1989). Why are we talking about discourse communities? Or, foundationalism rears its ugly head once more. In M. Cooper & M. Holzman (Eds.), Writing as a social practice (pp. 202 220). Portsmouth, NH: Boyton/Cook. Couture, B. (1999). Modeling and emulating. In T. Kent (Ed.), Post Process Theory (pp. 30 48). Southern Illinois University Pr ess. Creswell, J. W., Plano Clark, V. L. Gutmann, M. & Hanson, W. E. (2003). Advanced mixed method research designs. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social & behavioral research (pp. 209 240). Thousand Oaks, CA: Sage. Cr owhurst, M. (1987). Cohesion in argument and narration at three grade levels. Research in the Teaching of English, 21 (2), 185 201. Cumming, A. (1989). Writing expertise and second language proficiency. Language Learning, 39 81 141. Cumming, A., Kantor R. Powers D., Santos T., & Taylor, C. (2000). TOEFL 2000 writing framework: A working paper. TOEFL monograph series. ETS Retrieved on January 3, 2005, from: http://ftp.ets.org/pub/toefl/253719.pdf. De Beaugrande, R., & Dressler, W. (1981). Introduction to text linguistics London and New York, NY: Longman. De Graaf, R., Jauregi, K., & Nieuwenhuijsen, D. (2002). Collaborative learning in a virtual writing environment. In A. Mendez Vilas and J. A. Mesa Gonzales, I. Solo de Zaldivar (Eds.), Educational Technol ogy: International conference on ICTs in Education, 2002 (pp. 302 306). Badajoz, Spain. Dewey, J. (1964). Logic: The theory of inquiry New York, NY, H. Holt and Company.

PAGE 573

547 Dochy, F., Segers, M., & Sluijsmans, D. (1999). The use of self peer, and co assess ment in higher education: A review of literature. Studies in Higher Education, 24, 331 350. ELI CALL (2003). Annual Meeting Notes August 28, 2003. English Language Institute, University of South Florida ELI @ USF (2002). Mission Statement. Retrieved on June 29, 2006 from: http://www.cas.usf.edu/eli/mission.htm Ellis, R. (1994). The study of second language acquisition London: Oxford University Press. Emig, J. (1971). The composing processes of twelfth graders Urbana, IL: National Council of Teachers o f English. Enright, M. K., Grabe, W., Koda K., Mosenthal, K., Mulcahy Ernt, P., & Schedl, M, (2000 ). TOEFL 2000, Reading framework: A working paper. TOEFL monograph series. ETS Retrieved on January 3, 2005, from: http://ftp.ets.org/pub/toefl/253718.pdf. Faigley, L. (1994). Competing theories of process: A critique and a proposal. In S. Perl (Ed.), Landmark essays on writing process (pp. 149 164). Hermagoras Press. Faigley, L. (1990). Subverting the electronic workbook: Teaching writing using networked com puters. In D. Baker & M. Monenberg (Eds.), The writing teacher as researcher: Essays in the theory and practice of class based research (pp. 290 311). Portsmouth, NH: Heinemann. Ferris, D. (2003). Response to student writing: Implications for second langua ge students Mahwah, NJ: Lawrence Erlbaum Associates.

PAGE 574

548 Ferris D., & Hedgcock J. S. (1998). Teaching ESL composition Mahwah, NJ: Lawrence Erlbaum Associates. Flower, L., & Hayes, J. (1980). The cognition of discovery: Defining a rhetorical problem. College Composition and Communication, 31, 21 32. Freiermuth, M.R. (2002). Internet chat: Collaborating and learning via e conversations. TESOL Journal, 11 (3), 36 40. Gaies, S. J. (1980). T unit analysis in second language research: Applications, problems and limi tations. TESOL Quarterly, 14 (1), 53 60. Gall, M., Borg. W., & Gall, J. (1996). Educational research: An introduction (6th ed.). White Plains: Longman. Glass G. V., & Hopkins K. D. (1995). Statistical methods in education and psychology Boston: Allyn and B acon. Goodman, K. S. (1989). Whole language research: Foundations and development. The Elementary School Journal 90 208 221. Grabe W., & Kaplan R. B. (1996). Theory and practice of writing London: Longman. Graff, H. J. (1981). Introduction. In H. J. G raff (Ed.), Literacy and social development in the West (pp. 1 13). Cambridge University Press. Hairston, M. (1982). Winds of change: Thomas Kuhn and revolution in the teaching of writing. College Composition and Communication, 33 76 88. Halliday, M. A. K (1985). An introduction to functional grammar. Edward Arnold, London. Halliday, M. A. K. (1989). Spoken and written language Oxford: Oxford University Press.

PAGE 575

549 Hartman, K., Neuwirth, C., Kiesler, S., Sproull, L., Cochran, C., Palmquist, M., & Zubrow, D. ( 1991).Patterns of social interaction and learning to write: Some effects of networked technologies. Written Communication, 8 (1), 79 113. Hatch, J. A. (2002). Doing qualitative research in education settings State University of New York Press. Haynes M., & Baker, I. (1993). American and Chinese readers learning from lexical familiarization in English texts. In T. Huckin, M. Haynes, & J. Coady (Eds.), Second language reading and vocabulary acquisition (pp. 130 152). Norwood: Ablex. Henson, R. K. (1998, Nove mber). ANCOVA with intact groups: Don't do it! Paper presented at the annual meeting of the Mid South Educational Research Association, New Orleans, LA. (ERIC Document Reproduction Service No. ED 426086). Hildyard, A., & Hidi, S. (1985). The production and recall of narratives. In D. R. Olson, N. Torrance, & A. Hildyard (Eds.), Literacy, language, and learning: The nature and consequences of reading and writing (pp. 285 306). Cambridge: Cambridge University Press. Hunt, K. W. (1965). Grammatical structures written at three grade levels NCTE research report #3. Champaign, IL: NCTE. Hyland, K. & Hyland, F. (2006a). Feedback in second language writing: Contexts and issues Cambridge: Cambridge University Press.

PAGE 576

550 Hyland, K. & Hyland, F. (2006b). Contexts and iss ues in feedback on L2 writing: An introduction. In K. Hyland & F. Hyland (Eds.), Feedback in second language writing: Contexts and issues (pp. 1 15). Cambridge: Cambridge University Press. Jacobs, H. L., Zingraf, S., Wormuth, D., Hratfiel V., & Hughey, J. (1981). ESL composition profile. Retrieved on December 28, 2005, from: http://eli.tamu.edu/students/Profile.htm. Jepson, K. (2005). Conversations and negotiated interactions in text and voice chat rooms. Language learning and technology, 9 (3), 79 98. Jin L (2007). Computer mediated peer response in an ESL academic writing class: cultural historical activity theoretical perspective. Unpublished dissertation, University of South Florida, Tampa. Joh nson, K. (2003). Designing language teaching tasks Palgrave, NY: McMillan. Johnson, M. (2004). A Philosophy of second language acquisition Yale University Press Kanto Gakuin School Corporation, Japan (n.d.). Language educational chat system (L.E.C.S.). Retrieved on September 15, 2004, from http://www.kanto gakuin.ac.jp/~taoka/lecs/. Kaplan, R. B. (1987). Cultural thought patterns revisited. In U. M. Connor & R. B. Kaplan (Eds.), Writing across languages: Analysis of L2 text (pp. 9 27). Reading, MA: Addi son Wesley. Kasper, L. F. (2000). New technologies, new literacies: Focus discipline research and ESL learning communities. Language learning and technology, 4 (2), 105 128. Retrieved on January, 04 from: http://llt.msu.edu/vol4num2/kasper/default.html.

PAGE 577

551 Kem per E. A., Stringfield S., & Teddlie C. (2003). Mixed methods sampling strategies in social science research. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 273 296). Thousand Oaks: Sage. Kelm, O. R. (1992). The use of synchronous computer networks in second language instruction: A Preliminary Report. Foreign Language Annals, 25 441 454. Kent, T. (1999). Introduction. In T. Kent (Ed.), Post Process Theory (pp. 1 6 ). Southern Illinois University Press. Kern, R. (1995). Restructuring classroom interaction with network computers: effects on quantity and characteristics of language production. Modern Language Journal, 79 457 476. Koda, K. (1993). Task induced vocabulary in FL composition: Language specifi c perspectives. Foreign language annals, 26 332 346. Kroll, B. (1977). Combining ideas in written and spoken English: A look at subordination and coordination. In E. O Keenan & T. L. Bennett (Eds.), Discourse across time and space. Southern California Oc casional Papers in Linguistics 5 (pp. 69 108). Los Angeles: University of Southern California. Kroll, B. (1991). Teaching writing in the ESL context. In M. Celce Murcia (Ed.) Teaching English as a second or foreign language (2nd ed., pp. 245 263), Boston: Heinle. Kroll, B. & Reid, J. (1994). Guidelines for designing writing prompts: clarifications, caveats, and cautions. Journal of Second Language Writing, 3 (3), 231 255. Kytle, R. (1970). Pre writing by analysis. College Composition and Communication, 21 (5 ), 380 385.

PAGE 578

552 Larsen Freeman, D. (1983). Assessing global second language proficiency. In H. W. Seliger & M. H. Long (Eds.) Classroom oriented research in second language acquisition (pp. 287 304). Rowley, MA: Newbury House Publishers. Laufer, B. (1997). Th e lexical plight in second language reading: Words you don't know, words you think you know, and words you can't guess. In J. Coady & T. Huckin (Eds.), Second language vocabulary acquisition: A rationale for pedagogy (pp. 20 34). United Kingdom: Cambridge University Press. Lauren, U. (2002). Working papers 50 2002, 53 68. Retrieved on December 18, 2004, from: http://www.ling.lu.se/disseminations/pdf/50/Lauren.pdf. Lee, I. (2000). Explo ring reading writing connections through a pedagogical focus on coherence Canadian Modern Language Review 5 (2), 352 356 writing needs across the disciplines. TESOL Quar terly, 28 81 101. Lemke, L. J. (1985). Ideology, intertextuality, and the notion of register. In J. D. Benson & S. G. Greaves (Series Eds.), Advances in Discourse processes series: Vol. 15 Systemic perspectives on discourse: Vol. 1 (275 294). Ablex Publis hing. Leu, J., & Donald J. (2000). Literacy and technology: Deictic consequences for literacy education in an information age. In M. L. Kamil, P. Mosenthal, P. D. Pearson & R. Barr (Eds.), Handbook of reading research, Volume III Mahway, NY: Lawrence Erlb aum Associates. Levy, M. (1997). Computer assisted language learning: Context and conceptualization. Oxford: Clarendon.

PAGE 579

553 Lincoln, Y. S., & Guba E. G. (1985). Naturalistic inquiry Newbury Park, London, New Delhi: Sage. Lindenau, S. (1984). Improving electronic laboratory study in English as a second language pro grams: A case in point. NALLD Journal, 19 6 13. Liu, M., Moor, Z., Graham, L., & Lee, S. (2002) A look at the research on computer based technology use in second language learning: A review of the literature from 1990 2000. Journal of Research on Technol ogy in Education, 34 (3), 250 273. Liu, J, & Sadler, R. W. (2003). The effect and affect of peer review in electronic versus traditional modes on L2 writing. Journal of English for Academic Purposes 2, 193 227. Loftin, L. B., & Madison, S. Q. (1991). The ex treme dangers of covariance corrections. In B. Thompson (Ed.), Advances in educational research: Substantive findings, methodological developments Vol. 1. (pp. 133 147). Greenwich, CT: JAI Press. Lockhart, C. & Ng, P. (1995). Analyzing talk in ESL peer res ponse groups: Stances, functions and content. Language Learning, 45 (4), 605 655. MacDonald, S. P. (2002). Prose styles, genres, and levels of analysis. Styles, Winter 2002. Retrieved on January 01, 2005, from: http://www.findarticles.com/p/articles/mi_m234 2/is_4_36/ai_98167916/pg_7. Martin, J. (1989). Factual writing: Exploring and challenging social reality Oxford University Press. Matsuda, P. K. (2003). Process and post process: A discursive history. Journal of Second Language Writing, 12 (1), 65 83.

PAGE 580

5 54 Mat suda, P. K., Canagarajah, A. S., Harklau, L., Hyland, K., & Warschauer, M. (2003). Changing currents in second language writing research: A colloquium. Journal of Second Language Writing, 12 (2), 151 179. Maxcy, S. J. (2003). Pragmatic in mixed methods res earch in the social sciences: The research for multiple modes of inquiry and the end of the philosophy of formalism. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 51 89). Thousand Oaks, CA: Sage. Max well J. A., & Loomis, D. M. (2003). Mixed methods design: An alternative approach. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 241 271). Thousand Oaks, CA: Sage. McComiskey, B. (2000). Teaching co mposition as a social process Utah State University Press: Logan, Utah. McDonell, W. (1992). Language and cognitive developemnt through cooperative group work. In C. Kessler (Ed.), book (pp. 51 64). Pren tice Hall Regents. Mehnert, U. (1998). The effects of different lengths of time for planning on second language performance. Studies in Second Language Acquisition 20 83 108. Melvin, B. S., & Stout, D. F. (1987). Motivating language learners thought authe ntic material. In W.M. Rivers (Ed.), Interactive Language Teaching (pp. 44 56). New York: Cambridge, University Press. in ESL writing instruction. TESOL Quarterly, 28 (4 ).

PAGE 581

555 Merriam, S. B. (1998). Qualitative research and case study application in education San Francisco: Jossey Bass Publishers. Merriam Webster Dictionary Online (2006). Merriam Webster Inc. Retrieved August 30, 2006, from: http://www.m w.com/ software]. Microsoft Corporation. Miles, M., & Huberman, A. (1994). An Expanded Sourcebook: Qualitative Data Analysis Thousand Oaks: Sage Publications Thousand Oaks, CA: Sage. Miller, L. (2007). Review of calling on CALL: From theory and research to new directions in foreign language teaching. Language Learning and Technology, 11 (1), 40 44. Milton, J. (2006). Resource rich web based feedback: Helping learners become independent writers. Feedback in second language writing (pp. 123 139). Cambridge: Cambridge University Press. Morris, S. M. (1968). Pre writing. College Composition and Communication 19 (1), 24 26. Moskal, B. M. & Leydens J. A. (2000). Scor ing rubric development: validity and reliability. Practical Assessment, Research & Evaluation 7(10). Retrieved July 2, 2006 from http://PAREonline.net/getvn.asp?v=7&n=10 MySQL (2006). MySQL. Retrieved August 08, 2006, from: http://mysql.com/ Nathan, G (2001). Methodologies for Internet surveys and other telesurveys. Retrieved on September, 29, 2003: http://webfarm.jrc.cec.eu.int/ETK NTTS/Papers/final_papers/13.pdf.

PAGE 582

556 Nation, I. S. P. (2001). Learning vocabulary in another language Cambridge: Applied Ling uistics. Nelson, G.L., & Murphy, J, M. (1993). Peer response groups: Do L2 writers use peer comments in revising their drafts? TESOL Quarterly, 27 135 142. Nelson, G., & Carson, J. (2006). K. Hyla nd & F. Hyland (Eds.), Feedback in second language writing (pp. 42 59). Cambridge: Cambridge University Press. Nicenet (1998). Nicenet Internet classroom assistant Retrieved on September 29, 2003, from: http://www.nicenet.org. North, S. (1987). The makin g of knowledge in composition Portsmouth, NH: Heinemann. Nunan, D. (1989). Designing tasks for the communicative classroom Cambridge: Cambridge University Press. Ohta, A. S. (2000). Rethinking interaction in SLA: Developmentally appropriate assistance i n the zone of proximal development and the acquisition of L2 grammar. In J. P. Lantolf (Ed.), Sociocultural theory and second language learning (pp. 51 78). Oxford: Oxford University Press. Onwuegbuzie, A. J. (2003). Framework of internal and external vali dity. Research in the Schools 10 (1), 71 89. Onwuegbuzie, A. J., & Daniel, L. G. (2003, February 12). Typology of analytical and interpretational errors in quantitative and qualitative educational research. Current Issues in Education 6 (2). Retrieved on October 28, 2004: http://cie.asu.edu/volume6/number2/index.html.

PAGE 583

557 Onwuegbuzie, A. J., & Levin, J. R., (2005). Strategies for aggregating non significant outcomes of a single study. Research in the Schools 12 (1), 10 19. Ortega, L. (1997). Processes and outc omes in networked classroom interaction: Defining the research agenda for L2 computer assisted classroom discussion. Language Learning and Technology, 1 (1), 82 93. Oxford, R., Rivera Castillo, Y., Feyten, C., & Nutta, J. (1998). Computers and more: creati ve uses of technology for learning a second or foreign language. In V. D arleguy, A. Ding, & M. Svensson (Eds.), Les Nouvelles Technologies applications pratiques. Retrieved on Nov ember 15, 2002 retrieve from: http://www.insa lyon.fr/Departements/CDRL/computers.html. Painter, C. (1989). Learning language: A functional view of language development. In R. Hasan & J. Martin (Eds), Language development: Learning language, learning cultu re (pp. 18 65). Norwood, NJ: Ablex. Park, Y. M. (1988). Academic and ethnic background as factors affecting writing Performance. In A. C. Purves (Ed.), Writing across languages and cultures: issues in contrastive rhetoric. Written communication annual: an international survey of research and theory (Vol. 2. pp. 261 272). Newbury Park, CA: Sage. Patton, M. Q. (2002). Qualitative research and evaluation methods. (3rd ed.). Thousand Oaks, CA: Sage Petraglia, J. (1999). Is there life after process? The role o f social scientism in changing discipline. In T. Kent (Ed.), Post Process Theory (pp. 49 64 ). Southern Illinois University Press.

PAGE 584

558 PHP (2006). PHP5. Retrieved August 08, 2006, from: http://php.net/ Pica, T., Kanagy, R., & Falodun, J. (1993). Choosing and u sing communication tasks for second language instruction and research. In G. Crookes & S. M. Gass (Eds.), Tasks and language learning: Integrating theory and practice (pp. 9 34). Clevedon, UK: Multilingual Matters. Plough, I., & Gass, S. M. (1993). Interlo cutor and task familiarity: Effects in interactional structure. In G. Crookes & S. M. Gass (Eds.), Tasks and language learning: Integrating theory and practice (pp. 35 56). Clevedon: Multilingual Matters. Qualitative Solutions and Research Pty. (2001). NVI VO (Version 2.0)[Computer software]. Melbourne, Australia. Raimes, A. (1991). Out of the woods: Traditions in the teaching of writing. TESOL Quarterly, 25, 407 430. Read, J. (2000). Assessing vocabulary Cambridge: Cambridge University Press. Reid, J. (199 0). Responding to different topic types: A quantitative analysis form a contrastive rhetoric perspective. In B. Kroll (Ed.), Second language writing: Research insights for the classroom (pp. 191 210). Cambridge: Cambridge University Press. Reither, J. A. (1994). Writing and knowing: Towards redefining the writing process. In S. Perl (Ed.), Landmark essays on writing process (pp. 141 148). Hermagoras Press. Rodrigues, R. J. (1983). Tools for developing pre writing skills. English Journal, 72 (2), 58 60. Rohm an, D. G. (1965). Pre writing: the stage of discover in the pre writing process. College Composition and Communication, 16 (2), 106 112.

PAGE 585

559 Reference: Concepts. Cary, NC: SAS Institu te Inc. Salaberry, R. M. (2001). The use of technology for second language learning and teaching: A retrospective. The Modern Language Journal, 85, 39 56. students. TES OL Quarterly, 22 69 90. Seliger, H., & Shohamy, E. (1989). Second language research methods Oxford: London Oxford University Press. Sengupta, S. (2001). Exchanging ideas with peers in networked based classrooms: An aid or a pain? Language Learning and T echnology, 5 (1), 103 134. Shih, R. H. (2000). Compiling Taiwanese Learner Corpus of English. Computational Linguistics and Chinese Language Processing 5 (2), 89 102. Retrieved on December 28, 2004 from: http://rocling.iis.sinica.edu.tw/CLCLP/Vol5 2/paper4.p df. Shin, D. S. (2006). ESL students' computer mediated communication practices: context configuration. Language, Learning & Technology 10 (3), 65 85. Silva, T., & Matsuda, P. K. (Eds.). (2001). On second language writing. Mahwah, NJ: Lawrence Erlbaum As sociates. Simpson, J. (2005). Learning electronic literacy skills in an online language learning c ommunity. Computer Assisted Language Learning, 18 (4), 327 345. Smith, A.N. (1996). Using video and newspaper texts to provide topic schemata in the compositio n class. The French Review, 70 (2), 167 179.

PAGE 586

560 Smith, B. (2003). Computer mediated negotiated interaction: An expanded model. The Modern Language Journal, 87 (1), 38 57. Snellings, P., van Gelderen A., & de Glopper K. (2002). An aspect of fluent second languag e production that can be enhanced. Language Learning, 52 (4), 723 54. Sotillo, M. S. (2000). Discourse functions and syntactic complexity in synchronous and asynchronous communication. Language Learning and Technology 4 (1), 82 119. Spack, R. (1984). Inventi on strategies and the ESL college composition student. TESOL Quarterly, 18 (4), 640 670. St. John, E., & Cash, D. (1995). Language learning via e mail: Denstrable success in German. In M. Warschauer (Ed.), Virtual connections: Online activities and project s for networking language learners, (pp. 191 197). Honolulu, HI: University of Hawaii, Second Language Teaching and Curriculum Center. Stake, R. (1995). The art of case research Thousand Oaks, CA: Sage. Still, J., & Worton, M. (1990). Introduction. In M. Worton & J. Still (Eds.), Intertextuality: Theories and practices (pp. 1 44). Manchester University Press. Storch, N. (2002). Patterns of interaction in ESL pair work. Language Learning 52 (1), 119 158. Sullivan, N., & Pratt, E. (1996). A comparative study of two ESL writing environments: A computer assisted classroom and a traditional oral classroom. System, 29 491 501. Swaffar, J. (1998). Networking language learning: Introduction. In J. Swaffar, S. Romano, P. Markley, & K. Arens (Eds.), Language learnin g online: Theory and

PAGE 587

561 practice in the ESL and L2 computer classroom (pp. 1 15). Austin, TX: Labyrinth. Swain, M. (1985). Communicative competence: Some roles of comprehensive input and comprehensible output in its development. In S. M. Gass, & C. Madden (Ed s.). Input in second language acquisition (pp. 235 253). Cambridge, MA: Newbury House. Swales, J. M. (1990). Genre analysis: English in academic and research settings Cambridge: Cambridge University Press. Swales, J. (1993). Genre and engagement. Revue Be 71 (3), 687 698. Tashakkori, A., & Teddlie C. (1998). Mixed methodology. Combining qualitative and quantitative approaches. Applied Social Research Methods Series, Vol. 46. Thousand Oaks, CA: Sage. Tashakkori, A., & Teddlie C. (2003). Glossary. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social & behavioral research (pp. 703 718). Thousand Oaks, CA: Sage. Teasley, S., & Roschelle, J. (1993). Constructing a joined problem space: The computer as a tool fo r sharing knowledge. In S.P. Lajoie & S. J. Derry (Eds.), Computer as cognitive tools (pp. 229 258). Hillsdale, NJ: Lawrence Erlbaum Associate. Teddlie C., & Tashakkori, A. (2003). Major issues and controversies in the use of mixed methods in the social a nd behavioral sciences. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social & behavioral research (pp. 3 50). Thousand Oaks, CA: Sage.

PAGE 588

562 Trimbur, J. (1994). Taking the Social Turn: Teaching Writing Post Process. College Composition and Communication 45 108 18. Tyner, K. (1998). Literacy in a digital world: Teaching and learning in the age of information. Mahwah, NJ: Lawrence Erlbaum Associate Publishers. The University of Michigan, English Language Institute, Testing Certification Divi sion (2003). Test Publications. Retrieved September 09, 2004, from http://www.lsa.umich.edu/eli/testpub.htm. Ure, J. N. (1971). Lexical density and register differentiation. In E. G. Perren & J. L. Trim (Eds.), Applications of linguistics: Selected paper s of the second international congress of applied linguistics (pp. 443 452). Cambridge: Cambridge University Press. Voloshinov, V. N. (2001). Marxism and the philosophy of language. In P. Morris (Ed.), The Bakhtin reader (pp. 26 37). London: Arnold. Wang, Y. (1996). E mail dialog journaling in an ESL reading and writing classroom. Proceedings of the 18th National Convention of the Association for Educational Communications and Technology (pp. 766 781). Indianapolis, IN, Ware, P.D., & Warschauer, M. (2006) Electronic feedback and second language writing. In K. Hyland & F. Hyland (Eds.), Feedback in second language writing (pp. 105 122). Cambridge: Cambridge University Press. Warschauer, M. (1996). Comparing face to face and electronic discussion in the sec ond language classroom. CALICO Journal 13 (2), 7 26. Warschauer, M. (1997). Computer mediated communication. The Modern Language Journal 81, 470 481.

PAGE 589

563 Warschauer, M. (1999). Electronic literacies: Language, culture, and power in online education. Mahwah, NJ, London: Lawrence Erlbaum Associates, Publishers. Warschauer, M. (2004a). The rhetoric and reality of aid: Promoting educational technology in Egypt. Globalisation, Societies and Education 2 (3). Warschauer, M. (2004b). Technology and writing. In C. Davidso n and J. Cummins (Eds.), Handbook of English language teaching Kluwer: Dordrecht, Netherlands. Retrieved on November 15, 2004, from: http://www.gse.uci.edu/markw/technology.pdf. Warschauer, M. (2006). Literacy and technology: Bridging the divide. In D. Gibbs & K. L. Krause (Eds.), Cyberlines 2: Languages and cultures of the Internet, (pp. 163 176). Albert Park, Australia: James Nicholas Wertsch, J. V. (1985). Vygotsky and the social formation of mind. Harvard : Harvard University Press. White, E. M. (1 994). Teaching and assessing writing: Recent advances in understanding, evaluating, and improving student performance (2nd rev. ed.). San Francisco: Jossey Bass. Widdowson, H. (1978). Teaching language as communication Oxford: Oxford University Press. Won g, W. (2003). Textual enhancement and simplified input: Effects on L2 comprehension and acquisition of non meaningful grammatical form. Applied language learning 13 (2), 17 45. Worton, M, & Still, J. (1990). A note on vocabulary. In M. Worton & J. Still, In tertextuality: Theories and practices (pp. viii ix). Manchester University Press.

PAGE 590

564 Yin, R. K. (2003). Case st udy research: Design and method (3rd ed.). Thousand Oaks, London: Sage. Zamel, V. (1976). Teaching composition in the ESL classroom: What we can le arn f r o m research in the teaching of English. TESOL Quarterly, 10, 67 76. Z amel, V. (1980). Re evaluating sentence combining p ractice. TESOL Quarterly 14(1), 81 90. Zoran, A., & Sarieva, I. (2003). Extending your classroom online: The benefits of using I CA 2. Florida Foreign Language Journal, 1 (1), 67 74.

PAGE 591

565 APPENDICES

PAGE 592

566 Appendix 1 A: Intermediate Academic Writing Course: Goals and Objectives ACADEMIC WRITING III COURSE DESCRIPTION: This course focuses on developing academic writing ski lls. Topics with a cultural focus provide the stimulus for writing assignments. Students will review sentence types, focus on paragraph development, and begin working on essay development in this course. COURSE GOALS: 1. Students will read passages and use them as the basis for writing. 2. Students will review different types of sentences to prepare them for paragraph writing. 3. Students will be exposed to a variety of academic writing styles and patterns. 4. Students will express ideas or opinions b ased on the readings orally and in writing. COURSE OBJECTIVES: By the end of the course, students will be able to 1. Write simple, compound, complex, and compound complex sentences. 2. Write well structured coherent paragraphs with a topic sentence, su pport, and a conclusion. 3. Read an article or story and write a short summary or paraphrase. 4. Write a process report. 5. Write a paragraph of extended definition of a term used in their field of study. 6. Write multiple paragraphs using a variet y of rhetorical patterns, including a paragraph comparing/contrasting two graphs and a paragraph illustrating cause/effect using an academic field case study. 7. Expand a paragraph into an essay.

PAGE 593

567 8. Write fully developed paragraphs on given topics with in given time limits. 9. Write one in class essay. 10. Edit writing for correct grammar, punctuation, capitalization, and transitions. 11. Use library sources to identify, locate, and gather data for use in their writing. 12. Narrow a topic after brainst orming and organize into outline form. 13. Summarize, paraphrase, and cite information gathered through library sources. CALL OBJECTIVES : 1. Students will become familiar with the procedures for accessing ELI software and Web sites on the Internet. 2. St udents will improve their writing skills through the use of grammar, writing, and spelling software and through Web sites located on the Internet. 3. Students will become familiar with Web LUIS and Internet searches as sources of facts in the research dat a gathering process. 4. Students will share ideas and opinions in writing through e mail exchanges and on line chats with their classmates and instructor. CALL sources for Academic Writing III: Microsoft Word, Grammar 3D, Web LUIS, Web sites on the Inte rnet, Yahoo Messenger, e mail.

PAGE 594

568 Appendix 1 B: Comprehensive English Language Test (CELT) Grade Range: High School through Adult Administration: Individual or Group Time: Approximately 2 hours 15 minutes The Comprehensive English Language Test (CELT) has ea rned an excellent reputation as a valid and reliable instrument for measuring English language proficiency for English as a second language (ESL). It is especially appropriate for high school, college and adult ESL programs at the intermediate and advanced levels. Features and Benefits Listening, Structure and Vocabulary tests measure important dimensions of second language learning. Each section is of sufficient length to yield reliable results. Authors are experienced test writers in the field of ESL. A ll items were pretested on large samples of students of diverse language backgrounds enrolled in intensive, college level and high school ESL programs. Normative Information Student Percent Correct Scores may be referenced against normative data for severa l reference groups: Foreign students admitted to an American university ESL program; Adult ESL students attending a language upgrading program in a community college; Francophone ESL students from Quebec; High school ESL students; Adult ESL students in an intensive language institute; Adult non native speakers accepted for academic work. Components The Examiner's Instructions and Technical Manual contains instructions for test administration, a discussion of the procedures used in developing parallel forms, and norms for six different reference groups. Appendix 1 B, Comprehensive English Language Test (CELT) (cont.) Separate, reusable Test Books for each of Forms A and B contain all three sections of the test. Two parallel forms may be administered i n alternate years for placement purposes or

PAGE 595

569 as entry and exit tests within the same semester. Answer Keys include built in table for easy conversion of number of incorrect and omitted items expressed in a percentage correct score. C 60 Cassette is designe d for the Listening section of each form.

PAGE 596

570 Appendix 1 C: The Michigan Test of English Language Proficiency (MTELP) Advanced proficiency level The Michigan Test of English Language Proficiency (MTELP) is a 100 item multiple choice test containing grammar i tems in a conversational format, vocabulary items requiring selection of a synonym or completion of a sentence, and reading passages followed by comprehension questions. The MTELP is a retired component of the Michigan English Proficiency Battery, which wa s used to assess the English language proficiency of non native speakers of English who wish to pursue academic work at colleges or universities where English is the medium of instruction. This test may be appropriate for advanced secondary school students also, but norms available on the MTELP (see the MTELP Examiner's Manual) pertain only to adult non native speakers of English interested in college or university study. The test is inappropriate for those with elementary to low intermediate knowledge of E nglish and for those without a high school education. Scoring is by punched stencil. Purchasers may use their own scannable answer sheets. Raw scores can be converted to adjusted scores by using equation tables found in the MTELP Examiner's Manual. Time: 7 5 minutes 3 Forms available: P, Q, R

PAGE 597

571 Appendix 2: ESL/CALL Curriculum Objectives and Activities CALL Internet Software Word Processing Culture L/S Strategies Lower Objectives Suggested Activities (S.A.) Objectives S.A. Objectives S.A. Level I 1. Demons trate knowledge of facts about American culture 2. Locate specific information on maps, in pictures, and in directories 3. Read short passages and respond to them in writing. 1. or http://www.1 language.com/worldfactbook2002/index.html Answer T/F, find their own, check your 2. Join Nicenet class Listen to esl http://www.esl lab.com/intro2/intro2.htm Go into Nicenet and introduce yourself (send a personal message to all) 3. On Nicenet a short passage about American Culture In conferencing, short t/f questions, click on reply and answer with correct state ments (i.e. 1. Practice characteristic aspects of consonants and vowels which interfere with intelligibility and communic ability ELLIS Sound Recorder (dictation, record own voice) 1. Use affixes, word families, and word buildin g to develop and increase their basic vocabulary. 2. Use the learning strategies of scanning, sequencing, classification, and making analogies in order to organize and facilitate comprehension of authentic graphic materials and surroundings 3. Write a p aragraph with a title, margins, and indentation Crossword puzzles (file, open, save, new, link, hard copy,etc.) Dictation (typing and navigation) Follow directions (opening multiple windows) copy paragraphs (fonts, margins, indents, table space, bo ld) scavenger hunt (finding docs, directories, folders, save as) fill in form create various uses for Word (labels, envelope, Wizards, letter, memo, fax) Level II 1. Demonstrate an understanding of cultural references by the acquisition of new words in context. 2. Extract information from a variety of authentic materials, including maps, pictures, and directories. 1. Set up class in Nicenet On Nicenet write about their first week experience in Tampa and at ELI OR List of items to discuss (elevators, office, etc.) 2. ELI site find info OR library hours of operation OR ELI staff OR Go to tbo.com and scan for information on Questions provided under Conferencing or Personal Messages and students reply Compare with your country (Teacher then gives out slips of papers with names on it and students reply to the person that they received) home. 1. Practice characteristic aspects of consonants and vowels which interfere with intelligibility and c ommuni cability ELLIS Sound Recorder (dictation, record own voice) 1. Use synonyms, word parts, and word order patterns to enhance their vocabulary development and to guess meaning from context. 2. Extract information from maps, pictures, and directori es. 3. Write simple, compound, and complex sentences. 4. Follow the process of a paragraph to write well structured paragraphs with a topic sentence, support, and a conclusion. Level I+ Synonym, word parts, word order (grammar check, thesaurus) Sentences /order into paragr. (formatting paragraphs) Follow directions (header/footer, page numbering, renaming/deleting files) (word count, from a:// to x:// drive) Scavenger hunt with help function Insert tables/graphs/graphics/clip Usin g toolbars (parts vs. level) Access different menus/saving downloads, etc

PAGE 598

572 Appendix 2, ESL/CALL Curriculum Objectives and Activities (cont.) CALL Internet Multiple Applications Word Processing L/S Academic Interaction Test Prep Elective Academic Wri ting Writing Research Skills Upper Objectives Suggested Activities Objectives S.A. Objectives Suggested Activities Level III 1. Interpret and answer short essay questions based on listening passages. 2. Make inferences and understand new ideas from listening passages 1. Sign up for Nicenet/ interview a student, introduce your partner to Nicenet/read the posting about yourself and post any additional information 2. specific theme (i.e. Stress) 3. A three day v acation (weather, activities at destination) follow up present your ideas to other classmates mingling 1. Prepare for the computer based exam multiple applications. 1. Write simple, compound, complex, and compound complex sentences. 2. Identify cohere nt paragraphs with a topic sentence, support, and a conclusion. 4. Write a process report 5. Edit writing for correct grammar, punctuation, capitalization, and transitions 1. Rearrange (cut/paste) sentences to make a paragraph 2. Spell check, grammar c heck 3. Create a flyer with text box/graphics for events at the ELI 4. Open two windows on split pane and copy/cut/paste from one to another to create new doc. 5. copy graphic wrap text around it 6. Each student chooses part of Word and describes how t o use it. Level IV 1. Search WebLuis and internet sources to gather data for an oral presentation. 2. Identify key vocabulary used in the lectures and supplemental academic materials. 3. Identify main or key ideas and important details of a lecture 1. Sig n Up for Nicenet Write 3 true and 2 false statements about yourself and post under conferencing. Pair up students and the pairs quiz each other which one is true/false. The partner then reports on the true statements. 2. Library Scavenger Hunt (debrief in c lass or record findings in sound recorder) 3. News Event (predict/discuss) information from CNN, ABC, FOX with worksheet and debrief 1. Prepare for the computer based exam multiple applications. 1. Make use of standard rhetorical forms: narration, descrip tion, comparison/contrast, cause/effect, and/or persuasion/argumentation 2. Review and edit writing for content and mechanics, including: paragraph structure, rhetorical style, unity and coherence, capitalization and punctuation, transitions, and connec tors. Level III + 1. Give students sentences from paragraphs of different rhetorical styles. Students format and put in order and describe the style. Highlight characteristics, transition markers, etc. 2. Edit some sample paragraphs. Level V 1. Iden tify educational opportunities and resources at the University, including the library, computer resources, and support centers 2. Read and understand a schedule of classes. 3. Locate resources, including books, journals, newspapers, and magazines, in the USF library via WebLuis and on site and via the Internet 1. Sign Up for Nicenet Open up sound recorder and record a short introduction about yourself. Save the file and send it via email to your sten to it and post a summary onto Nicenet. 2. University Scavenger Hunt (university resources, buildings, etc.) 3. Gather information from 2 other universities (select universities based on their needs). Students (in pairs) create worksheets for evaluati ng the university. Follow up: present their university choices as a poster presentation. 1. Prepare for the computer based exam multiple applications. 1. Understand bibliographic citations and use them as sources of information for research. 2. Summari ze and/or paraphrase journal articles and/or in field readings related to the topic of the research paper. Level III + 1. Give bibliographic information and they put in correct formats/templates 2. Write short essay; exchange; insert comments, highlight with color coding. 3. Create templates for bibliographic entries.

PAGE 599

573 Appendix 3A: Multiple Trait Rubrics for the Pre Treatment Paragraph Writing Prompt: The International Student organization has invited you to write a note for their newsletter. The topi c of your note is: My Favorite U.S. holiday. Based on your personal experience and the pre writing discussion you had with your partner, write the your first draft of a descriptive paragraph. You will have 30 minutes. Title: My Favorite U.S. Holiday S R hetorical soundness Presentation and development of the main point and its support Overall language use 6 The paragraph allows the reader to create a full and clear understanding of the statements are well supported with explicit descriptions. The main idea of the paragraph directly addresses the topic and is stated clearly and succinctly in a logically organized manner. Explicit transitions and/or connectors allow achieving the coherence of the paragraph. Supporting ideas examples, and/or explanations strengthen the main idea presentation. Language use is direct, fluid, and generally accurate. 5 The paragraph allows the reader to create favorite US holiday. However, some of the statem ents are not supported with an explicit description. The main idea of the paragraph is related to the topic. The paragraph is logically organized through the use of coherence markers. Explicit transitions and/or connectors allow achieving the coherence of the paragraph. The paragraph contains at least two supporting ideas, examples or explanations that are related to the main idea. Language control is good; vocabulary use is nicely varied. 4 The paragraph allows the reader to create a sufficiently clear un derstanding of the the description is sketchy which does not allow the reader to create a full mental image of the described feelings and/or events. The main idea of the paragraph is related to the topic but it could be stated in a more clear and explicit way. The paragraph may lack logic or coherence because connectors and transitions are not used consistently and effectively. The supporting ideas, examples or explanations that are related to the main idea are underde veloped. Language shows satisfactory but inconsistent control; vocabulary use shows adequate variety. 3 The paragraph allows the reader to create a somewhat sufficiently clear holiday. However, very limited des cription is provided and this description is not always related to the topic, this does not allow the reader to create a mental image of the described feelings and/or events. The main idea of the paragraph is only marginally related to the topic or it is d ifficult to identify. The paragraph does not have a clear organizational structure; coherence is weak due to absent or inappropriate transition signals. The paragraph lacks specific supporting ideas, examples or explanations that are directly related to th e main idea. Language shows inconsistencies that distract the reader; vocabulary shows lack of variety. 2 The paragraph does not allow the reader to create a sufficiently clear holiday due to lack of description t hat is related to the topic. The paragraph lacks main idea that is related to the topic. The supporting ideas, examples, or explanations are inappropriate. The text lacks organization and coherence. Language control is weak and frequently distracts the reader; vocabulary use is highly restricted. 1 Not a ratable sample Not a ratable sample Not a ratable sample

PAGE 600

574 Appendix 3B: Multiple Trait Rubrics for the Post Treatment Paragraph Writing Prompt: The International Student organization has invited you t o write a note for their newsletter. The topic of your presentation is : The People of the United States Based on the discussion you had with your partner using L.E.C.S. or ICA2, write the first draft of a descriptive paragraph th at would present the peo ple of the United States and will correspond with the pictures given for the discussion. Make sure to include in your description all the ideas presented by all the pictures. You can use the records of your L.E.C.S. or ICA2 discussion when com posing the pa ragraph. You will have 35 minutes to write this paragraph. Title: The People of the United States S Rhetorical soundness Presentation and development of the main point and its support Overall language use 6 The paragraph allows the reader to create a f ull and clear understanding of US developed based on the visual prompts given for discussion. All statements are well supported with explicit descriptions. The main idea of the paragraph directly addresses the topi c and is stated clearly and succinctly in a logically organized manner. Explicit transitions and/or connectors allow achieving the coherence of the paragraph. Supporting ideas, examples, and/or explanations strengthen the main idea presentation. Language u se is direct, fluid, and generally accurate. 5 The paragraph allows the reader to create a good understanding of the developed based on the visual prompts given for discussion. However, some of the statements are n ot supported with an explicit description. The main idea of the paragraph is related to the topic. The paragraph is logically organized through the use of coherence markers. Explicit transitions and/or connectors allow achieving the coherence of the paragr aph. The paragraph contains at least two supporting ideas, examples or explanations that are related to the main idea. Language control is good; vocabulary use is nicely varied. 4 The paragraph allows the reader to create a sufficiently clear understandin g the US developed based on the visual prompts given for discussion. However, the description is sketchy which does not allow the reader to create a full the topic. The main idea of the paragraph is related to the topic but it could be stated in a more clear and explicit way. The paragraph may lack logic or coherence because connectors and transitions are not used consistently and effectively. The supporting ideas, examples or exp lanations that are related to the main idea are underdeveloped. Language shows satisfactory but inconsistent control; vocabulary use shows adequate variety. 3 The paragraph allows the reader to create a somehow sufficiently clear understanding of the writ the people of the US developed based on the visual prompts given for discussion. However, very limited description is provided and this description is not always related to the topic, this does not allow the reader to create a mental image view on the topic. The main idea of the paragraph is only marginally related to the topic or it is difficult to identify. The paragraph does not have a clear organizational structure; coherence is weak due to absent or inappropriate transit ion signals. The paragraph lacks specific supporting ideas, examples or explanations that are directly related to the main idea. Language shows inconsistencies that distract the reader; vocabulary shows lack of variety. 2 The paragraph does not allow the reader to create a sufficiently clear the people of the US developed based on the visual prompts given for discussion. due to lack of description that is related to the topic. The paragraph lacks main idea that is related to the topic. The supporting ideas, examples, or explanations are inappropriate. The text lacks organization and coherence. Language control is weak and frequently distracts the reader; vocabulary use is highly restricted. 1 Not a ratable sample Not a ratable sample Not a ratable sample

PAGE 601

575 Appendix 3B, Multiple Trait Rubrics for the Post Treatment Paragraph (cont.) *The Rubrics listed in Appendix 3 A & B were designed closely following paragraph rating scale and multiple trait scoring guide pres ented by Ferris and Hedgcock (1998). This would increase the criterion related validity of the instruments.

PAGE 602

576 Appendix 4: Treatment CMC Task Task ACMC -Nicenet You will be asked to write a descriptive paragraph about the people of the US. To prepare for this writing assignment, you will brainstorm and plan with your partner using Nicenet. You and your partner have six pictures presenting specific ideas about the topic. Together you need to: (1) Discuss what ideas you should include in a paragraph that descri bes the people of the US and includes the ideas represented on the six pictures. (2) Discuss the paragraph organization, main and supporting ideas, and topic sentence. Note: It is important to participate actively and regularly in the discussion via Nicenet during the whole discussion period: starting today and completing the discussion either Monday night or Tuesday morning. Please discuss the pictures ONLY with your partner and ONLY on Nicenet. This assignment is due on Tuesday. Task ACMC L.E.C.S. You will be asked to write a descriptive paragraph about the people of the US. To prepare for this writing assignment, you will brainstorm and plan with your partner using L.E.C.S. You and your partner have six pictures presenting specific ideas about the topi c. Together you need to: (1) Discuss what ideas you should include in a paragraph that describes the people of the US and includes the ideas represented on the six pictures. (1) Discuss the paragraph organization, main and supporting ideas, and topic sentence Note: It is important NOT TO talk about the pictures with anybody. Please discuss them ONLY with your partner and ONLY on L.E.C.S. This assignment is due in 50 minutes.

PAGE 603

577 Appendix 4,Treatment CMC Task (cont.)

PAGE 604

578 Appendix 5: Additional Assignment Engl ish as a World Language ELI, Intermediate Writing Class Activity designed by Raymond Cepko Read the following text and decide if you agree with the author's point of view or not. Then go to the conferencing section and post your response under agree or d isagree. After you have finished posting your response, read what other students had to say and comment on them. Then go back to your post and see what responses you received. (NOTE: When replying to another students message, be sure to hit the REPLY butto n.) It is widely believed that "English is truly the world language." English seems to be emerging as a global language. If this were to become official, it would reduce the amount of mis communications and would make communication across cultures m uch easier. Language may be a cornerstone of culture, but the culture itself would not have to disappear if English were used as a second or third language for the purpose of communicating globally. Currently, "English is the only language used in i nternational air traffic control and is virtually the only language of a whole range of other activities from scientific research to pop music." English may not be the best choice, but it is the obvious choice, for an international language. Whether we like it or not, the English language is becoming the global language. adapted from http://iml.jou.ufl.edu/projects/students/Wheaton/page1.htm

PAGE 605

579 Appendix 6: Sample NVIVO Report NVivo revision 2.0.163 Project: CL_Analysis_Themes User: Administrator D ate: 1/26/2007 1:29:04 AM DOCUMENT CODING REPORT Document: ObservationNotes_ALL_CL Created: 1/10/2007 9:20:00 PM Modified: 1/26/2007 1:07:38 AM Description: Writing III A 2:00 4:00 Node: (1 5 1) /ClassDynamic_General/Group Work/Collaborati ve Work Passage 1 of 14 Section 0, Para 200, 286 chars. 200: Andrea Gets up and goes to the teacher to verify a vocabulary item in her outline. Passage 2 of 14 Section 0, Para 206, 171 chars. 206: The teacher gives the students a choice to use a computer and work alone or to work with a partner on the same computer. All of the students chose to work with a

PAGE 606

580 partner: Passage 3 of 14 Section 0, Para 209, 159 chars. 209: an her partner follows her adding to her statements and expands them. Passage 4 of 14 Section 0, Para 212, 33 chars. 212: This p air is very collaborative. Passage 5 of 14 Section 0, Para 212, 115 chars. 212: discussed and collaboratively constructed. Passage 6 of 14 Section 0, Para 213, 82 chars. 213: Passage 7 of 14 Section 0, Para 217, 97 chars. 217:

PAGE 607

581 Passage 8 of 14 Section 0, Para 220, 347 chars. 220: W e directed, while Passage 9 of 14 Section 1, Para 436, 39 chars. 436: [studen and work nicely together [3] [3] Internal DB: Within group collaboration with one student not participating. Passage 10 of 14 Section 1, Para 440, 89 chars. 440: composition process that is rather slow Passage 11 of 14 Section 1, Para 444, 111 chars. 444:

PAGE 608

582 Passage 12 of 14 Section 1, Paras 445 to 446, 248 chars. 445: For this group all the communication seems to go through the comp uter. 446: They communicate about the things that are already on the screen not the things Pass age 13 of 14 Section 1, Paras 618 to 619, 261 chars. 618: this will help them to d o the analysis easier. 619: Passage 14 of 14 Section 1, Paras 630 to 631, 160 chars. 630: name]. 631: about the different elements of the text.

PAGE 609

583 scores by Measure Participant TUL IUNum IUL L D VC RH PD Lang Average Gain Shin ACMC Low 1.072904144 1.010719788 1.288982119 2.268470683 4.99092967 1.591944144 0.948754042 0.492870616 1.21832844 Azad 0.722248647 0.596969891 0.892633663 2.088639634 0.047450799 1.104883524 0.816554665 1 .130901741 0.74447316 Isabella ACMC High 0.081919713 2.262577354 1.928343428 1.000449951 2.276017642 0.894340092 0.801043203 1.421222759 1.31275934 Felipa 1.753424533 0.473754858 0.956107686 3.096551718 0.173507933 3.944156651 4.870859907 0.29369732 1.78344188 Gains ACMC 0.017540208 0.039809887 0.122909217 0.08704579 0.063707522 0.021701206 0.111291485 0.148536338 0.01355427 Kamil SCMC Low 0.191620053 0.555898213 2.025924631 0.585669431 0.097329662 1.104883524 0.409554752 1.130901741 0.6 3600165 Ajwad 0.276310822 0.145181437 0.694937204 0.412313528 0.636829904 2.129985899 2.042664083 1.130901741 0.65682789 Suong ACMC High 0.657180269 0.679199241 0.500260276 0.028187002 2.901174683 0.919830659 0.816554665 1.130901741 0.46025022 Kang 0.766533188 0.17870166 0.547247349 1.859607826 2.115912528 0.919830659 0.409554752 1.346954354 0.64492224 Gains SCMC 0.095458738 0.071886564 0.262806804 0.208194316 0.155181904 0.0283785 0.145535019 0.098937669 0.05179591 Gains CMC A LL 0.031426002 0.008591908 0.044234392 0.040891589 0.031144563 0 0.00E+00 0.041297602 0.01476414

PAGE 610

584 scores by Measure (Continued) (1) TUL the syntactic complexity (measured by calculating the mean length of t units) (2) IUL the amount of information present in a single focus (measured by mean length of idea units) (3) IUNum the quantity of overall information present (measured by the number of idea units) (4) LD lexical information per clause (measured through lexical density a nalysis) (5) VC vocabulary complexity (measured by analyzing the frequency of the unique words used) (6) Rh rhetorical soundness (7) PD presentation and development of main ideas (8) Lang overall language use (the last three criteria were assessed using a multipl e trait rubric).

PAGE 611

585 Appendix 8: Azad ACMC Interview with the Researcher 1. How old are you? I'm 19. 2. In which country did you receive your education? In my country Qatar. 3. What did you study before coming to the ELI? In my elemanry school I studied ma ny subjucts like, History science and, but in my high school I choose science only. 4. How were your classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about you r classes and what you did not?) the teacher. 5. What is the importance of being a good writer for an educated person? No answer. 6. For you as a language learner, how imp ortant is it to develop good writing skills in English? Please explain your answer. No answer. 7. What is the most difficult thing for you when you write in English? No answer. 8. What is the most enjoyable thing for you when writing in English? No answe r. 9. How comfortable are you when working with computers? I'm can use the computer easiely, but not in every programs because there are some programs it is diffecult to me. 10. How were computers used at your previous school(s)? used the computers only in computer class.

PAGE 612

586 11. Did you use computers to communicate with other people (chat, email, list serves) before this class? I use the copmuter often. Chat and email. Not listserve 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and also posted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? No answer. 13. How good computer communication skills in English can help you (if at all) in the future? No answer. How did you feel completing the discussion of the pictures? How was it different from when you discuss things in class? I did a good job in that pictures. 14. Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your mes sages? No. 15. Did you find that your typing skills hindered your discussion on line? Yes it slou. 16. Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please explain your answer: What was very hel pful? What was distractive or not helpful? I did a good job. Easy. my partner is good we had not a time to talk. I like that he tell and I tell then. We do it together. 17. Do you think computer discussion activities are good for learning how to write bett er? Please explain your opinion. It was not very helpful, bacause we did not have a time to write many things. Yes,because it improved my spelling.

PAGE 613

587 18. Did you include in your writing some of the ideas that your partner shared with you? I write new ideas. 19. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures? Yes. 20. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thes aurus, etc.). Spell check.

PAGE 614

588 Appendix 9: Azad Training Task: ACMC Posting Task: Read the following text. Do you agree or disagree with the author? Post your opinion on Nicenet, Conference topic: English as a World Language: Agree or Disagree. It is wi dely believed that "English is truly the world language." English seems to be emerging as a global language. If this were to become official, it would reduce the amount of mis communications and would make communication across cultures much easier. Langu age may be a cornerstone of culture, but the culture itself would not have to disappear if English were used as a second or third language for the purpose of communicating globally. Currently, "English is the only language used in international air traffic control and is virtually the only language of a whole range of other activities from scientific research to pop music." English may not be the best choice, but it is the obvious choice, for an international language. Whether we like it or not, the Engli sh language is becoming the global language. Text adapted from : http://iml.jou.ufl.edu/projects/students/Wheaton/page1.htm

PAGE 615

589 FROM: Azad (02/16/06 1:28 PM GMT 06:00) [ Send a personal message to Azad ] SUBJECT: English as a world language [ Edit | Dele te ] I think that English can be the world language very soon for several reasons.First,English is one of the easiest languages in the world. Second, it is very famous language and it is common to use it in many places. Third, many schools, companies,and U niversities use it in different ways around the world.finally, there are many students study in English in some Universities in U.S. or in other countries which use the English as the first language. English will be the world language, but it is will be th e second or the third language in these countries.

PAGE 616

590 Appendix 10: Shin ACMC Interview with the Researcher 1. How old are you? I'm 38years old 2. In which country did you receive your education? Korea 3. What is the highest degree you have? In what area? I graduated a university. My major was nursing. 4. How were your classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about your classes and what you d id not?) In our country, classes has organized the same in america. In nursing schol, We have two kinds of things that are practice and theory. 5. What is the importance of being a good writer for an educated person? Very important for study 6. For you as a language learner, how important is it to develop good writing skills in English? Please explain your answer. The language learner have to study a lot of vocabulary and words for improve good writing skills. 7. What is the most difficult thing for you when you write in English? I am too difficult how I can use world order correctly. 8. What is the most enjoyable thing for you when writing in English? I like to write about the trip. 9. How comfortable are you when working with computers? When I began the ELI, I was uncomfortable. However, now I can use computer more easer than past. 10. How were computers used at your previous school(s)? I did not use computer before I study the ELI. 11. Did you use computers to communicate with other people ( chat, email, list serves) before this class?

PAGE 617

591 I usually use computer when I send to E mail with my family and friend. 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and also posted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? I did more enjoy Nicenet because I can not type quickly 13. How good computer communication skills in English can help you (if at all) in the future? I need to have a good computer comm unication skill because I have to give important information to my patients. 14. How did you feel completing the discussion of the pictures? How was it different from when you discuss things in class? I was uncomfortable because I can not use computer very well. In class I can not speak if I want. In computer I have to. 15. Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your messages? I used electronic dictionary. 16. Did you find that you r typing skills hindered your discussion on line? I did find. 17. Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please explain your answer: What was very helpful? What was distractive or not helpful? I was helpful about sharing information with my partner. 18. Do you think CMC activities are good for learning how to write better? Please explain your opinion. I think that discussing face to face is better then on line discussion. Easi er no type. 19. Did you include in your writing some of the ideas that your partner shared with you? Yes I did. 20. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures?

PAGE 618

592 When I did not understand some information, I asked to my partner. 21. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thesaurus, etc.). I used grammar ,spell check and formatting.

PAGE 619

593 Appendix 1 1: Ordered Comparison Matrix Name Gend er Educatio n & Field of Studies Treatment CMC Pattern Treatmen t Partner's Pattern Computer Experience Task Helpful Task Short comes Task Attitude CMC Paragra ph Organiz ation Paragr aph Organi zation Overall Number of T urns Actor 's Turn s Paragr aph IU # ACMC Low Azad M HS science orientatio n Dominant Dominant Experienced Yes Satisfied & 11 5 11 Shin F P nursing Dominant Dominant Limited Experience Yes Uncomfortabl e O own = 5 3 20 ACMC High Felipa F C college Dominant Dominant Experienced Yes Positive O & 2 1 18 Isabella F P marine biology Do minant Collaborat ive Experienced Yes ? (about general CMC tasks) Positive O P= 2 1 38 SCMC Low Ajwad M P industrial ingeneerin g Collaborati ve Collaborat ive Yes Positive O & 82 39 21 Kamil M C college Pa ssive Dominant Experienced OK boring Low motivation and engagement & 10 5 13 SCMC High Kang M C college Collaborati ve Collaborat ive Experienced Not known Not known Not known & 31 19 12 Sun F P informatio n systems Collaborati ve Collaborat ive Ex perienced Yes Good for discussio n not writing Comfortable & 83 41 29

PAGE 620

594 Appendix 11: Ordered Comparison Matrix (continued) Name Gend er CMC IU# Total Partn er CMC IU# Total Actor CMC IU Conte nt Partne r CMC IU Conte nt Actor CMC IU Partn er's Exact % of t otal Pragr aph IU CMC IU Partne r's Rephr ased % total Pragra ph IU CMC IU own Exact % of total Pragrap h IU CMC IU Own Rephras ed % total own IU CMC IU New % total IU Lexical Interte xtual Conne ctions Rephrasin g Strategise New Ideas Improved Use of pre treat ment discu ssion ACMC Low Azad M 15 10 11 8 0 2 18.18% 1 9.1 % 5 45.45% 3 27.27 % 26.09 Low Syn Synt Short IU CCh NT IUL LD DIB Shin F 18 15 8 8 0 5 25% 1 5% 5 25% 9 45% 34.95 LowAv Syn Synt, Short IU CCh CIU EIU Mreph NT N NC PD CB+ Under i ACMC High Felipa F 39 35 32 26 0 0 4 22.22% 8 44.44% 6 33.33 % 27.27 Low Syn Synt EIU CIU Gr NT N NC TUL IUL LD Rh PD Lang DIB Isabel la F 28 38 13 21 0 2 5.26% 0 14 36.84% 22 57.89 % 44.68 Higher DevIU Gr EIU ShortIU, Syn Synt NC NOSum, NOPSum NT N IUL IUNum LD VC Rh PD and Lang DIB+ CB SCMC Low Ajwa d M 57 50 19 20 0 1 4.76% 0 1 4.76% 19 90.48 % 20.87 Low Syn Synt N NT IUL LD CB+ Under i Kamil M 9 8 7 3 0 0 0 0 13 (100 %) 18.92 Low 0 N NT NC LD PD DOI ACMC High Kang M 32 28 16 14 0 1 8.33% 1 8.33% 3 25% 8 66.67 % 21.43 Low Syn & Synt N NT NIU LD VC Rh PD Lang CB Sun F 65 54 25 13 0 3(10.34 %) 0 4 (13.79) 22 (75.86 %) 16 .52 Low Syn & Synt Syn, Synt, & Merge, ShortIU & Synt ShortIU, EUI, DevIU N, NT, NIU TUL IUL IUNum VC Rh CB

PAGE 621

595 Appendix 11: Ordered Comparison Matrix (continued) CMC IU Content Partner CMC IU Content Own CMC IU Partner's Exact total Paragra ph IU CMC IU Partner's Rephrased total Paragraph IU CMC IU own Exact total Paragraph IU CMC IU Own Rephrased total own IU CMC IU New % total own IU Lexical Intertextual Connections Rephrasing Strategies New Ideas Improved Use of pre treatment discus sion 11 8 0 2 18.18% 1 9.1 % 5 45.45% 3 27.27% 26.09 Low Syn (4 36.4%), Synt (4 36.4%), Short IU (3 27.3) CCh (1 9.1) NT (1 9.1) IUL LD DOI 8 8 0 5 25% 1 5% 5 25% 9 45% 34.95 LowAv Syn (1 5%), Synt (3 15%), Short IU (2 10% ), CCh (1 5%), CIU (1 5%), EIU (2 10%), Mreph (5 25%) NT (4 20%), N (4 20%), NC (1 5% ) PD CB+Under i 32 26 0 0 4 22.22% 8 44.44% 6 33.33% 27.27 Low Syn (6 33.33%) Synt (5 27.78%), EIU (4 22.22%), CIU (1 5.56), Gr (1 5.56%) NT (4 44.44%), N (1 5.56%), NC (1 5.56%) TUL IUL LD Rh PD Lang DOI 13 21 0 2 5.26% 0 14 36.84% 22 57.89% 44.68 Higher DevIU (6 15.79%), Gr (4 10.53%), EIU (3 7.89%), ShortIU (3 7.89%), Syn (6 15.79%), Synt (2 5.26%) NC (2 5.26%), NOS um (4 10.53%), NOPSum (1 2.63), NT (6 15.79%), N (10 26.32%) IUL IU LD VC Rh PD and Lang DOI+CB

PAGE 622

596 Appendix 11: Ordered Comparison Matrix (continued) Key Abbreviations: Collaborative brainstorming CB Individual brainstorming IB Background: (1) College C, (2) Professional P, High school HS, Keyboard skills CMC Task Perceptions: (1) Limited Time (3)Confused in the beginning, may create confusion ?, (4) Underdeveloped ideas Underi Gains (1) TUL the syntactic complexity (measured by calculating the mean length of t units), (2) IUL the amount of information present in a single focus (measured by mean length of idea units), (3) IUNum the quantity of overall information present (measured by the number of idea units), (4) LD lexical information per clause (measured t hrough lexical density analysis), (5) VC vocabulary complexity (measured by analyzing the frequency of the unique words used), (6) Rh rhetorical soundness, (7) PD presentation and development of main ideas, (8) Lang overall language use (the last t hree criteria were assessed using a multiple trait rubric). Lower than average LowAv CMC Discussion Use -> Paragraph: (1) Development of own ideas DOI, (2) Individual Brainstorming IB, (3) Discussed O, (4) Not discussed (5) Partner's suggestio n followed =P, (6) Order from CMC followed =, (7) Order from CMC shuffled &, (8) New extends discussed topic NT, (9) New entirely N, (10) New for cohesion NC (10) New Summary of own discussion, own partner, partner NOSum, NOPSum, NPSum Reph rasing Strategies: (1) Shortening idea units Short IU, (2) Synonyms and synonymic phrases Syn ; (3) Syntactic change Synt ; (4) Combining IU CIU; (5) Extending IU EIU; (6) Major Rephrasing Mreph, (7) Content Change CCh; ( 8 ) D evelopment of one ide a unit into more DevIU ; (9) Successful grammar correction Gr; (10) Coherence/cohesion devices addition Coh

PAGE 623

597 Appendix 12: Felipa ACMC Interview with the Researcher 1. How old are you? 2. In which country did you receive your education? 3. What is the highest degree you have? In what area? 4. How were your classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about your classes and what you did not?) No answer. 5. What is the importance of being a good writer for an educated person? No answer. 6. For you as a language learner, how important is it to d evelop good writing skills in English? Please explain your answer. No answer. 7. What is the most difficult thing for you when you write in English? No answer. 8. What is the most enjoyable thing for you when writing in English? No answer. 9. How comfo rtable are you when working with computers? I feel very good because I like the computers and the internet because you can find everything that you like shopping friends etc. 10. How were computers used at your previous school(s)?

PAGE 624

598 use to much the computers because my school was a little monotonous and antiquated (nuns) but I learn a little about computers at home with my sister because she study about this in the university and about if. 11. Did you use computers to communicate with oth er people (chat, email, list serves) before this class? This type of communication is called Computer Mediated Communication and I will use CMC when talking about it. You can do the same in your answers. I like use the chat yes I love chat with my family after the class because it is only my free time for use CMC and this is very important for me because in this way you never are feeling alone 12. How did you feel completing the discussion of the pictures? How was it different from when you discuss things in class? I felt very good but i needed time for understood each question but the work was very nise because my brain was confused 13. Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your messages ? My husband and the dictionary were help me a little but thanks because in this form i can learn more about the US. 14. Did you find that your typing skills hindered your discussion on line? No I was ok. 15. Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please explain your answer: What was very helpful? What was distractive or not helpful? No answer. 16. Do you think CMC activities are good for learning how to write better? Please explain your opinion. No answer. 17. Did you include in your writing some of the ideas that your partner shared with you?

PAGE 625

599 No answer 18. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures? No answer. 19. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thesaurus, etc.). No answer.

PAGE 626

600 Appendix13: Felipa FROM: Teacher (01/16/06 2:19 PM GMT 06:00) SUBJECT: Writing Class Please tell me what you hope to learn in this writing class. What are some language skills that are difficult for you that you would like to improve? What are some English skills that are easier for you that you could help other students with? Click on "Post a New Message" below to type your response and remember that other students will read what your write. FROM: Felipa (01/23/06 1:29 PM GMT 06:00) SUBJECT: Felipa hello everybody I would like to tell you. That i want to know everything about writing. sometimes i am comfuse with the vocabulary and the order of the words. i want to know about how i can write in past and present progresive. and also i what to know about punctuation because mine it's terrible. thanks a lot Felipa

PAGE 627

601 Appendix 14: Isabella ACMC Interview with the Researcher 1. How old are you? I am 28 2. In which country did you receive your education? I received my education in Italy. 3. What did you study before coming to the ELI? I have a degree in biology and I have done an sperimental thesis in marine biology specifically in bioacoustics. 4. How were your classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your clas ses? What did you like about your classes and what you did not?) The teaching in Italy is very different if you are in the high school or in the University.In the high school you are much followed: th e teacher says you what you have to do during the cla ss and at home. You have the interrogation also every day and there are some afternoon activity like sports, informatics or theather. At the University often you can have any contact whit your teacher. You can prepare you exam alone in your home for the fi rst years. The teacher racommend to the alumns to be present but it is not an obbligation. For the last years when you should choose your specialization the attendance is necessary. The teacher should know you becouse often will be him to follow you during the thesi's degree. You can also choose to do the thesis outside the institution of the University, like me, and in my opinion this is the best mood to know who the world of work go on and after the thesis often you can stay in that laboratory for working whit them.. 5. What is the importance of being a good writer for an educated person? For me being a good write means to known exactly what I want to explain and to be able to addend a lot of possible example and argument to enrich the main idea 6. For you a s a language learner, how important is it to develop good writing skills in English? Please explain your answer. The important thing to develop is the grammar becouse if you have a good idea to explain but you don't know how to explain it nobady could un derstand you

PAGE 628

602 7. What is the most difficult thing for you when you write in English? The most difficult I have when I write in English is to rende in corret grammar what I want to say. 8. What is the most enjoyable thing for you when writing in English? T he most enjoyable thing is excercising writing in English to impruve my writing capacity. Have a good writing capacity means to use a good language property, use right words and term and be more clear as possible in order to explain you idea. 9. How comforta ble are you when working with computers? I became to use the computer only 2 years ago when I started to write my thesi's degree and I have to say that now I can't stop to use. I always use computer: to send e mail, to study and to have news from world e very day. Now I can be able to use some software that before I didn't know and the use of them has changed my mood of study. For example I use Exel for calculation or word for writting all thing that before I use to do on paper but using less time. Other I often do presentations in my language about the biology and I use Power point (that I love!). In conclusion I am really comfortable working with computers now. 10. How were computers used at your previous school(s)? In my school the computer was few used. There were specific schools in which you could learn to use it but for example in scientific high scool, that I attended, is not used or teached.In the University we use to write our paper for most of teacher on paper yet. Now the reality is changing becou se already from the medium school the guys learn the use of computer. 11. Did you use computers to communicate with other people (chat, email, list serves) before this class? I started to use the computer using Internet and to send e mail. Through the comp uter infact I can get news about master degree in biology, where the research is better and communitate with some researchers who live in the other part of the world. 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and also po sted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? Normally I don't you a lot the chat becouse I don't have much time in the school and I don't have personal computer in my apartment yet. So rig ht now I have only used the ICA2. I can't say what I prefer but I know that also in Italy where I have more time I usually don't use the chat.

PAGE 629

603 13. How good computer communication skills in English can help you (if at all) in the future? I think that the co mputer will be always more important: people could chat using the computer, send their documents when they aren't in the office and study at home using Internet. 14. How did you feel completing the discussion of the pictures? How was it different from when y ou discuss things in class? HAVE PROBLEM ABOUT THE CHOICE OF THE ORGANIZATION OR WHAT PUT IN THE PARAGRAPH. WE WERE SUDDELY AGREE ABOUT THE CHANGES TO DO IN OUR PARAGRAPH AND HOW PROCE ED. I FEEL REALLY GOOD USING THE ONLINE DISCUSSION BECAUSE YOU CAN EXPLAIN YOUR OPINION WITHOUT BEING IN THE SAME PLACE AT THE SAME TIME. 15. Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your mes sages? YES I USED INTERNET IN ORDER TO FIND SYNONISM, AND NEW WORDS. 16. Did you find that your typing skills hindered your discussion on line? No answer. 17. Was the computer discussion activity you did helpful for you when getting ready to write the assig ned paragraph? Please explain your answer: What was very helpful? What was distractive or not helpful? Sometimes it could be helpful because you can share your ideas with your partner and he/she could help you writting better. Somethimes, on the contrar y, it could be distractive because you could miss much time to chat and discuss about the topic loosing time needed for writting. 18. Do you think computer discussion activities are good for learning how to write better? Please explain your opinion. in my o pinion is different if you speak stundent to student or if you speak on line with your teacher. sometimes When you speak with your classmate could be difficult to make understand and you can loose your time nedeed for the writting. On the contrary when you speak with your teacher is always helpful because your teacher could help you to impruve your writting but also your grammar.

PAGE 630

604 19. Did you include in your writing some of the ideas that your partner shared with you? Yes, my partner helped me changing the o rganization of the paragraph and it was really helpful. 20. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures? I compared my suggestions and I review these in order to impruve the writting It was 21. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thesaurus, etc.). I use all of them and I also use a site m w.com in order to find the definitions of words I don't know.At the end for me is really important to use the grammatical correction and counting of the words.

PAGE 631

605 Appendix 15: Ajwad ACMC Interview with the Researcher and Notes from the Face to Face Interview 1. How old are you? 27 2. In which country did you receive your education? King Abdulaziz University Saudi Arabia 3. What did you study before coming to the ELI? industrial Engineering 4. How were your classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about your classes and what you did not?) No answer 5. What is the importance of being a good writer for an educated person? For me it i idea to large number from the people I focus to them, so if I am a good writer and I have good idea the people will resave my idea. 6. For you as a language learner, how importa nt is it to develop good writing skills in English? Please explain your answer. It is very important to me, because if I write in any subject even if I tack deferent position I will found many writer wrote in the same subject so I need in that time to b e unique in that subject, In idea, in the steppes to present my subject, make my subject easy to understand witch mean good skills in writer. 7. What is the most difficult thing for you when you write in English? Spilling and how to make my idea in Engl ish 8. What is the most enjoyable thing for you when writing in English? If I can delivery my idea to the people, and I can effect to the people so I make change.

PAGE 632

606 9. How comfortable are you when working with computers? Very comfortable, and I like do do every thong by computer, even if I start like new writing. 10. How were computers used at your previous school(s)? we used the computer as a basic thing. In the first class in my field, the chairmen of industrial engineering was the instructors, and I rem ember what he said about using home, he can transfers to another department, and we are welcome to help any student to I remember the first time I use the c omputer since I was 13 years old, my father send me to the computer center in King Abdulaziz University in the summer to take Course about DOS and then windows 3.1 skills, and then Microsoft office. And I remember the first time I deal with the enter net was to open my email. 11. Did you use computers to communicate with other people (chat, email, list serves) before this class? Yes, to contact or take appointment 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and also po sted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? Answer provided during the face to face interview (see the end of the Appendix). 13. How good computer communication skills in English can help y ou (if at all) in the future? No answer provided. 14. How did you feel completing the discussion of the pictures? How was it different from when you discuss things in class? It was something new for me, to see the word in the static positions, and som e time I need to stand out of habit, just to cut the rotten. Also it give me good chance to listen to the people in deferent ways so at least I can look to the picture from deferent ways. 15. Did you use anything additional an electronic dictionary, the In ternet, a regular dictionary, when writing your messages? NO.

PAGE 633

607 16. Did you find that your typing skills hindered your discussion on line? Absolutely yes, because when I speak with the people we have million and million idea to discus, and we need to spea k continuously and quickly to represent all these idea, but in the computer I thing it is limited. 17. Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please explain your answer: What was very helpful? What was distractive or not helpful? No answer provided. 18. Do you think computer discussion activities are good for learning how to write better? Please explain your opinion. No answer provided. 19. Did you include in your writing some of the ideas that your partner shared with you? No answer provided. 20. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures? Answer provided during the face to face interview (see the end of the Appendix) 21. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thesaurus, etc.). Additional face to face interview Ajwad liked more chatting but did not like the LECS in terface. He thinks that completing discussion online is important for learning how to write and allows to organize the ideas better. He uses spell check. Used the ideas of his partner According to Ajwad, people are difficult to lead because they need a lot of choices and will obey but I will not be creative because this is his idea, but if he shares a project with me in the beginning, start working with him and this project will be my project too and I Further, Ajwad continued the analogy, he said that the ELI was very structured and did not consider the specifics of students. He was referring mainly to the Saudi students because their relativ ely large group of the ELI. He thought that more space for being creative would help him and the other students from Saudi Arabia. The structure of the classes lowered his motivation and the motivation of others. He thought that this was

PAGE 634

608 disrespectful from children who were unable to make their own decision. Ajwad volunteered for this short interview at the end of one of the classes I observed after the treatment was over, he was eag er to share his opinion with me.

PAGE 635

609 Appendix 16: Kamil ACMC Interview with the Researcher 1. How old are you? I`m 19 years 2. In which country did you receive your education? Kuwait 3. What did you study before coming to the ELI? Finance 4. How were y our classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about your classes and what you did not?) It was really very good organized yeah some times my tea cher lecture duing in my class, actually I like the system in my class and I hate when some one try to bother the student 5. What is the importance of being a good writer for an educated person? I think We must write a lot to improve writing skills 6. Fo r you as a language learner, how important is it to develop good writing skills in English? Please explain your answer. Read books and write a lot 7. What is the most difficult thing for you when you write in English? I've problem with the spieling 8. W hat is the most enjoyable thing for you when writing in English? Actually I write anything cause it makes me bored ,But I enjoyable just when I write my name in English. 9. How comfortable are you when working with computers?

PAGE 636

610 Very comfortable 10. How we re computers used at your previous school(s)? No answer. 11. Did you use computers to communicate with other people (chat, email, list serves) before this class? Y es 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and al so posted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? No answer. 13. How good computer communication skills in English can help you (if at all) in the future? Actually the computer is really important for our life and it`s very easy to done the works with it that why i prefer to work by computer because i get used to work with it in my high school and in my life .Also the computer chek my spelling mistake finnaly i prefer the computer becau se we learn more with it 14. How did you feel completing the discussion of the pictures? How was it different from when you discuss things in class? It was ok. Little boring 15. Did you use anything additional an electronic dictionary, the Internet, a regu lar dictionary, when writing your messages? Nothing 16. Did you find that your typing skills hindered your discussion on line? N o 17. Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please exp lain your answer: What was very helpful? What was distractive or not helpful? Some of my partner ideas were helpful.

PAGE 637

611 18. Do you think computer discussion activities are good for learning how to write better? Please explain your opinion. k so. nobody says if I make mistakes or not. 19. Did you include in your writing some of the ideas that your partner shared with you? not all of them only if agree. 20. When you wrote your paragraph, how did you use the messages you and your partner wrote whe n discussing the pictures? Only best ideas we had 21. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thesaurus, etc.). Spell check and synonyms

PAGE 638

612 Appendix 17: Kang ACMC I nterview with the Researcher 1. How old are you? I'm 24 years old. I was born in Dec,1981 2. In which country did you receive your education? I have educated in korea since 1988. 3. What did you study before coming to the ELI? My major is mach anical engineering. So, I studied a lot of subject about engineering. And I have been study english to get a score TOEIC(Test Of English for International Communication) test which is mandotory to get a job in South Korea. 4. How were your classes organi zed in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about your classes and what you did not?) In korea, the classes is usually lecture. It have a merit that a student can learn many content for a short time. But because of that many krean student have a passive trend. I think it is a problem. 5. What is the importance of being a good writer for an educated person? To be a good writer, there are essential codition. Such as vocabulary, gammer, phase verb,. 6. For you as a language learner, how important is it to develop good writing skills in English? Please explain your answer. We should read many kinds of English book. And Trying to writting somthing english at all time. 7. What is the most difficult thing for you when you write in English? translation. I think of writing somthing as Korean. and then try to translate from korean to english. 8. What is the most enjoyable thing for you when writing in English?

PAGE 639

613 I don't know. Because I feel difficuty to write english. But, I will be satisfied when someone can understand what to say. 9. How comfortable are you when working with computers? No answer. 10. How were computers used at your previous school(s )? No answer. 11. Did you use computers to communicate with other people (chat, email, list serves) before this class? No answer. 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and also posted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? No answer. 13. How good computer communication skills in English can help you (if at all) in the future? No answer. 14. How did you feel completing the discussio n of the pictures? How was it different from when you discuss things in class? No answer. 15. Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your messages? No answer. 16. Did you find t hat your typing skills hindered your discussion on line? No answer. 17. Was the computer discussion activity you did helpful for you when getting ready to write the assigned paragraph? Please explain your answer: What was very helpful? What was distra ctive or not helpful?

PAGE 640

614 No answer. 18. Do you think computer discussion activities are good for learning how to write better? Please explain your opinion. No answer. 19. Did you include in your writing some of the ideas that your partner shared with you? No answer. 20. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures? No answer. 21. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & pas te, formatting, spell check, Thesaurus, etc.). No answer.

PAGE 641

615 Appendix 18: Sun ACMC Interview with the Researcher 1. How old are you? 27 2. In which country did you receive your education? T wiwan 3. What did you study before coming to the ELI? Manageme nt information system 4. How were your classes organized in your country? (Did your teacher lecture during your classes? Did you do group activities during your classes? What did you like about your classes and what you did not?) no lecture. yes, we do gr oup activities like > discuss with professro and classmate not like > only listen what the professor said. 5. What is the importance of being a good writer for an educated person? have personal and specific opinion; have a lot of vocabulary to describe what he/she want to say 6. For you as a language learner, how important is it to develop good writing skills in English? Please explain your answer. read more make me know how the other author write 7. What is the most difficult thing for you when you write in English? vocabulary are not enough 8. What is the most enjoyable thing for you when writing in English? I can learn more and more vocabulary and let the other know what is my opinion. 9. How comfortable are you when working with computers? I am p retty comfortable when working withe computers. Actually, my job is computer software designer and I work use computers everyday. I will uncomfortable if no

PAGE 642

616 computers can be used anymore. 10. How were computers used at your previous school(s)? We used com puter to design software, do homework, look for information for presentation. 11. Did you use computers to communicate with other people (chat, email, list serves) before this class? yes, I used computers to communicate with other people, 1) chat with my cl assmates, friends, customers and my team members 2) email to my classmates, friends, customers and my team members. For business purposes and personal communication. 3) Using the online banking services, communicate with banks 4) using the online stores to buy something, like books. 12. You have chatted with L.E.C.S. (the program you used to interview a partner) and also posted on Nicenet in this class. What did you like more: chatting with L.E.C.S. or conferencing on Nicenet? Why? t like LECS 13. How good computer communication skills in English can help you (if at all) in the future? No answer. 14. How did you feel completing the discussion of the pictures? How was it different from when you discuss things in class? I found it wa s easy to discussion of the pictures. Nothing expect I must increase my vocabulary. 15. Did you use anything additional an electronic dictionary, the Internet, a regular dictionary, when writing your messages? yes, I used an electronic dictionary and the online dictionary (webster online) to help me when writing my messages. 16. Did you find that your typing skills hindered your discussion on line? no, never. My typing skills are very well. 17. Was the computer discussion activity you did helpful for you when getting ready

PAGE 643

617 to write the assigned paragraph? Please explain your answer: What was very helpful? What was distractive or not helpful? On my opinion, it is helpful. When I discussioning activity, I can get more information and different opinions for ot her people. 18. Do you think computer discussion activities are good for learning how to write better? Please explain your opinion. No, I don't think so. Discussion activities are like chats. I can get for information but it's not useful for learning how to write better. 19. Did you include in your writing some of the ideas that your partner shared with you? Yes. 20. When you wrote your paragraph, how did you use the messages you and your partner wrote when discussing the pictures? To write more sentenc es and words to describe the ideas which we shared. 21. When you wrote your paragraph, which help features of Microsoft Word did you use (i.e. cut & paste, formatting, spell check, Thesaurus, etc.). I use cut, paste, spell check to help me in writing.

PAGE 644

618 Appendix 19: Summary of Quantitative Stage Findings Research Question Analysis Findings Research Question 1: What is the difference in the syntactic complexity present in the post treatment paragraphs of the students who participated in the synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of ANCOVA Failed to reject the Null Hypothesis at Research Question 2: What is the difference in the amount of information present within a single focu s of the post treatment paragraphs of students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of ANCOVA Failed to reject the Null Hypothesis at Research Question 3: Wh at is the difference in the quantity of the overall information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of ANCOVA Failed to reject the Null Hypothesis at Research Question 4: What is the difference in the lexical information present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of t drafts: ANCOVA Failed to reject the Null Hypothesis at Research Question 5: What is the difference in the vocabulary complexity present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of ANCOVA Failed to reject the Null Hypothesis at Research Question 6: What is the difference in the rhetorical soundness present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of Mann Whitney U test Failed to reject the Null Hypothesis at Research Question 7: What is the difference in the present ation and development of the main point and its support present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of Mann Whitney U test Failed to reject the Null Hypothesis at Research Question 8: What is the difference in the overall language use present in the post treatment paragraphs of the students who participated in synchronous versus asynchronous CMC pre writing discussion? Quantitative analysis of rst drafts: Mann Whitney U test Failed to reject the Null Hypothesis at

PAGE 645

619 Appendix 20: Summary of Qualitative Research Findings Research Question Analysis Findings Research Question 1: perceptions of the role of CMC in the process of establishing their writing skills? Qualitative analysis of interviews 1. CMC discussion supported the writing of their first paragraph drafts in the following ways: Supported the generation of ideas Students understood better the t opic of the written assignment The written format of the discussion supported their paragraph writing 2. There were students who perceived as more appropriate to incorporate only their own idea s into the paragraph. The majority of these students participated in an ACMC discussion and demonstrated dominant patterns of interaction. 3. Four main reasons for the CMC pre writing being unnecessary were pointed out: They experienced difficulties formulati ng ideas in a text based environment The text based mode of communication was time consuming Teacher feedback was identified as more valuable than interactions with peers Writing was defined as an individual activity that does not require peer collaborati on Research Question 2: What patterns of dyadic interaction do participants manifest during the asynchronous and synchronous CMC interaction process? Qualitative analysis of interactions The patterns of dyadic interaction were strongly inf luenced by mode of interaction: The SCMC discussion promoted opportunities for more collaboration The ACMC discussion, led to more dominant patterns A mutuality mismatch the trend was stronger in ACMC discussions Disengagement (observed mostly in the A CMC group ) students abandoned the posted only once and would not engage in further discussion.

PAGE 646

620 Appendix 20: Summary of Qualitative Research Findings (continued) Rese arch Question Analysis Findings Research Question 3: What are the factors that influence the CMC pre writing interaction process? How do these factors influence the interaction process? Qualitative analysis of interactions & interviews 1. CM C mode of interaction the most dominant factor that influenced the pre writing discussion: The ACMC mode: triggered postings that used longer strings of words often organized in complete sentences *the number of unique words was lower led to more d ominant patterns of interaction the collaborative students contributed longer postings demonstrating multiple collaborative moves in one posting *the dominant students would more often abandon the discussion space after posting once *there were more inst ance of mutuality mismatch The SCMC mode: *the electronic utterances were much shorter the electronic utterances did not always comprise complete sentences generated longer discussions, with higher number of turns students demonstrated a stronger t endency to collaborate with each other, the collaborative moves were frequently spread out through separate postings. *the dominant students would make several postings but would demonstrate inability to engage into a discussion with their partner 2. The u se of computer as an interaction medium: students reported difficulties related to typing skills difficulties expressing themselves in a text based environment CMC interaction being more time consuming and challenging Some students were less likely to use the computer in the early stages of the writing process

PAGE 647

621 Appendix 20: Summary of Qualitative Research Findings (continued) Research Question Analysis Findings Research Question 3: (cont.) What are the factors that influence the CMC pre writing int eraction process? How do these factors influence the interaction process? Qualitative analysis of interactions & interviews 3. Partner interaction inability or the unwillingness of one of the partners to engage in the interaction resulted in in teractions that would not develop completely the topic 4. Opinion of the nature of writing no relation found between patterns of dyadic interaction and attitudes towards writin g in the group of students who expressed overall positive opinion about writing there was a relation between the dominant dyadic pattern of interaction and the less view on writing as an individual experience

PAGE 648

622 Appendix 21: Summary of Case Study Re search Findings Research Question Analysis Findings Research Question 1: How do peers participate in synchronous and asynchronous CMC interactions? Instrumental Case studies: interviews, intertextual analysis of CMC interactions and first drafts 1. The CM C mode of interaction affected strongly the nature of dyadic patterns of interaction. 2. The ACMC actors, regardless of their post treatment gains, were mostly concerned with their own postings, demonstrating dominant patterns of interaction. 3. The SCMC actors were more likely to participate in a collaborative mode of interaction. 4. The collaborative dyadic patterns SCMC actors demonstrated during the interactions were not necessarily related to their post treatment gains. 5. The passive SCMC actor had a low treatm ent gain. 6. The mode of interaction may have influenced the number of postings. 7. The high gain ACMC actors and the collaborative SCMC actors, regardless of their gain, contributed significantly to the discussion. 8. The low gain ACMC actors and the passive SCMC actor failed to contribute significantly to the interaction on the assigned topic.

PAGE 649

623 Appendix 21 : Summary of Case Study Research Findings (continued) Research Question Analysis Findings Research Question 2: (a) How do they use the specific ideas and l anguage generated during these interactions in their writings? (b) What are the differences and similarities in the implementation of these ideas? Instrumental Case studies: intertextual analysis of CMC interactions and first drafts 1. The high gain ACMC ac tors demonstrated higher level of lexical intertextuality as compared with the rest of the actors. 2. On organizational level most of the actors, regardless their mode and gain level showed no intertextual connections. In addition, organization was more lik ely to be discussed by ACMC actors. 3. Trends of intertextual connections on idea units level: (1) The ACMC actors tended to use more high level intertextual connections by borrowing verbatim or rephrasing idea units from the discussion; (2) Medium level i ntertextual connections were more likely to be established in (3) High gain SCMC participants were more likely to establish high and medium levels of intertextual connections, as compared to the low gain participants who parti cipated in the same CMC mode; (4) High gain actors, regardless to the CMC mode, were more likely to move beyond the single idea unit boundaries when paraphrasing; (5) High gain ACMC actors were able to implement some grammar correction into the idea unit s when rephrasing them; (6) Collaborative actors, who were all SCMC actors, were able to consider and (7) The collaborative actors, who also had high gains, were able to conside r the low gain collaborative SCMC actor; (8) Most of the actors developed their paragraphs beyond the topics discussed during the pre writing interactions.

PAGE 650

624 Appendix 22: Ad ditional Quantitative Findings Research Question Analysis Findings Additional Research Question: Do the ACMC and the SCMC modes of the pre writing discussion influence to a different degree the texts produced after those discussions as measured by matchi ng distinct lexical items? Quantitative analysis of drafts ( t test) intertextual connections the ACMC and the SCMC groups as measured by DLIs. The intertextuality of the writing discussions was stronger for the post treatment (first draft) writings of the ACMC group participants as compared to the writings of the participants in the SCMC group.

PAGE 651

625 ABOUT THE AUTHOR Iona Sarie St. Kliment Ochridsky and post graduate certificate for teaching English as a Foreign Language. After a number of years of language teaching, she continued her education in the Seco nd Language Acquisition and Instructional Technologies doctoral program at University of South Florida. Her research interests are in the field of second language writing and computer assisted language learning.