USF Libraries
USF Digital Collections

The effects of processing instruction, structured input, and visual input enhancement on the acquisition of the subjunct...

MISSING IMAGE

Material Information

Title:
The effects of processing instruction, structured input, and visual input enhancement on the acquisition of the subjunctive in adjectival clauses by intermediate-level distance learners of Spanish
Physical Description:
Book
Language:
English
Creator:
Russell, Victoria
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Complex L2 grammar
Distance language learning
Focus on form
Spanish as a foreign language
Web based instruction
Word animation
Dissertations, Academic -- Second Language Acquisition and Instructional Technology -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: This study investigated the effects of processing instruction (PI) on the acquisition of the subjunctive in adjectival clauses by 92 intermediate-level distance learners of Spanish. PI is a novel instructional technique that is based on VanPatten's principles of input processing (1993, 1996, 2002, 2004), and it has three key components: (a) an explicit explanation of grammar that is not paradigmatic, (b) information on processing strategies, and (c) structured input tasks and activities. Structured input activities were isolated and combined with computerized visual input enhancement (VIE) in an attempt to increase the salience of targeted grammatical forms for web based delivery. VIE was operationalized as word animation of subjunctive forms through flash programming language.An experiment comparing four experimental groups with traditional instruction indicates that for interpretation and production tasks, there were no significant differences between PI and traditional instruction. However, learners who received PI combined with VIE outperformed learners who received structured input activities without VIE for interpretation tasks. In addition, the present study examined the effects of PI when learners encountered targeted forms that were embedded in an authentic input passage that was received following the experimental exposure. Thus far, studies in the PI strand have only examined how learners interact with structured, or manipulated, input.The results of the present study indicate that participants who received PI in combination with VIE noticed targeted forms in subsequent authentic input with metalinguistic awareness, and they demonstrated a significantly higher level of awareness than participants who received traditional instruction or structured input activities. Further, learners who received PI, with or without VIE, were better processors of targeted forms that were embedded in subsequent authentic input than learners who received structured input activities without VIE.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2009.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Victoria Russell.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 524 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 - 002069309
oclc - 608202621
usfldc doi - E14-SFE0003189
usfldc handle - e14.3189
System ID:
SFS0027505:00001


This item is only available as the following downloads:


Full Text

PAGE 1

The Effects of Processing Instruction, Structured Input and Visual Input Enhancement on the Acquisition of the Subjunctive in Adjectival Clauses by Intermediate Level Distance Learners of Spanish b y Victoria Russell A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Secondary Education College of Education and Department of World Language Education College of Arts & Sciences University of South Florida Co Major Professor: Marcela van Olphen, Ph.D. Co Major Professor: Camilla Vsquez, Ph.D. Constance Hines, Ph.D. Glenn Smith, Ph.D. Date of Approval: November 2 2009 Keywords: complex L2 grammar, distance language learning, focus on form, Spanish as a foreign language, web based instruction, word animation Copyright 2009, Victoria Russell

PAGE 2

Dedication This work is dedicated to the three most important people in my life: my husband, Mark, and our two sons, Nicholas and Stephen.

PAGE 3

Acknowledgments I would like to thank a number people whose assistance made this dissertation study possible. First, I would like to thank my co major professors, Dr. Marcela van Olphen and Dr. Camilla V squez who provided constant support, encourageme nt, and speedy help whenever it was needed. It has been a privilege to work closely with both of these outstanding professors over the past year and a half. I could not have asked for better mentorship, advice, and inspiration than I received from them. I would also like to thank Dr. Constance Hines, who taught me everything that I know about statistics. Her guidance and expertise ensured that this study was rigorous, and for that I thank her very much. Dr. Glenn Smith also deserves my thanks for teach ing me the principles of instructional design and lending his expertise to this study. I feel so fortunate to have had these professors serve on my committee, and I am grateful that they shared their time and talents with me. I would like to give a speci al thanks to Beatrz Franco and Dr. Nancy Blain for allowing me access to their students and for modifying the syllabus in their courses to accommodate my research study. This study would not have been possible without their help and support. I am also g rateful to all of the Spanish II students who took the time and effort to participate in the study. In addition, I would like to thank all of the students and faculty who are part of the Second Language Acquisition and Instructional Technology (SLAIT) Ph. D. program for their encouragement and support. In particular, I would like to thank my colleagues

PAGE 4

Shengrong Cai, Angela Cresswell, Jane Harvey, Radhika Lothe, and Aylin Tekiner for sharing this journey and their friendship with me. I also thank Corina O wens from the Educational Measurement and Research Department and Maritza Chinea Thornberry, Carla Zayas Santiago, and Manuela Garcia Luque from the World Language Education Department for their assistance with this project. I would also like to thank Dr. Carrine Feyten and Dr. Jeffra Flaitz for having the vision to establish the SLAIT program. This program has brought together many diverse and talented people from all over the world, and it has been a privilege to be a part of it for the past four and ha lf years. Finally, I would like to thank my husband, Mark. I would not have been able to complete the Ph.D. program and the dissertation without his unselfish love and unconditional support. I would also like to thank my sons, Nicholas and Stephen. The y never complained once even though working on this dissertation took away time that I could have spent with them. I hope that they will reflect back on my example and know that with hard work and determination, they too can reach whatever goals they set for themselves in life.

PAGE 5

i Table of Contents List of Tables v List of Figures viii Abstract ix Chapter 1 Introduction 1 Background 1 Theoretical Framework 13 Statement of the Problem 18 Purpose 21 Research Questions 23 Research Hypotheses 24 Significance of the Study 26 Definition of Terms 27 Delimitations 31 Limit ations 32 Chapter 2 Review of Literature 33 Introduction 33 34 Processing Instruction 46 Empirical Studies on Processing Instruction 51 Authentic Versus Flawed Structured Input Activities 72 Processing Instruction and Meaningful Output Based Instruction 80 Processing Instruction and Type of Feedback 86 Summary of Findings and Implications for Future Research 88 Input Enhancement 93 Empirical Studies on Input Enhancement 94 Computer Based Visua l Input Enhancement 101 Summary of Findings and Implications for Future Research 104 Output and Language Learning 106 Empirical Studies Examining the Output Hypothesis 109 Summary of Findings and Implications for Future Research 116

PAGE 6

ii Distance Foreign Language Learners 117 The Spanish Subjunctive as a Complex Grammatical Feature 120 Chapter 3 Method 127 Introduction 127 Research Design 128 Population and Sample 131 Instruments and Measures 135 Pretreatment Questionnaire 135 Subjunctive Knowledge Test 136 Interpretation Subtest 137 Production Subtest 142 Comprehension Test 146 Validity and Reliability of the Tests 150 Evidence of Test Content 150 Response Process 151 Internal Structure 153 Reliability Evidence 155 Internal Consistency Reliability 157 Note Sheets 159 Authentic Input Text 160 Posttreatment Questionnaire 161 Variabl es 164 Instructional Materials 165 Multimedia User Interface Design 165 Alpha/Usability Testing of Instructional Materials 173 Beta and Pilot Testing of the Web Based Instructional Materials 183 Traditional Instruction 188 Processing Instruction 190 Structured Input 197 Visual Input Enhancement 197 Data Collection Procedures 199 Data Analysis 206 Chapter 4 Results 213 Introduction 213 Summary of the Pretreatment Ques tionnaire 213 Time on Task 217 Analyses of Pre and Posttests 218 Establishing Pretreatment Equivalence of Groups 219

PAGE 7

iii Analysis of Scores for Interpretation of the Subjunctive 219 Ana lysis of Scores for Production of the Subjunctive 228 Analysis of Scores for Interpretation of the Indicative 235 Analysis o f Scores for Production of the Indicative 241 Analysis of Note and Awareness Scores 247 Analysis of Text and Grammar Comprehe nsion Scores 254 Correlational Analyse s of Comprehension Test Scores 260 Summary of the Posttreatment Questionnaire 260 Ended Responses from the Posttreatment Questionnaire 265 Overall Summary of the Results of the Major Statistical Analyses 272 Repeated Measures Analyses 272 Multivariate Analyses 274 Correlational Analyses 275 Chapter 5 Discussion 276 Introduction 276 Discussion of Results 276 Discussion of Findings in Relation to the Research Questions 276 Discussio n of Findings in Relation to the Research Hypotheses 285 Discussion of Results for Interpretation and Production of the Indicative 291 Theoretical Implications 293 Processing Instruction and the Spanish Subjunctive 293 Processing Instruction and Input Enhancement 296 Processing Instruction and Authentic Input 299 Attention and Awareness in SLA 300 Pedagogical Implications 306 Limitations 309 Suggestions for Future Research 315 Conclusion 316 List of References 318 Appendix A Inform ed Consent Form 338 Appendix B Pretreatment Questionnaire 344 Appendix C Forms A, B, and C of the Subjunctive Knowledge Test 346 Appendix D Comprehension Test 377

PAGE 8

iv Appendix E Note Sheet 381 Appendix F Authentic Input Text 383 Appendix G Posttreatment Questionnaire 385 Appendix H Satisfaction Survey 390 Appendix I Vocabulary Practice Activity 392 Appendix J Traditional Instruction Explicit Grammar Explanation: Subjunctive Formation 396 Appendix K Traditional Instruction Explicit Grammar Explanation: Subjunctive Use 399 Appendix L Traditional Instruction Treatment Package 403 Appendix M Processing Instruction Explicit Grammar Explanation: Subjunctive Formation 421 Appendix N P rocessing Instruction Explicit Grammar Explanation: Subjunctive Use 423 Appendix O Information on Processing Strategies 427 Ap pendix P Processing Instruction Treatment Package 430 Appendix Q Structured Input Treatment Package 452 Appendix R Example of Computerized Visual Input Enhancement 473 App endix S Processing Instruction w ith Visual Input Enhancement Treatment Package 475 Appendix T Structured Input with Visual Input Enhancement Treatment Package 499 Appendix U Listening Scripts for the Instructional Treatments and the Subjunctive Knowledge Test: Forms A, B, & C 520 About the Author End Page

PAGE 9

v List of Tables Table 1.1 Ordering 3 Table 1.2 Type of Instruction by Group Examined in the Present Research Study 17 Table 2.1 res and Structural Outcomes 122 Table 3.1 Number of Student Participants by Class 133 Table 3.2 Number and Percent of Final Sample Assigned to Instructional Groups 134 Table 3.3 Means and Standard Deviations of Scores on Three Forms of the Interpretation and Production Subtests 157 Table 3.4 Means and Standard Deviations of Scores on the Comprehension Test Across Two Times of Testing 158 Table 3.5 Frequency Count of Subjunctive Ve rbs in the Authentic Input Text 162 Table 3.6 Learnability Assessment for Each Task by Instructional Treatment Package 175 Table 3.7 Error Score for Each Task by Instructional Treatment Package 177 Table 3.8 Results of the Satisfaction Survey by Instructional Treatment Package 180 Table 3.9 Time Needed to Complete the Instructional Treatment Packages 186 Table 3.10 Order of Test Delivery by Class 202

PAGE 10

vi Table 4.1 Participant Background Information from the Pretreatment Questionnaire 214 Table 4.2 Time Spent on Task by Instructional Group 218 Table 4.3 Descriptive Statistics for Scores on the Interpretation Subtest by Group at Pretest, Posttest 1, and Posttest 2 for Interpretation of the Subjunctive 221 Table 4.4 Analysis of Variance of Interpretation Test Scores by Instructional Treatment Group and Time for Interpretation of the Subjunctive 225 Table 4.5 De scriptive Statistics for Scores on the Production Subtest by Group at Pretest, Posttest 1, and Posttest 2 for Production of the Subjunctive 229 Table 4.6 Analysis of Variance of Productio n Test Scores by Instructional Treatment Group and Time for Production of the Subjunctive 233 Table 4.7 Descriptive Statistics for Scores on the Interpretation Subtest by Group at Pretes t, Posttest 1, and Posttest 2 for Interpretation of the Indicative 236 Table 4.8 Analysis of Variance of Interpretation Test Scores by Instructional Treatment Group and Time for Interpreta tion of the Indicative 2 39 Table 4.9 Descriptive Statistics for Scores on the Production Subtest by Group at Pretest, Posttest 1, and Posttest 2 for Production of the Indicative 242 Table 4.10 Analysis of Variance of Production Test Scores by Instructional Treatment Group and Time for Production of the Indicative 246 Table 4.11 Descriptive Statistics on Note and Awareness Scores by Group 248 Table 4.12 ANOVA Summary Table for Awareness Scores by Instructional Treatment Group 253

PAGE 11

vii Table 4.13 Descriptive Statistics on Text and Grammar Comprehension Scores by Group 254 Table 4.14 ANOVA Summary Table for Grammar Comp rehension Scores by Instructional Treatment Group 259 Table 4.15 Participant Responses from the Posttreatment Questionnaire 266

PAGE 12

viii List of Figures Figure 2.1. A Sketch of Basic Processes in Acquisition (from VanPatten, 1993) 36 Figure 2.2a. Processing Instruction in Foreign La nguage Teaching (from VanPatten & Cadierno, 1993b) 47 Figure 2.2b. Traditional Explicit Grammar Instruction in Foreign Language Teaching (from VanPatten & Cadierno, 1993b) 47 Figure 3.1. Research Design 130 Figure 3.2. Constructs Measured on the Subjunctive Knowledge Test 137 Figure 3.3. Major Types of Structured Input Activities (from J.F. Lee & VanPatten, 2003) 195 Figure 3.4. Experimental Schedule 207 Figure 4.1. Interaction Plot for Instruction Type and Time for Interpretation of the Subjunctive 226 Figure 4.2. Graph of the Main Effect for Time for the Mean Interpretation Test Score at Pretest, Posttest 1, and Posttest 2 for Interpretation of the Subjunctive 227 Figure 4.3. Graph of the Main Effect for Time for the Mean Producti on Test Score at Pretest, Posttest 1, and Posttest 2 for Production of the Subjunctive 234 Figure 4.4. Graph of the Main Effect for Time for the Mean Interpretation Test Score at Pretest Posttest 1, and Posttest 2 for Interpretation of the Indicative 240 Figure 4.5. Graph of the Main Effect for Time for the Mean Production Test Score at Pretest, Posttest 1, and Posttest 2 for Production of the Indicative 247

PAGE 13

ix The Effects of Processing Instruction, Structured Input, and Visual Input Enhancement on the Acquisition of the Subjunctive in Adjectival Clauses by Intermediate Level Distanc e Learners of Spanish Victoria Russell ABSTRACT This study investigated the effects of processing instruction (PI) on the acquisition of the subjunctive in adjectival clauses by 92 intermediate level distance learners of Spanish. PI is a novel instruct input processing (1993, 1996, 2002, 2004), and it has three key components: (a) an explicit explanation of grammar that is not paradigmatic, (b) information o n processing strategies and (c) struct ured input tasks and activities. Structured input activities were isolated and combined with computerized visual input enhancement (VIE) in an attempt to increase the salience of targeted grammatical forms for web based delivery. VIE was operationalize d as word animation of subjunctive forms through flash programming language. An experiment comparing four experimental groups with traditional instruction indicates that for interpretation and production tasks, there were no significant differences betwee n PI and traditional instruction. However, learners who received PI combined with VIE outperformed learners who received structured input activities without VIE for interpretation tasks.

PAGE 14

x In addition, the present study examined the effects of PI when le arners encountered targeted forms that were embedded in an authentic input passage that was received following the experimental exposure. Thus far, studies in the PI strand have only examined how learners interact with structured, or manipulated, input. The results of the present study indicate that participants who received PI in combination with VIE noticed targeted forms in subsequent authentic input with metalinguistic awareness, and they demonstrated a significantly higher level of awareness than par ticipants who received traditional instruction or structured input activities. Further, learners who received PI, with or without VIE, were better processors of targeted forms that were embedded in subsequent authentic input than learners who received str uctured input activities without VIE.

PAGE 15

1 Chapter 1 Introduction Background Traditionally, foreign languages (FL) in secondary schools, colleges, and universities across the United States use textbooks and materials that present learners with explicit grammar explanations followed by practice activities where students are required to produce target language (TL) output from the moment that they first enter class. When FL students first encounter a new language, in effect all linguistic input in the second language (L2) is new to them. Novice language learners typically struggle to extract meaning from their L2 input (Frch & Kasper, 1986; Krashen, 1982), and a large part of their attentional resources are consumed during the comprehension pro cess (Just & Carpenter, 1992). Traditional FL textbooks place a heavy emphasis on grammar instruction and output based practice, which requires beginning level learners to focus on the formal features of language. When novice FL learners are required to focus on grammatical forms and structures, they may not have enough attentional resources to attend to both meaning and grammatical form simultaneously. Therefore, while FL learners attempt to comprehend grammar, they often miss the intended message of t heir TL input.

PAGE 16

2 With traditional instruction, teachers and textbook authors commonly use mechanical drill activities to encourage students to focus on a t argeted grammatical form. During mechanical drills, the teacher/textbook author has complete control over the response, and there is only one possible correct answer. According to Paulston (1972), the goal of the mechanical drill is to give students practice with TL structure in order to assist them in moving from repetition to self expression without m aking gra mmatical errors 1. A drawback of mechanical drills is that students do not have to understand the stimulus to produce a correct answer. Therefore, it is not uncommon f or students to fail to understand the meaning of even their own TL utterances when they are engaged in mechanical drills. Some scholars (Krashen, 1980, 1981, 1982; VanPatten, 2003) claim that language learners need adequate time to process linguistic in put before they are required to produce output in a second language (L2). According to VanPatten (2003), linguistic t can be interactional, as in the input learners hear during their communicative exchanges, or noninteractional, as in input that is either not specifically directed to the individual learner or where the learner is not part of the communicative exchange ( R. Ellis, 1994). In order for input to be available for acquisition, it must be made comprehensible to the learner because incomprehensible language is not useful for SLA (Krashen, 1980, 1981, 1982). Further, VanPatten claims

PAGE 17

3 Table 1.1 my of Practice Types and Their Sequential Ordering Sequencing Characteristics mechanical 1. Learner does not need to attach meaning to sentences in order to complete the practice. 2. There is one and only right correct response Ex: transformation drill meaningful 1. Learner needs to attach meaning to both stimulus and response. 2. There is one and only right correct response; the intended meaning of the learner is already known by the instructor (or fellow learner). Ex: answering questions such as, "What time does class begin? communicative 1. Learner needs to attach meaning to both stimulus and response. 2. Intended meaning of the learner is not known by the instructor (or fellow learner). Ex: answering questions such as, "Do you have posters in your Note From Paulston (1972). Structural pattern drills: A classification In H. B. Allen & R. N. Campbell (Eds.), Teaching Engli sh as a second language (pp. 129 138). New York: McGraw Hill.

PAGE 18

4 that input is the single most important factor for Second Language Acquisition (SLA), with all theories of SLA relying on input, in some way, to explain language acquisition. of input processing (1993, 1996, 2002, 2004) proposes a set of principles that describe the strategies that learners use to process L2 input. The principles of this model also serve as the foundation for processing instruction a novel instructional techn ique that is informe d by SLA research According to VanPatten (1993, 1996, 2002, 2004) there are two subprocesses involved with input processing: (a ) making form meaning connections, and (b ) parsing. Form meaning connections refer to the cognitive mappi ng that learners make between a formal feature of language and its referential meaning. For example, with the Spanish verb ir which is rendered will go in English, the ending on the verb ir encodes a future tense meaning. VanPatten (1996) asserts th at L2 learners must notice a formal feature of the language and the referential meaning that the form encodes for SLA to take place. In the previous example, L2 learners of Spanish would need to notice the ending attached to the infinitive ir and be ab le to connect the ending with the future tense meaning. This process is known as making a form meaning connection. The future tense inflectional morpheme in Spanish has a high communicative value, meaning that its presence contributes to the overall r eferential meaning of the sentence ( J.F. Lee, 2002). VanPatten (1996) claims that L2 learners are more likely to make the necessary form meaning connections when a linguistic form has a high communicative value. VanPatten (1996) also asserts that when l earners make form meaning connections, the related input becomes intake for learning and has the potential to

PAGE 19

5 become internalized into the developing interlanguage system. Intake is defined as the subset input data that is available to learners for furthe r language processing (Gass, 1988, linguistic information must first be incorporated into the developing IL system before any output processing is possible. (VanPatten, 2003, p 118). Parsing is an important element in input processing because what elements learners expect to encounter in their linguistic input influences comprehension. Due to the principles that guide parsing, L2 learners of Spanish whose first language is Engl ish will typically assume that the first noun that they hear or read in a sentence is the subject of the sentence. VanPatten (1993, 1996, 2002, 2004) refers to this as the first noun principle. Due to this faulty assumption, it is expected that L2 learner s of Spanish whose L1 is English will have delayed acquisition of pronouns, case markings, and passives (VanPatten, 2003). In order to overcome delays in acquisition, VanPatten developed processing instruction (PI), which draws upon the principles that guide input processing. The main objective of PI is to provide ample opportunities for L2 learners to make form meaning connections through tasks and activities that supply them with structured input, in particular input that elevates the communicative va lue of specific linguistic forms. PI has three key elements: (a) explicit information regarding how a grammatical form or

PAGE 20

6 structure works, (b) information on processing strategies and how to avoid faulty input processing, and (c) structured input tasks an d activities. Structured input activities are their processing strategies in order to facilitate the conversion of input to intake. Researchers such as Collentine (1 998 a ), and Hwu (2004) suggest that PI is a perfect fit for computer assisted language learning (CALL) activities because teachers can easily manipulate the linguistic input that they provide to learners. Structured input is an input enhancement technique of a linguistic item relative to its position in the surrounding sentence, and it is thought to increase the likelihood that input will be converted to intake for learning (VanPatten, 1993, 1995, 1996) Hwu (2004) claims that computer based instruction is superior to teacher delivered instruction for input based activities, such as structured input tasks, computer tec which makes them an ideal fit for web based instruction (WBI). Research that has compared PI to traditional instruction has found that PI is superior to traditional instructio n when learners are engage d in interpretation activities and that PI is equal to traditional instruction when learners are engaged in production activities (Benati, 2001, 2005; Cheng, 2002; Cadierno, 1995; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wo ng, 2004). These results are somewhat surprising because the L2 learners who received PI in the aforementioned studies did not engage in any output activities during the instructional treatments, yet their production in the TL

PAGE 21

7 was equivalent to that of th e L2 learners who produced output during instructional treatments. Further, it would appear that PI is superior to traditional instruction because participants who received PI demonstrated superior interpretation of targeted grammatical forms compared to their counterparts who received traditional instruction. In addition, research by VanPatten and Fernndez (2004) indicates that the beneficial effects of PI are durative, with participants still demonstrating learning gains eight months after their instru ctional treatments. A drawback of PI is that not all TL forms are amenable to structured input activities, a key component of PI. In order to create structured input activities, all contextual cues that are redundant to the targeted grammatical form mu st be removed in Mary walked to the bank yesterday the word yesterday would be removed from the input so that the learners must derive the past tense meaning from the bound inf lectional morpheme ed rather than from the lexical item yesterday In addition, forms that are devoid of meaning, such as definite articles in Spanish, are not suitable for PI because they only carry grammatical information and no semantic meaning. In o rder to benefit from processing instruction, a grammatical form or structure must convey some type of semantic meaning. Another instructional technique that attempts to help learners focus on form is input enhancement, which is a pedagogical attempt to p romote SLA by increasing L2 input. Input enhancement is an input based technique that does not attempt to alter

PAGE 22

8 input enhancement is to make certain aspects of the L2 input more salient for learners, whether the enhancement technique is explicit and elaborate, such as the provision of metalinguistic rule explanations, or implicit and simple, such as using colored ma rkers or chalk to highlight targeted forms on the board (Sharwood Smith, 1991). Sharwood Smith (1981) claims that grammar instruction does not have to take the traditional form of metalinguistic discussions; rather, teachers can help their students pay at tention to grammar through a variety of input enhancement techniques. Rutherford and Sharwood Smith (1985) proposed a number of input enhancement techniques, including: input flood, typographical enhancement, and grammatical consciousness raising. Curren tly, the most common forms of input enhancement techniques are structured input, input flood, and textual or visual input enhancement (Wong, 2005) Structured input, a key element in processing instruction, is a technique that elevates the communicative v al ue of linguistic forms by eliminating any lexical redun dancies in the input and by simplifying the input sur rounding the targeted structure and make a form meaning connecti on. Input flood is an input enhancement technique where the teacher or researcher manipulates the input in order to saturate it with the targeted linguistic form. Theoretically, L2 learners are more likely to notice the targeted form due to its increased frequency in their i nput (Gass, 1997, Wong, 2005). Grammatical consciousness raising is an input enhancement technique that utilizes inductive grammar activities. L2 learners are encouraged to discover grammatical rules on their own by interacting with the input while performing some task ( R. Ellis, 1997).

PAGE 23

9 This instructional method operates on the premise that when students understand how certain features of grammar work in the L2, then they will be more likely to notice those features in subsequent inp ut ( R. Ellis, 1997; Fotos, 2002). Visual Input enhancement (VIE) is one of the simplest forms of input enhancement. It is used to make certain features of written L2 input more salient for L2 learners through formatting techniques such as bolding, capita lizing, highlighting, and/or a change in font style or size. Of all the previously mentioned input enhancement techniques, VIE is perhaps the most controversial in SLA research. The results of res earch studies on whether VIE is facilitative for SLA have been largely mixed, as some studies have demonstrated a positive effect for VIE (Doughty, 1988, 1991 ; Jourdenais, Ota, Stauffer, Boyson, & Doughty, 1995; Shook, 1994; Williams, 199 9 ; Wong 2002), some have found only a minimal effect (Alanen, 1995; Izumi, 2002; Robinson, 1997; J. White, 1998), and still others have demonstrated no beneficial effect for VIE (Leow, 1997 2001; Leow, Nuevo & Tsai 2003; Jordenais, 1998; Overstreet, 1998 ; Wong, 2003) Of note, Sharwood Smith (1981, 1991) posits that input e nhancement techniques are designed to help L2 learners pay attention to the formal features of language. However, he cautions form learning may not occur in the presence of input enhancement because even if L2 learners do consciously attend to a linguisti c form due to the presence of input enhancement, there is no guarantee that intake into the developing IL system will occur, as learners may make incorrect form meaning connections. While there is no guarantee of a beneficial instructional outcome through the use of input enhancement

PAGE 24

10 techniques, there is an increased likelihood that learners will notice the enhanced form, which may or may not lead to further language processing. Both PI and VIE are instructional techniques that are designed to focus learn attention on the formal features of their input in order to facilitate language acquisition. Although these instructional methods are both input based and emphasize comprehension over production strategies, PI is not a comprehension based teaching me thod. Teaching methods such as the natural approach a nd total physical response which are comprehension based teaching methods, do not take into account input processing al prompting learners to make correct form meaning mappings According to VanPatten (1996), language re provided with repeated examples of correct form meaning mappings that result when learners process their input correctly. A key difference between PI and VIE is that the goal of input enhancement is to help L2 learners notice certain features of their linguistic input by making the targeted features more salient for them. Once key elements are noticed, however, input enhancement does not provide a way to help learners understand the meaning of the noticed input, and it is possible that learners may mak e incorrect form meaning mappings in the presence of input enhancement techniques. In contrast to the aforementioned input based instructional methods, traditional to be a key factor in developing fluency and accuracy in the L2 (Swain, 1985). Traditional

PAGE 25

11 instruction typically presents grammar points via an explicit explanation of grammar followed by output practice. Paulston (1972) describes traditional FL instruct ion as presentation/explanation of a targeted grammatical form followed by mechanical, then meaningful, then communicative activities. VanPatten (2004) asserts that the model language classrooms in the U.S. and is the model followed by almost every major language textbook published for the secondary and post Research by Swain (1985, 1993, 1995, 1998) supports the role of output in the FL classroom. Swain (1985) developed the Output Hypothesis based on her observations of long term French immersion students in bilingual education programs in Canada. She concluded that comprehensible input, although necessary for SLA to take place, was not sufficient for learners to fully develop native like proficiency in a second language. Swain (1985) found that long term French immersion students who received large amounts of comprehensible input in the L2 developed high levels of comprehension and native like acc ents, but failed to attain native like production, especially in the area of grammatical accuracy. She observed that long term immersion learners were not prompted or pushed by their teachers to produce linguistic output in the L2 during subject matter in struction. Immersion learners tended to only produce a very small amount of language in the L2, which Swain described as the minimum amount of language that was Further, the immersion classes that she observed were mainly teacher fronted, and teachers failed to correct grammatical errors if students were able to adequately convey

PAGE 26

12 meaning in the L2. Swain (1985) claims that L2 learners need to be pushed to produce output in the target language because the provision of abun dant amounts of linguistic input (as in the case in bilingual immersion education) is not enough to develop native like grammatical accuracy. She posits that when L2 learners produce output, they shift from semantic to syntactic processing, which is a dee per level of language processing. Swain asserts that when students are not pushed to produce output in the L2, they will only attend to linguistic meaning at the expense of grammatical form. While the Output Hypothesis is not a theoretical unpinning of t raditional instruction, it lends weight to the importance that traditional instruction places on production practice (DeKeyser & Sokalski, 1996). Swain (1993, 1995, 1998) extended the Output Hypothesis when she identified three functions that output serve s in SLA: ( a ) the hypothesis testing function, ( b ) the metalinguistic function, and ( c ) the noticing / triggering function. The hypothesis testing accuracy in the TL. T hrough the first two functions, learners become aware of and test out their theories regarding the TL rules and structures. The noticing function of output Schmidt & Frota, 1 986), which states that L2 learners must first notice target language forms in order for input to be converted into intake for learning. Schmidt claims that correct target language form for SLA to take place (Schmidt & Frota, 1986). According to Swain and Lapkin (1995), when learners attempt to produce output in the L2, they may

PAGE 27

13 not know or remember the necessary linguistic forms and structures that they need to communicat their IL knowledge. Thus, by attempting to produce output, learners are forced into noticing what they do not know, or what they know only partially. Swain (1995) posits that no ticing holes or gaps in their IL knowledge primes learners to pay more careful attention to the relevant forms in their future input. A criticism of the Output Hypothesis is that there has been little empirical research to support it (Swain, 1998; Swain & Lapkin, 1995). Further, only a handful of studies have investigated the noticing function of output by providing participants with relevant input following output based activities (Izumi, 2002; Izumi & Bigelow 2000; Izumi, Bigelow, Fujiwara & Fearnow, 19 99). Moreover, the results of only one published study (Izumi, 2002) partially support the noticing function of the output hypothesis. Theoretical Framework input, with the re cognition that not all of the linguistic input that learners are exposed to becomes intake for learning (Gass, 1988). Intake is defined as the subset input data that is available to learners for further language processing (Gass, 1988, VanPatten, 1996). Schmidt (1990, 1995, 2001; Schmidt & Frota, 1986) claims that noticing is what mediates input and intake, and that noticing is a necessary condition for second language acquisition. In other words, Schmidt claims that conscious rather than subliminal proc esses drive SLA. Schmidt (2001) also asserts that noticing requires focal attention

PAGE 28

14 and awareness on the part of the learner, and learners must notice their TL input with understanding for language acquisition to take place. One of the primary goals of b oth that they would not otherwise notice. Thus, the Noticing Hypothesis serves as a 1991 1993 m odel of i nput p rocessing (1993, 1996, 2002, 2004) It is important to note that according to VanPatten (2004), input processing assumes that learners have perceived and noticed the targeted grammatical forms; however, notici ng alone does not signify that learners have processed the forms in their working memories. VanPatten (2004) posits that intake does not occur until learners make form meaning ories m odel of i nput p rocessing m odel of i nput p rocessing presume that intern al or mental processes are responsible for SLA and that new knowledge must be integrated organized cognitive systems or networks to accommodate new knowledge, restruc turing is said to take place. The integration of new knowledge and restructuring are key components of a cognitive approach to SLA, which views all language learning as a mental construct. Thus, both input enhancement and processing instruction are instr uctional techniques that are based upon a cognitivist framework for SLA.

PAGE 29

15 The major theoretical underpinnings of traditional instruction are behaviorism and skill acquisition theory (Anderson, 1976). The mechanical drill activities that are at the core of tradition al instruction are vestiges of the audiolinguial teaching method, which dominated FL instruction in the United States in the 1950s and 1960s. The principles of behaviorism underlie the basic tenets of the audiolingual ism (Omaggio Hadley, 2001). that depended on imitation, practice, and reinforcement. Skinner described language as a sophisticated stimulus response system, and the goal of instruction was to es tablish and strengthen stimulus response connections. In essence, the principles of behaviorism propose that language is best learned through extensive drills and practice In addition, skill acquisition theory posits that all knowledge begins in a declarative form and is converted to procedural knowledge through practice. Thus, the role of output practice in traditional instruction is paramount. Although the audiolingual method fell into disfavor in the 1980s, many FL teachers and most textb ook authors still rely heavily on the output based mechanical drill activities that stemmed from this methodology. These activities emphasize the teaching of structural patterns through the use of repetitive drill activities. According to Chastain (1976) every audiolingual textbook included pattern drills, of which there were two main types: (a) (b) model. The latter type of drill was subsequently reinforced by the teacher or by an audio recording.

PAGE 30

16 While all of the instructional techniques that were examined in this study (PI, structured input, visual input enhancement, and traditional instruction ) attempt to focus theoretical underpinnings. Both input enhancement and PI fall within a cognitive framework for SLA and rely heavily on the Noticing Hypothesis; ho wever, PI is also m odel of i nput p rocessing (1993, 1996, 2002, 2004), as the main other two methods, traditional instruction relies heavily on skill acquisition theory and behaviorism. While skill acquisition theory and behaviorism are still cognitive approaches to SLA, behaviorist theory emphasizes the importance of repetition and rote practice for learning foreign languages, and skill acquisit ion theory stresses the importance of practice in order to convert declarative knowledge to procedural knowledge. With traditional instruction, the primary way to engage students in rote practice of novel forms and structures is through the use of mechani cal and pattern drill activities, both of which require an extensive amount of TL output in the oral and written modalities. Table 1.2 provides a visual display of the various instructional techniques that were examined in the present study.

PAGE 31

17 Table 1 .2 Type of Instruction by Group Examined in the Present Research Study Instructional method Components Approach Type Processing instruction (+PI VIE) 1. Explicit grammar explanation 2. Information on processing strategies 3. Structur ed input activities Deductive Input based Processing instruction with visual input enhancement (+PI+VIE) 1. Explicit grammar explanation 2. Information on processing strategies 3. Structured input activities with visual input enhancement Deductive Input b ased Structured input (+SI VIE) Structured input activities Inductive Input based Structured input with visual input enhancement (+SI+VIE) Structured input activities with visual input enhancement Inductive Input based Traditional in struction (comparison group) 1. Explicit grammar explanation 2. Mechanical output based activities 3. Meaningful output based activities 4. Communicative (open ended) output based activities Deductive Output based

PAGE 32

18 Statement of the Problem Today, more than 3.9 million college students take at least one class online, and both web facilitated and blended/hybrid courses are growing in popularity ( E. Allen & Seaman, 2008). While traditional face to face (FTF) instruction does not rely on web based technology to deliver instruction, web facilitated courses are becoming more popular due to the ease of using courseware management systems to post syllabi and assignments. A new type of course is the hybrid, where online instruction is blended with FTF delivery. In hybrid courses 3 0% to 79% of instruction is delivered online ( E. Allen & Seaman, 2008). With the growth of WBI, it is important to look for new instructional techniques that are suitable for online delivery. Currently, many textbook companies provide web based versions of their textbooks for online or distance e ducation classes. According to Fraser (1999) the developers of most of these products offer little more than an identical version of the printed textbook that is adapt ed for online use This com mon practice of indiscriminately converting information from one format into another without regard for ease of use, appearance, or capabilities is known as without cons idering how the learning environment may be different or opti mized in the new format M. Allen (2003) recommends that time and resources should be invested into the process of planning, analysis, and design before developing e learning materials. Since w eb facilitated and web delivered courses are growing in popularity, it is important to examine the body of research on computer assisted language learning (CALL), in particular research that seeks to determine which instructional methods and techniques

PAGE 33

19 yie learning materials onto the web. An overview of the literature on PI and VIE suggest that these techniques might be beneficial for web based foreign language instruction; howev er, there is scant research to support this claim. More research is needed on the components of PI, in particular on structured input activities, which may be a good fit for WBI because teachers have to FTF classrooms where learners receive linguistic input from their peers as well as their teacher. Furthermore, when studying a FL online, the input that is provided to the learner is a key feature of instruction because online learning typically provid es fewer opportunities for students to produce verbal output and to interact with one another and/or their teacher than traditional FTF instruction. Thus, the role of linguistic input takes on even greater significance for distance FL learning. VanPatten and Oikkenon (1996) examined the components of PI to determine which part (explicit grammar explanation, structured input, or both) is the most beneficial object pron ouns and word order in Spanish, and they found that structured input activities alone are as effective as PI. Interestingly, VanPatten and Oikkenon did not find the same effect for explicit grammar explanation alone. Benati (2004 b 2005) replicated VanPa tten and Oikkenon with a different TL and a different grammatical form and obtained the same results. Doughty (2004) posits that the explicit instruction component of PI only leads to the learning of metalinguistic knowledge, or knowledge about the

PAGE 34

20 langua for the explicit explanation component of PI is to orient the learners to the processing problem. Since structured input activities (the second component of PI) also or ient learners to the processing problem, Doughty suggests that the explicit explanation component is not necessary. Farley (2004 b ), however, found that explicit instruction plays a key role in PI when the targeted grammatical form is complex. He replicat ed VanPatten and Oikkenon (1996) with the present subjunctive in Spanish in noun clauses following expressions of doubt or denial, a much less transparent form than object pronouns. Farley found that participants that had an explicit grammar explanation w ith structured input activities performed significantly better on posttests than participants who only received structured input activities. Thus, the findings are mixed on the effectiveness of the explicit grammar explanation component of PI, and it is u nclear if learners require an explicit explanation when the targeted grammatical form is complex. Further, it is presently unknown whether other types of input enhancement in combination with structured input activities, as suggested by Collentine ( 1998 a ) Doughty (2004), and Hwu (2004), are as beneficial for L2 learners as PI. In addition, novel forms of input enhancement need to be developed and investigated in order to determine if they ay be that traditional forms of textual and visual input enhancement such as bolding and underling text no longer M. Allen (2003) asserts that learners filter out stimuli that are perceived as uninteresting in computer based media. Visual enhancement in computer and web based environments can take on

PAGE 35

21 many different forms. Text can be enhanced acoustically, with color, with graphics, and/or animations. Thus far, no study has investigated t he combined effects of VIE and PI in a web based learning environment Finally, all of the research to date in the PI strand has failed to investigate whether PI is able to affect how learners process authentic input subsequent to receiving their instruct ional treatments. Collentine ( 2004) forms constituting the targeted grammatical phenomenon in normal input conditions (i.e., ). Theoretically, L2 learners will notice targeted forms in subsequent authentic input, avoid faulty processing strategies, and make correct form meaning mappings following PI; however, the current body of research has failed to demonstrate this facilitat ive effect as learners engage with authentic, rather than structured, input after the experimental exposure. Purpose Motivated by previous research on processing instruction (PI) and visual input enhancement (VIE) and the existing gaps in the literature in these areas, the overarching purpose of this study was to investigate novel instructional techniques (processing instruction structured input, and visual input enhancement) for teaching complex grammar to distance learners of Spanish and to compare th ese methods to traditional instruction, the dominant instructional paradigm in both FTF and web based formats in the United States today.

PAGE 36

22 In summary, a review of the relevant research in the field predicts that PI and structured input activities will be su perior to traditional instruction for interpretation tasks, but that PI and structured input activities will be relatively equal to traditional instruction for production tasks (Benati, 2001, 2005; Cadierno, 1995; Cheng, 1995; Farley, 2001a; VanPatten & Ca dierno, 1993a, 1993b; VanPatten & Wong, 2004). The present study combined structured input activitie s with VIE to determine if additional enhancements increase the effectiveness of these techniques in a web based learning environment. Theoretically, the presence of two types of input enhancement (structured input with VIE ) should be superior to only one type (structured input alone) for research is mixed regarding the beneficial effect of structured input alone on the acquisition of complex grammatical features, the present study also isolated structured input activities from PI to determine whether the explicit explanation component of PI is necessary when the tar geted grammatical form is complex. subjunctive verb forms that were embedded in an authentic TL text that participa nts received post experimental exposure This portion of the study attempted to determine if exposure to a particular instructional technique had an effect on how learners noticed, processed, and comprehended subsequent L2 input that contained the targeted grammatical forms. In addition, the relationship betwee n comprehension and input processing was examined to determine if there was a relationship between these two

PAGE 37

23 constructs. Although J.F. Lee and VanPatten (1995) assert that these are two distinct processes (making meaning versus making form meaning connec tions), they are likely to overlap. Research Questions Five instructional treatments were examined in the present study as follows: processing instruction without visual input enhancement (+PI VIE), processing instruction with visual input enhancement (+PI +VIE), structured input with visual input enhancement (+SI +VIE), structured input without visual input enhancement (+SI VIE), and traditional output based instruction (TI). The following research questions were addressed within the context of this study: 1. Is there a differential performance between treatment groups for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks over time ? 2. Is there a differential performance between treatment grou ps for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks over time? 3. ability to notice targeted forms in subsequent aut he ntic input as measured by note scores and awareness scores? 4. Following the instructional treatments, is there a differential performance between the targeted grammatical form (input processing) and the message of the authentic

PAGE 38

24 input text in which it is embedded as measured by grammar comprehension and text comprehension scores? 5. What is the relationship between text comprehension and input processing when learners encount er the targeted grammatical form in subsequent authentic input? Research Hypotheses Hypothesis 1 is that l earners who are exposed to processing instruction and structured input activities will outperform learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks. ( +PI VIE and +SI VIE > TI ) Hypothesis 2: L earners who are exposed to processing instruction and structured input activities will p erform as well as learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks. ( +PI VIE and +SI VIE = TI ) Hypothesis 3: L earners who are exposed to processing instruction and structured input activities with visual input enhancement will outperform learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by inter pretation tasks. ( +PI +VIE and +SI + VIE > TI ) Hypothesis 4: L earners who are exposed to processing instruction and structured input activities with visual input enhancement will perform as well as learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks. ( +SI +VIE and +SI +VIE = TI )

PAGE 39

25 Hypothesis 5: L earners who are exposed to processing instruction and structured input activities with visual inpu t enhancement will outperform learners who are exposed to processing instruction and structured input activities without visual input enhancement for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks. ( +PI +VIE and +SI +VIE > +PI VIE and +SI VIE ) Hypothesis 6: L earners who are exposed to structured input activities alone will not perform as well as learners who are exposed to processing instruction for the acquisition of the present subjunc tive in adjectival clauses in Spanish as measured by interpretation tasks. ( +SI VIE < +PI VIE) Hypothesis 7: L earners who are exposed to structured input activities combined with visual input enhancement will perform as well as learners who are expose d to processing instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks. ( +SI +VIE = +PI VIE ) Hypothesis 8a: L earners who are exposed to processing instruction and structured i nput activities with visual input enhancement will notice more targeted verb forms that are embedded in a subsequent authentic input passage than learners who are exposed to traditional instruction and learners who are exposed to processing instruction and structured input activities without visual input enhancement. ( +PI +VIE, +SI +VIE > TI, +PI V IE and +SI V IE) Hypothesis 8b: L earners who are exposed to processing instruction and structured input activities with visual input enhancement will have a h igher level of awareness (or a deeper level of noticing) of the targeted verb forms that are embedded in

PAGE 40

26 a subsequent authentic input passage than learners who are exposed to traditional instruction and learners who are exposed to processing instruction an d structured input activities without visual input enhancement. ( +PI +VIE, +SI +VIE > TI, +PI VIE and +SI VIE ) Hypothesis 9a: L earners who are exposed to processing instruction and structured input activities with and without visual input enhancemen t will perform as well as learners who are exposed to traditi onal instruction for comprehending the message of a subsequent authentic input text in which the targeted grammatical form is embedded as measured by text comprehension scores. ( +PI +VIE, +PI VI +SI +VIE, +SI VIE = TI ) Hypothesis 9b: L earners who are exposed to processing instruction and structured input activities with and without visual input enhancement will outperform learners who are exposed to traditional instruction for processing ta rgeted forms that are embedded in a subsequent authentic input text as measured by g rammar comprehension scores ( +PI +VIE, +PI VI, +SI +VIE, +SI VIE > TI ) Hypothesis 10: T here will be a significant positive correlation between input processing and tex t comprehension. Significance of the Study This study is significant because it contributed to the present body of knowledge on PI structured input, and VIE in the field of SLA. More specifically, the present study was the first to investigate the effe cts of PI with distance language learners, which is of particular importance since more than 3.9 million undergraduates in the United States

PAGE 41

27 currently take at least one course online, and the numbers of online learners continues to grow ( E. Allen & Seaman, 2008). In addition, all previous studies that examined VIE employed simple typographical e nhancements or enhancements through the use of highlighting and/or color. The present study was the first of its kind to operationalize VIE with word animation t hrough the use of flash programming language, which took advantage of the capabilities of the web based learning environment. Further the present study is significant because it examined whether PI and VIE rocessing of targeted forms in subsequent authentic input post experimental exposure. Past studies in the PI strand only examined input. Finally the findings of the present st udy have the potential to improve FL pedagogy. The results will assist FL practitioners in determining which instructional techniques are the most beneficial for t eaching complex grammar in web based and blended learning environment s Definition of Terms Awareness: A particular state of mind in which an individual has undergone a specific objective experience of some cognitive content or external stimulus (Tomlin & Villa, 1994). Awareness can occur at the level of noticing, which indicates meta awarenes level of awareness, or depth of noticing, was measured by a Posttreatment Q uestionna i re

PAGE 42

28 (an off line measure). The Posttreatment Q uestionnaire was a retrospective measure of input, and it required participants to provide metalinguistic information about the use of the sub junctive in adjectival clauses. Participants demonstrated meta awareness, or awareness at the level of noticing, if they were able to articulate that the subjunctive mood was present in the authentic input text. They were also asked to give a TL example of the grammatical form as proof of their meta awareness. Participants demonstrated awareness at the level of understanding if they were able to state the morphological rule for using the subjunctive in adjectival clauses. Comprehensio n: A Comprehension T est that was created for this study measured two types of comprehension: (a) of the input passage, and (b) comprehension of the referential meaning of the targeted verb forms. Comprehension of the propositional content of the input passage was measured by multiple choice test items that queried information from different levels of the passage including the main idea and specific details (Wolf, 1993). Comprehension of the referential meaning of the targeted verb forms was measured by multiple choice and short answer questions. The multiple choice q uestions were intended to measure whether participants were able to identify the grammatical form of the conjugated verb (present subjunctive or present indicative). The short answer ques tions were intended to measure whether participants were able to comprehend the referential meaning of the subjunctive forms.

PAGE 43

29 Detection: ar and specific bit of Form Meaning Connection: The connection between a grammatical form and the referential meaning that it encodes (VanPatten, 1993, 1996, 2002, 2004). Input: The linguistic data that i s available to the learner. Input Processing: meaning connections from the linguistic data in J.F. Lee & VanPatten, 1995, p. 96). Intake: T he subset input data that is avai lable to learners for further language processing Interpretation: grammatical feature (i.e. too help them carry out a form function mapping). In this case, the goal is grammar comprehension to di stinguish what might be termed message comprehension which can take place without the learner having to attend to the R. Ellis, 1995, p. 94). Interpretatio n was measured by three forms of an Interpretation Subtest that were created for this study. These tests were de signed to measure whether participants understood that the subjunctive in adjectival clauses in Spanish connotes a non referential antecedent (a referent that is hypothetical, uncertain, or unknown to the speaker). Noticin g: What learners detect in their linguistic input with conscious awareness (Schmidt, 1990). Noticing was measured by notes that participa nts took as they read an authentic input text following their instructional treatments. Participants were instructed

PAGE 44

30 to note the words that they cons idered to be important to comprehend an authentic input passage that was delivered online. They recorded their notes in a text box, and their notes were converted to a note score. One point was awarded for each instance o f the targeted grammatical form that was noted. Output: What learners produce, verbally or in writing, in the target language. Processing Instruction: In the present study, processing i ns truction was operationalized as an explicit grammar explanation of the targe ted grammatical form that was not paradigmatic followed by information about processing strategies and how to avoid faulty input processing followed by referential and affective structured input activities. Production: Three versions of a P roduction Subtest were created for this study to measure production. production of the present subjunctive in adjectival clauses in Spanish when the referent is unknown to the speaker. Participants were requir ed to determine if a verb had to be conjugated in the present indicative or in the present subjunctive in Spanish depending upon the context of the sentence, and they had to produce the correct verb form in writing. The Production Subtest requ ired participants to produce both regular and irregular verbs in the subjunctive as well as verbs that take an orthographical change. Text Comprehension: content in the input for the purpose of interpret J.F. Lee & VanPatten, 1995, p. 96).

PAGE 45

31 Traditional Instruction: The present study operationalized traditional instruction as an explicit grammar explanation of the targeted grammatical form that was paradigmatic followed by mechanical, then meaningful, then communicative activities. L earners were required to produce target language output immediately following the explicit grammar explanation. Visual Input Enhancement: A visual means of rendering certain features of the written input dat attention In the present study VIE was operationalized as the enhancement of targeted grammatica l forms through the use word animation as learners read input sentences online The subjunctive verb forms in e ach input sentence grew larger and smaller over a period of seven seconds to attract s as they worked online. Animated words were delivered consecutively rather than simultaneously in order not to distract parti attention from other static elements on the screen. Delimitations The findings of this study are not generalizable to the entire population of Spanish language students in the United States because there was not random selection of participants from universities across the country. The findings are only generalizable to students of Spanish from urban/suburban universities in the southeast who complete their foreign language coursework online. Two teachers of second semester Spanish language stu dents who deliver their instruction online from two institutions (one urban university and one suburban university) in the southeast were invited to have their classes take part in the study.

PAGE 46

32 Limitations As with all studies, the present study was not free from threats to internal validity. The pre and posttests were delivered over the web since the online Spanish courses from which students were selected had few face to face class meetings (for orientations, reviews, and exams only). In order to keep th e students from consulting their texts, notes, other individuals, or resources on the Internet, the pre and posttests were timed, only allowing enough time for students to answer each question rapidly and from their own working memories. Due to the natur e of timed tests, some students may have experienced greater test anxiety than others, which may have inhibited their performance on the pre and posttests. In addition, it is likely that extraneous variables such as gender, age, SES, and language aptitud e may have exerted some influ ence on the outcome measures. However, t hese variables were controlled by random assignment to groups. As all three instructional methods were delivered via WBI, teacher was not an extraneous variable in the present study.

PAGE 47

33 Chapter 2 Review of Literature Introduction m odel of i nput p rocessing (1993, 1996, 2002, 2004), which serves as the foundation for processing instruction (PI), a relatively novel type of focus on form i nstructional technique that utilizes tasks and processing of specific grammatical forms. PI is unique in that it is one of the few instructional methods that is informed by second language acquisition (SLA) research. A thorough description of PI and its characteristics are provided in this chapter. In addition, the relevant research on the efficacy of P I, and structured input, a component of PI, are reviewed, and any lim itations, design flaws, or gaps in the literature are explicated. As input enhancement and traditional (output based) instruction are also examined in the present study, the pertinent literature on these topics is also reviewed in this chapter. It is the aim of the present study to build upon and add to the current body of knowledge on the efficacy of each of the aforementioned instructional techniques. The chapter ends with an examination of the targeted grammatical form under investigation in the prese nt study, the present subjunctive in adjectival clauses in Spanish when the referent is hypothetical or unknown.

PAGE 48

34 m odel of i nput p rocessing (1993, 1996, 2002, 2004), by his own admission, is a working model and has not yet evolved to the level of theory in SLA. Although the model is not a theory of SLA, the principles and subprinciples that are attached to it serve as the foundation for PI. VanPatten (2004) defines input processing m meaning mappings that occur during real time comprehension of TL input Thus, a prerequisite for the model is the provision of comprehensible input to the learner. Form meaning mappings occur when learners assign meaning to a grammatical form, which is defined as a surface feature of language. Grammatical forms include inflectional morphology as well as function words such as pronouns, working memories; thus, it is not directly observable. In order for input processing to occur, learners must first perceive and notice a grammatical form in their TL input, and then they must assign meaning to it. It is important to note that L2 learners may process forms incorrec tly by assigning incorrect meanings or functions to forms. Also, input processing is constrained by working memory limitations, or the amount of information that can be perceived, noticed, and processed as L2 learners attempt to comprehend TL input in rea l time (Just & Carpenter, 1992). VanPatten (1993, 1996, 2002, 2004) defines input processing as the initial phase of the acquisition process, and he is careful to explain that his m odel of i nput p rocessing is not a model of SLA. See Figure 1 for VanPatte Acquisition. According to VanPatten (2004), once a grammatical form is processed, it

PAGE 49

35 becomes intake, which is input that has been filtered by the learner and that is available for further language processing. Once ini tially processed, a form may be either partially or fully accommodated into the developing linguistic system (McLaughlin, 1990). representations that as an aggregate constitutes the create an unconscious system of rules that govern phonology, morphology, syntax, and semantics. This unconscious system of rules is referred to as an implicit linguistic (or developing) system, which is made up of complex and varied components that interact with one another ( J.F. Lee & VanPatten, 2003). As learners encounter new forms and structures, in order to acquire them, they mu st accommodate new knowledge into their already existing implicit linguistic systems. If accommodation occurs, then it may trigger L2 learner s to restructure their internal gramm ars. Restructuring is a necessa ry precursor to production, which requires le arners to access their developing systems in order to produce target language (TL) forms (Gass, 1988, 1997; Terrell, 1991) According to VanPatten (2004), output is not p art of the basic processes in language acquisition. Rather, the production of output is a result of the acquisition process. He claims that acquisition occurs when learners use their input to take in and store pairs of form meaning relationships. Conversely, when learners produce output they must retrieve TL forms that are already part of SLA is presented in Figure 2.1.

PAGE 50

36 I II III input intake developing system output I = input processing II = accommodation, restructuring III = access, production procedures Fig ure 2.1 A Sketch of B asic Processes in A cquisition ( f rom VanPatten, 1993) Input processing is only concerned with phase I, or how input becomes intake, which is the starting point for acquisition. Because human beings are limited capacity processors (J ust & Carpenter, 1992; McLaughlin, Rossman, & McLeod, 1983; Schneider further language processing. Gass (1988) and VanPatten (1993, 2004) posit that learners activel y contribute to the selection of their input that is noticed. Gass claims that input is apperceived, or noticed, when learners are able to relate it to their prior knowledge. tion during input processing. Further, he posits that learners process meaningful input first, such as lexical items and grammatical forms that have a high communicative value. The communicative value of a grammatical form refers to the extent to which t he form contributes to the overall referential meaning of a sentence or utterance (VanPatten, 1996).

PAGE 51

37 A form has a high communicative value if it has inherent semantic value (it conveys some type of meaning) and lacks redundancy [+semantic value] [ redund ancy] within the sentence or utterance. For example, in the sentence Mary watched television for two hours yesterday, the grammatical morpheme ed has inherent semantic value [+semantic value] because it conveys a past tense meaning. However, the lexical item yesterday also conveys past tense meaning; therefore the ed verb ending in this sentence is redundant [+redundant]. Removing lexical items that express the same meaning as the mantic value cannot be manipulated or changed. m odel of i nput p rocessing (1993, 1996, 2002, 2004) forms with a low communicative value can be processed, but only when learners are able to process other items in the sentence or ut terance without difficulty (without draining all of their processing resources), thus leaving enough resources available to process the grammatical form with a low communicative value. m odel of i nput p rocessing (2004) is founded up on two main principles and several subprinciples VanPatten states them as follows : Principle 1. The Primacy of Meaning Principle. Learners process input for meaning before they process it for form. Principle 1a. The Primacy of Content Words Principle Learners process content words in the input before anything else. Principle 1b. The Lexical Preference Principle. Learners will tend to rely on lexical items as opposed to grammatical form to get meaning when both encode the same semantic information Principle 1c. The Preference for Nonredundancy principle. Learners are more likely to process nonredundant meaningful grammatical forms before they process redundant meaningful forms.

PAGE 52

38 Principle 1d. The Meaning Before Nonmeaning Principle. Learners are more likely to process meaningful grammatical forms before nonmeaningful forms irrespective of redundancy. Principle 1e. The Availability of Resources Principle. For learners to process either redundant meaningful grammatical forms or nonmeaningful forms, the processing of overall sentential meaning must not drain available processing resources. Principle 1f. The Sentence Location Principle. Learners tend to process items in sentence initial position before those in final position and those in med ial position. Principle 2. The First Noun Principle. Learners tend to process the first noun or pronoun they encounter in a sentence as the subject/agent. Principle 2a. The Lexical Semantics Principle. Learners may rely on lexical semantics, where pos sible, instead of word order to interpret sentences. Principle 2b. The Event Probabilities Principle. Learners may rely on event probabilities, where possible, instead of word order to interpret sentences. Principle 2c. The Contextual Constraint Princ iple. Learners may rely less on the First Noun Principle if preceding context constrains the possible interpretation of a clause or a sentence. ( 2004, p. 14) Principle 1, or the Primacy of Meaning Principle, and the first two subprinciples are based upo n research on L1 and L2 acquisition, which found that both L1 and L2 learners attempt to seek out the communicative intent of their input at the expense of processing grammatical form, and that learners primarily extract meaning from content words ( Frch & Kasper, 1986; Klein, 1986; Sharwood Smith, 1986; Peters, 1985; Wong Fillmore, 1976). The aforementioned research revealed that learners tend to skip over function words and inflections while parsing the communicative intent of sentences or

PAGE 53

39 utterances fro m content words (usually in the form of lexical items). In other words, learners are pushed to extract meaning from their input, and the quickest way to do so is to focus on content words. The researchers also found that function words and inflections ma y be chunked, or fused to the content words with which they normally appear. Alternatively, function words and inflections may be only partially processed exhausted by the task demands of processing lexical items. Furthermore, some grammatical forms may be perceived and noticed by learners, but due to constraints on working memory from processing content words, L2 learners, especially novice learners, are often unable to connect meaning to noticed grammatical forms (VanPatten, 2004). meaning from content words in their input at the expense of processing grammatical forms, VanPatten devised th e Primacy of Meaning Principle along with subprinciples P1a words and lexical items while skipping over grammatical forms during input processing. Subprinciples P1c and P1d refer to how the relative communicative value of grammatical forms affect the way that learners process them. In essence, VanPatten (2004) claims that meaningful forms that are not redundant [+semantic value] [ redundancy] are proc essed before meaningful forms that are redundant [+semantic value] [+redundancy]. Additionally, VanPatten asserts that meaningful forms [+semantic value] are processed before nonmeaningful forms [ semantic value], regardless of whether or not they are red undant. Research conducted by J.F. Lee (1987, 2002)

PAGE 54

40 supports subprinciples P1c and P1d. J.F. Lee (1987) found that L2 learners skip over grammatical forms with a low communicative value [ semantic value] [+redundancy] during input processing, and J. F. L ee (2002) found that L2 learners are able to process novel forms when they have a high communicative value [+semantic value] [ redundancy], even when they have not been presented formally through instruction. Subprinciple P1e is based upon research by Blau (1990), Hatch (1983) and Long acquisition also increases. Subprinciple P1e reflects this finding by stating that when learners are able to process the over all meaning of a sentence or utterance with little or no cost to attention, then there is an increased likelihood that they will be able to process either redundant meaningful forms [+semantic value] [+redundancy] or nonmeaningful forms [ semantic value]. VanPatten (2004) asserts that certain factors may influence the amount of processing resources that are available to learners; these include proficiency level and familiarity of lexical items in the input string. If L2 learners are already familiar with the lexical items in their input and they can be easily accessed during comprehension, they will have more processing resources available for redundant and nonmeaningful grammatical forms (VanPatten, 2004). Subprinciple P1f is based on research by Ba r crof t and VanPatten (1997) and Klein (1986), who found that elements that are in the initial position of sentences or utterances are more salient for learners than elements that are in sentence medial or sentence final positions. VanPatten (2004) claims that grammatical forms that are in the sentence initial position are more likely to be processed before items that are in the

PAGE 55

41 not yet exhausted at the beginning of an input string. Further, VanPatten (2004) asserts that items in sentence final position are processed more readily by learners than items in the sentence medial position because when learners reach the middle of the string of input, their processing resources hav e already been exhausted by processing items in sentence initial position. However, when learners reach the end of an input string, they are inclined to redirect their attention to processing the input string, and some attentional resources may become fre ed up to process items in the sentence final position (VanPatten, 2004). Thus, items in the sentence medial position have the least chance of being processed by learners. He qualifies this assertion by stating that the length of the sentence or utterance sentences or utterances being more difficult to process than shorter ones, especially for novice L2 learners. While the Primacy of Meaning Principle and its subprinciples refer to aspects of second language morphology, Principle 2, or the First Noun Principle, and its subprinciples refer to the interpretation of second language syntax. The First Noun Principle describes how many L2 learners whose first language is SVO (subject verb object) or SOV (subject object verb) often interpret the first word or noun that they encounter in a sentence or utterance in the L2 as the subject. VanPatten (1993, 2004) an d passives. Research by Ervin Tripp (1974), J.F. Lee (1987), LoCoco (1987) and VanPatten (1984) appears to support this claim. Ervin Tripp found that L2 learners of

PAGE 56

42 French whose L1 was English (an SVO language) confused the first noun in the French passi ve structure as the agent rather than the patient even though English and French passive constructions have the same word order. Both J.F. Lee (1987) and VanPatten ( 198 4) found that L2 learners of Spanish whose L1 was English commonly confused the object of the sentence or utterance as the subject when the subject pronoun was omitted, which is a common occurrence with Spanish and other pro research indicates that L2 learners of German whose L1 was English ignored case markings and interpreted the first noun as the subject of the sentence when it was marked as the object even after they had been formally taught case markings in class. Although the First Noun Principle does not account for learners whose L1 is OVS (object verb subje ct), VanPatten (2004) asserts that the majority of world languages are either SOV or SVO languages, where the canonical order is subject before object. Therefore, he claims that the tendency is for most L2 learners to interpret the first noun that they en counter as the subject of the sentence or utterance. Further, he suggests that the default parameter for syntax from the perspective of Universal Grammar (UG) may be SVO, but he cautions that research needs to be conducted with L1 and L2 learners of OVS l anguages before this claim could be supported. Subprinciples P2a, P2b, and P2c refer to how the First Noun principle may be constrained by other factors. Subprinciple P2a describes how lexico semantic information can weaken the First Noun principle, na mely by assisting L2 learners with the correct interpretation of pronouns, object pronouns, and/or case markings. For example, in the English passive construction the ball was kicked by the boy rather than

PAGE 57

43 interpret the first noun the ball as the subject or agent of the sentence, L2 learners of English will be assisted by the lexical semantics of the verb to kick which requires that the agent of the verb be an animate being. However, in the sentence the boy was kicked by the horse, L2 learners of Englis h are likely to misinterpret the boy as the subject or agent of the sentence because both boys and horses are animate and capable of kicking (VanPatten, 2004). According to subprinciple P2a, since there is a lack of lexical semantics to constrain the agen cy of the verb in the previous example, L2 learners would tend to process the first noun that they encounter as the subject or agent of the sentence. well as contextual cues as they interpret sentences and utterances in the TL. VanPatten sentence the ant was stepped on by the girl of ants and girls will assist them with the correct interpretation of the aforementioned example regardless of the First Noun Principle. Subprinciple P2c is similar to subprinciple P2b ; however, the former refers to elements in the preceding context that constrain the possible interpretation of sentences or utterances while the latter refers to interpretation of sentences or utterances. VanPatten (2004) asserts that the principles may interact or combine to delay Research by VanPatten (1984) indicates that L2 learners of Spa nish usually skip over

PAGE 58

44 nonredundant object markers that occur at the beginning of sentences, which causes them to misinterpret the object pronoun as the subject of the sentence. For example, in the Spanish sentence A Paco ve Mara or Mara sees Paco the Sentence Location Principle does not assist L2 learners of Spanish with the correct interpretation of the sentence even though the object marker a occurs at the sentence initial position, which clearly indicates that Paco is the object and not the subject of the sentence (VanPatten, 2004). Rather, VanPatten (1984) found that Spanish language learners whose L1 is English skip over and fail to process the object marker a because the First Noun Principle incorrectly drives them to process Paco the first noun that they encounter, as the subject of the sentence. Based on this finding, VanPatten (2004) suggests that the First Noun Principle, is stronger than P1f, the Sentence Location Principle. Although VanPatten (2004) claims that more than one principle i n his model of i nput p rocessing can operate at the same time and that some principle s are more powerful than others; thus far there is scant research to support these claims. More research studies are needed that specifically examine if and how the princ iples interact, and/or which ones are more powerful than others. At this time it is not possible to make any definitive assertions regarding the interaction and relative power of the principles in i nput p roce ssing (1993, 1996, 2002, 2004) appears to describe how L2 learners initially process their input, several scholars have questioned the basic assumptions of the model (DeKeyser, Salaberry, Robinson, & Harrington, i nput p rocessin g ha s been criticized because it is does not

PAGE 59

45 take into account current models of sentence processing (parsing). Although VanPatten maintains that his model of i nput p rocessing is not a psycholinguistic model of sentence processing, Harrington (2004) and D eKeyser, Salaberry, Robinson, and Harrington (2002) assert that any model of input processing should be consistent with what is presently known about sentence processing. According to Harrington (2001), sentence processing research attempts to understand comprehend and/or produce sentences in real time. The fields of artificial intelligence, psycholinguistics, and computational linguistics conduct research on sentence processing in an attempt to understand ho w various sources of linguistic and extralinguistic knowledge interact to yield meaning in real time (Clifton, Frazier, & Rayner, 1994). There are three cognitive mechanisms that are responsible for sentence processing: algorithms, heuristics, and represe ntations (Harrington, 2001). Algorithms are IF THEN production rules that are responsible for making meaning out of linguistic input, and heuristics are principles that constrain how algorithms function (Caron, 1992; Harrington 2001). Harrington (2001) d m odel of i nput p rocessing do not correspond with algorithms, heuristics, and/or representations as they a re traditionally used to describe sentence processing mechanisms. However, sentence processing research almost exclusively focuses on how mature individuals parse input in their L1. VanPatten contends that L1 parsing models are not helpful for L2 learner s because unlike L1 speakers, they do not have intact

PAGE 60

46 cannot and is not the same process as that for beginning non natives who must not only comprehend but also come to Further, he makes no claim that his model of i nput p rocessing is a model of sentence processing; rather, he states that the model is a description of the strategies that L2 learners use as they at tempt to comprehend TL input in real time. In other words, VanPatten is not concerned with how the L2 parser develops; rather, he is more concerned with identifying the faulty processing strategies in which L2 learners tend to engage. The goal of PI is t o help learners avoid faulty input processing strategies and move toward more favorable ones. Processing Instruction model of i nput p rocessing (1993, 1996, 2002, 2004), which ent ails a set of principles that describe the processing strategies that L2 learners use to make meaning out of their TL input. In his model, VanPatten describes how L2 learners engage in the initial processing of TL input, which he refers to as making form meaning connections. The goal of PI is to change or manipulate the way that learners initially perceive and process TL input, and it is a completely input based approach to FL instruction. PI is in direct contrast to traditional instruction, which attemp ts to facilitate acquisition by focusing on the a and 2.2b present a visual depiction of these two contrasting instructional methods.

PAGE 61

47 ____________________________________________ __________________________ Input intake developing system output processing mechanisms focused practice _________________________________________ _____________________________ Figure 2.2a Processing Instruction in Foreign Language T eaching ( f rom VanPatten & Cadierno, 1993 b ) ______________________________________________________________________ Input intake developing system output focused practice ______________________________________________________________________ Figure 2.2b Traditional Explicit Grammar Instruction in Foreign Language T eaching (from VanPatten & Cadierno, 1993 b )

PAGE 62

48 i nput p rocessing ( 1993, 1996, 2002, 2004 ) L2 learners are driven to extract meaning from their input at the expense of processing grammatical form (the Primacy of Meaning principle), and L2 learners often misinterpret their input based upon the order of words in sentences and/or utterances (the First Noun Principle). Due to these tendencies, L2 learners will often engage in inefficient and/or faulty processing of their TL input. PI was dev eloped to push learners away from flawed processing strategies in favor of more optimal ones (Wong, 2004). The first step in PI is identified, instructors can then alert their students to th e processing problem and provide information about c orrect input processing strategies. The final step in PI is to create structured input tasks and activities, which encourage learners to abandon their faulty strategies in favor of correct input processing strategies. There are three characteristics of PI: (a) explicit instruction on the targeted grammatical form or structure, (b) information about processing strategies that may be causing delays in acquisition, and (c) structur ed input tasks and activities. Although explanation of TL forms is a component of PI, it may not be a critical feature. Research by VanPatten and Oikkenon (1996) isolated the components of PI and found that grammatical explanation is not a necessary feat ure of PI. In contrast, Farley (2004) found that the provision of an explicit explanation of grammar i s a key component of PI, especially when the targeted grammatical form is complex. The e xplicit grammar explanation in PI is different from traditional and other types of instruction because it is not paradigmatic. With PI, grammar is presented with a focus on only one form at a time

PAGE 63

49 Traditional instruction typically presents the full paradigm of a grammatical form or structure on the same day, while PI breaks up paradigms into smaller chunks. with determining which processing problems learners are likely to experience given the targeted form or structure that is being taught. After the provision of an explicit explanation of the grammar point, learners are provided with specific information about the faulty processing strategies that they are likely to engage in and that may hinder their acquisition of the TL form. After receiving information about processing strategies, students are given structured input activities. Although somewhat complex to design, structured input activities are intended to push learners away from faulty processing strategies and move them towards more efficient ones. Structured input is an input enhancement technique that elevates the communicative valu e of a linguistic form by eliminating any lexical redun d ancies in the input and by simplifying the input surrounding the targeted structure. Structured input is an enhancement technique that increases the likelihood that L2 meaning con nection. Further, structured input activities provide learners with multiple opportunities to make correct form meaning mappings, which according to VanPatten (1996) increases the likelihood that acquisition will take place. There are two types of struct ured input activities: referential and affective. With referential activities, students are required to interpret the targeted grammatical form (or realize the meaning a specific

PAGE 64

50 feature encodes) in order to answer questions correctly. Conversely, affect ive structured input activities focus more on using the targeted grammatical forms in meaningful L2 communication. With these activities, learners complete tasks that help them to become hood that they will process the targeted forms. This involvement can include a variety of activities such as matching, answering yes/no, or checking off items on a list. It has been argued, however, that learners may not notice the relevant grammatical f orm during affective structured input activities because the focus in on meaning rather than on form Hwu (2004) has criticized affective structured input activities because learners do not typically have to notice or comprehend the targeted form in order to understand the communicative intent of each sentence. Thus, Hwu advocates using some type of input enhancement to help learners notice targeted grammatical forms in structured input activities where meaning can be extracted from other elements in the sentence. It is important to note that learners never produce the grammatical form during either type of structured input activity (referential or affective), as both types are input based. J. F. Lee and VanPatten (1995, 2003) proposed six specific guide lines for developing structured input activities, which will be discussed further in Chapter 3. A criticism of structured input activities is that most practitioners are unfamiliar with i nput p rocessing ( 1993, 1996, 2002, 2004 ) and th e faulty input processing strategies in which L2 learners are likely to engage. Further, structured input activities have specific and somewhat complicated guidelines that must be followed in order to create them accurately. Although they may be highly e ffective, structured input

PAGE 65

51 activities are not yet currently widely employed in practice due to aforementioned limitations. Empirical Studies on Processing Instruction The first study on PI was conducted by VanPatten and Cadierno (1993a), and they investig ated P 2 or the First Noun Principle, which states that L2 learners tend to process the first noun or pronoun that they encounter in a sentence as the subject or agent of the sentence. VanPatten and Cadierno ( 1993a) examined this principle with 80 L2 lear ners of Spanish who were in their second year of university level language study. The targeted grammatical form was object pronouns and syntax (word order) in Spanish. Since Spanish is a pro drop language, in sentences containing object pronouns, they ar e often the first word. When object pronouns appear at the sentence initial position, L2 learners of Spanish often tend to process them as the subject of the sentence. This faulty processing strategy often leads to a misunderstanding of the meaning of the sentence. For example, the Spanish sentence Lo bes Mara is often interpreted He kissed Mara rather than Mara kissed him (the correct interpretation). VanPatten and Cadierno (1993a) compared PI with traditional instruction for the acquisition of obj ect pronouns in Spanish. They also included a control group that received no instruction. Traditional instruction was operationalized as an explicit grammar explanation of the targeted grammatical form followed by mechanical, then meaningful, then open e nded communicative activities. Of note, in traditional instruction, the full paradigm of direct object pronouns were presented to the learners at one time, and following this initial grammar presentation/explanation of the novel forms,

PAGE 66

52 the learners were r equired to produce them immediately in speech and in writing. The manual as the basis for the traditional instruction materials. PI consisted of an explicit grammar expl anation, information on processing strategies, and structured input activities. The grammar explanation in PI differed from traditional instruction in two important ways: (a) p articipants received grammar instruction that broke up the full paradigm of dir ect object pronouns into two parts, and (b) p articipants received information on processing strategies that helped them to differentiate subject and object pronouns. By providing training on processing strategies, the researchers attempted to circumvent t he First Noun Strategy, which states that L2 learners will process the first noun that they encounter in a sentence or an utterance as the subject or agent. Thus, participants were trained to interpret subject and object pronouns correctly before receivin g any structured input activities. All of the activities in the PI materials were input based, and at no time did any of the participants in the PI group produce any of the targeted grammatical forms. Both instructional treatments were delivered via penc il and paper instructional activity packets that were completed over a period of two class sessions. The of the researchers Cadierno taught both experimental groups. Participants completed all of the study related materials in their classrooms and did not have any homework during the study period. VanPatten and Cadierno (1993a) balanced both experimental treatment packages for the total number of activities, percenta ge of aural versus written activities,

PAGE 67

53 percentage of whole class versus paired activities, number of tokens (number of input sentences containing the targeted forms interpreted by the PI group versus number of sentences containing the targeted forms produc ed by the traditional instruction group). Also, the researchers adjusted the vocabulary in both treatment packets so that they contained roughly the same vocabulary items, which consisted of high frequency words that were already familiar to participants from previous Spanish language coursework. Finally, the researchers also checked the vocabulary items in the assessment tasks against the vocabulary items in both instructional treatment packets in order to ensure that there was no vocabulary bias for eit her of the two treatment groups. VanPatten and Cadierno (1993a) employed an experimental design with a true control group that received no instruction on object pronouns in Spanish. The 80 participants were given pre and posttests that included interpre tation an d production tasks. The pre and posttest scores on each task were submitted to two repeated measures Analyses of Variance (ANOVAs), which revealed a between subjects effect for instruction on each task type. Post hoc tests revealed that for the production task, the test scores for PI and traditional instruction were not significantly different from each other, but both of these groups scored significantly higher than the control group. Additionally, the post hoc tests revealed that for the inte rpretation task, the PI group made significant gains, but the traditional instruction group and the control group did not. VanPatten and Cadierno (1993a) claim that these results indicate that PI is superior to traditional instruction because although par ticipants in the PI group never produced the targeted forms during the instructional treatments, they scored equally as well as those

PAGE 68

54 participants who did. Further, participants in the PI group scored significantly higher than both the traditional instruc tion group and the control group on the interpretation task. Therefore, PI was found to be effective for assisting learners in interpreting object pronouns and word order correctly in Spanish. When both types of tasks are taken into consideration (interp retation and production), the results of this study indicate that PI is superior to traditional instruction and to no instruction. VanPatten and Cadierno (1993b) replicated their study with the same grammatical form (object pronouns) and 49 second year un iversity level learners of Spanish. They employed the same research design, instructional treatments, procedures, and assessment tasks as VanPatten and Cadierno (1993a); however, their results were slightly different in the 1993b study. Similar to VanPat ten and Cadierno (1993a) the researchers found that PI was superior to traditional instruction and to the control group for interpretation task measures. However, the results of the production task measures were less apparent. A repeated measures ANOVA o n production task scores revealed a significant between subjects effect for type of instruction, a main effect for time, and an interaction effect between type of instruction and time. Post hoc Sheff tests revealed that the main effect for type of instr u ction was due to one contrast: trad itional instruction outperformed control (no instruction). The researchers point out, however, that the post hoc tests also revealed that there were no significant differences between PI and traditional instruction, nor were there any significant differences between PI and control on production task

PAGE 69

55 not superior to processing instruction on the production task, and on the second and t hird With the results of both studies taken together (VanPatten & Cadierno, 1993a, 1993b), the researchers claim that PI is superior to traditional instruction and to n o instruction. VanPatten and Cadierno (1993a) suggest that PI has a beneficial effect on new knowledge for L2 production, even though they never actually produced th e targeted forms during their instructional treatment. They also claim that traditional instruction did gains on production tasks and not on interpretation tasks (VanP atten & Cadierno, 1993a, 1993b). VanPatten and Cadierno (1993b) assert that the results of their study are although linguistic rules can be learned, they do not he lp learners use or produce language during real time communication. In other words, languages cannot be learned, they can only be acquired through exposure to input, and acquisition is an unconscious process. The researchers in the present study claim tha t PI results in acquisition of the targeted form while traditional instruction leads only to form learning and not to acquisition. VanPatten and Cadierno (1993a) state: [T] raditional grammar presentation and practice do not feed into the developing syste m directly but instead result in a different knowledge system. Krashen (1982) has suggested that learners may develop two systems an acquired competence and a learned competence and has claimed that traditional instruction results in learned competenc e, but only by accessing comprehensible input can the acquired system build up.

PAGE 70

56 however, have been heavily criticized in the literature as not being falsifiable (Morgan Short & Wood Bowden, 2006; Salaberry, 1997; DeKeyser, Salaberry, Robinson, & Harrington, 2002). Morgan Short and Wood Bowden (2006) suggest that both PI and traditional instruction may have resulted in the same type of knowledge, but with different strengths, a mounts, or degrees. They also state that it is impossible to claim that and posttest scores because they received expl icit instruction and feedback addition to struct ured input. Therefore, given the research design of VanPatten and Cadierno (1993a), it is impossible to determine which component facilitated acquisition. Of note, in subsequent studies that investigated PI, VanPatten made no such claims regarding how PI (1982) learning versus acquisition distinction. Subsequent studies that compared PI and traditional instruction include Cadierno (1995) who investigated PI and Spanish preterit tense morphology. She employ ed an experimental research design that paralleled VanPatten and Cadierno (1993a, 1993b), with two treatment groups (PI and traditional instruction) and a true control group that received no instruction. Her participants were third semester undergraduate students of Spanish. Cadierno addressed Principle 1b, or the Lexical Preference principle, in m odel of i nput p rocessing (1993, 1996, 2002, 2004). This principle states that learners will rely on lexical items instead of grammatical form to de rive meaning from a sentence or an utterance when both encode the same semantic information. In

PAGE 71

57 order to circumvent this faulty processing strategy, Cadierno raised the communicative exical redundancies (specifically temporal adverbs). Thus, the PI activities that she created forced participants to attend to and process tense markers for the preterit, which are usually nonsalient for learners. For example, she elimated words such as ayer or yesterday and la semana pasada or last week in order to compel learners to examine the preterit tense inflectional morphemes to interpret the time reference of each sentence (past or present). This distinction is particularly difficult in Spanish because the first person singular inflectional morpheme for the present tense is an unaccented o and the third person singular inflectional morpheme for the preterit tense for verbs ending in ar (the most common type of verb in Spanish) is an accented For example, in Spanish I speak is rendered hablo while he spoke is rendered habl It is only an acoustic stress that distinguishes these two verb forms in speech, and a written accent mark that distinguishes them in writing, which is often problemat ic for novice L2 learners of Spanish. The traditional instruction materials contained a combination of mechanical, meaningful, and communicative (open ended) activities where the participants were required to produce output in the TL immediately following a grammar explanation. As with the previous two studies that examined PI (VanPatten & Cadierno 1993a, 1993b), participants never once produced the targeted form during the PI treatment. The research design, treatment procedures, and assessment tasks in Cadierno (1995) paralleled those of VanPatten and Cadierno (1993a, 1993b). In order to analyze the interpretation and the production assessment tasks scores on the pre and posttests,

PAGE 72

58 Cadierno empl oyed a repeated measures ANOVA with type of instruction as the between subject s factor and time as the within s ubjects variable. The ANOVA revealed a significant between subjects effect for type of instruction, a significant main effect for time of testing and a significant interaction effect between type of Ins truction and time. Post hoc Sheff tests revealed that PI was superior to TI and to control on the interpretation task for all posttests. A repeated measures ANOVA was also performed on production test scores with type of instruction serving as the betwe en subjects variable and time serving as the with subjects variables. The results of the ANOVA revealed a significant between subjects effect for instruction, a significant main effect for time, and a significant interaction effect between type of instruc tion and time. Post hoc Sheff tests revealed that PI and traditional instruction were superior to no instruction, and that there was no significant difference between PI and traditional instruction on the production task scores. These results corroborat ed the findings of VanPatten and Cadierno (1993a, 1993b) even though Cadierno (1995) investigated a different grammatical form. Similar to VanPatten and Cadierno (1993a, 1993b), Cadierno (1995) also claims c systems and results in acquisition of the targeted grammatical form while traditional instruction only results in form learning, which is not useful during real time communication in the TL. As mentioned previously, this assertion has been heavily criti cized by several SLA scholars (Morgan Short & Wood Bowden, 2006; Salaberry, 1997; DeKeyser, Salaberry, Robinson, & Harrington, materials were meaningful, while onl y some of the traditional instruction materials were

PAGE 73

59 meaningful. About half of the traditional instruction materials were mechanical, where participants only had to attend to form in order to answers questions correctly. It may be that the significant di fferences found between the treatment groups was due to the meaningful tasks that were present in the PI materials versus the mechanical tasks that were present in the traditional instruction materials. Cadierno (1995) asserts, however, that traditional i nstruction as it was operationalized in her study (mechanical, followed by meaningful, followed by open ended activities) possible t o compare PI with traditional instruction without including mechanical activities. Further, it is also unclear which components of PI (explicit grammar explanation, information on processing strategies, or structured input) may have been responsible for p to posttests since the researcher did not isolate these components. Given that VanPatten and Cadierno (1993a, 1993b) employed the same research design as Cadierno (1995), these same c riticisms apply to the previous studies as well. The previous three PI studies reviewed (VanPatten & Cadierno, 1993a, 1993b; Cadierno, 1995) had findings that lend weight to the claim that PI is superior to traditional instruction for interpretation tasks and that PI is equal to traditional instruction for production tasks. The finding that participants in PI groups performed as well as participants in traditional instruction groups on production tasks is remarkable, especially given that PI participants n ever once produced any of the tar geted grammatical forms during instructional t reatments. Conversely, traditional instruction participants produced

PAGE 74

60 targeted forms immediately following an initial explanation of grammar, and they continued to produce targe ted forms both orally and in writing throughout the instructional treatments. The production tasks in these studies (VanPatten & Cadierno, 1993a, 1993b; Cadierno, 1995), however, were limited and only comprised sentence level tasks. In order to determin communicative tasks, VanPatten & Sanz (1995) investigated the acquisition of syntax and word order (the same target form as VanPatten & Cadierno, 1993a, 1993b) on discourse level pr oduction tasks with undergraduate learners of Spanish in their second year of language study The researchers used the same instructional treatment and testing materials as VanPatten and Cadie rno (1993a, 1993b), which they expanded upon for the ir study. VanPatten and Sanz created two additional production tasks to measure (a) a video narration task, and (b) a structured question answer interview. Each production task had an oral and a written version. Thus, there were six assessmen t tasks in total; three production tasks will two versions each. The video narration task and the question level production tasks that were used in past PI studies. In other words, parti cipants were encouraged to produce the targeted forms in order to communicate a message in the TL without an unnatural repetition of object pronouns. Of note, VanPatten and Sanz included oral production tasks on their posttests. Past PI studies only meas compare PI with traditional instruction; rather, they only compared a PI group and a

PAGE 75

61 control group that received no instruction. Given that VanPatten and Sanz were only interested in pa instrument in their study. On the pretest, participants who sc ored higher than 60% were excluded from the study. In addition, participants who did not show a gain from pre to posttest on the interpretation task were not included in the study. VanPatten and Sanz (1995) attempted to examine only the production of those participants who benefited from PI as meas ured by the interpretation task. As the researchers were examining the effec ts of PI on output, t hey wanted to determine if learners who benefited from PI on an interpretation task (where no output was required) were also able to benefit on communicative tasks where both oral and written output was required The researchers submi tted each production task to separate repeated measur es ANOVA s. The PI group demonstrated significant gains from pre to posttests on the sentence level completion task and on the video narration task in the written mode; however, they only made slight ga ins on the question answer task in the written mode. For the oral mode, the PI gro up performed had significant improvements on the sentence level task across time ; however, they made no significant gains on the video narration task in the oral mode. The control group made no significant gains from pre to posttest s on any of the production tasks in either mode. These results indicate that PI is beneficial for production in the oral and written modes when the assessment tasks are less structured and more communicative in nature. VanPatten and Sanz also point out that sentence level tasks in the written mode are easier for beginning and intermediate level language

PAGE 76

62 learners than open ended communicative tasks in the oral mode. As expected, PI participants made the greatest gains on sentence level tasks in the written mode. A criticism of VanPatten and Sanz (1995) is that they did not report the total number of participants who benefited from the PI treatment as measured by the interpretation task Partici pants in the study were required to show a gain on this measure following the instructional treatment ( from pre to posttest ) in order to participate in the screening, partici pation in all phases of the study, and performance at 60% or below on the pretest, the final pool consisted of 44 subjects: 27 in the processing group and 17 in the control grou failed to report the total number of particip ants in their initial pool, it is unclear how many participants impr oved on the interpretation task Further, the number of participants assigned to each group was not equal, with the participants in the PI group outnumbering the participants in the cont rol group by 63%. It is typically easier to detect group differences, if they exist, when cell sizes are larger. (1995) study was quite small (only 17), yet the researchers employed univariate statistical tests (AN OVAs) where larger groups ar e often recommended in order for the test to have sufficient statistical power to detect group differences if they exist In addition, the researchers threw out the data on the question answer task. They stated that the data i t yielded was problematic because participants tended to repeat direct objects rather than replace them in the sentence with direct object

PAGE 77

63 to posttests for the PI group on this measure (p. 107). Although the gains may have been insignificant for the PI group on this measure, it is still important to report the data even VanPatten and San z threw out the data on one of the production assessment tasks, they still claimed that: [S] ubjects receiving processing instruction made gains on all tasks in the written mode and on two of the three tasks in the oral mode. Only on the oral video narrati on task did the analysis fail to yield a significant difference between pre and posttest performance (p. 111) This assertion appears to be incorrect and overly strong as the results were not significant for the oral video narration task and the data was thrown out for the question answer task. In other words, the only task that yielded significant differences in the oral mode was the sentence level production task. Thus, while reporting the results of this study, researcher bias may have been at play, especially given that the developer of PI was one of the principle investigators. VanPatten and Oikkenon (1996) investigated the components of PI to determine pro duction of targeted grammatical forms. The instructional materials and assessment tasks used in VanPatten and Oikkenon were taken from VanPatten and Cadierno (1993a, 1993b), and their study examined the same targeted form, Spanish direct object pronouns a nd syntax. The 59 participants were second year high school students of Spanish whose L1 was English. In their study, VanPatten and Oikkenon isolated explicit instruction and structured input by dividing participants into three treatment groups as follow s: (a) a

PAGE 78

64 group that received full PI, (b) a group that received explicit grammar explanation only, and (c) a group that received structured input only. Information on processing strategies, another component of PI, does not appear to be a variable in this study. Participants in the explicit explanation group read over the grammar explanation as the instructor reviewed it. The grammar explanation was followed by TL examples. Participants in this group were permitted to ask questions or request clarificat ions as needed. Although information on processing strategies would constitute explicit information, the research report does not specifically state that information on processing strategies was included in the grammar explanation for th e explicit instruc tion group. Further, t hese participants did not complete any instructional activities that focused on the targeted grammatical form following their grammar lesson. Conversely, participants in the structured input group only completed referential and stru ctured input activities and did not receive any type of grammar explanation during their treatme nt. The instructor indicated whether them. If participants asked abo ut the targeted forms, the instructor stated that they would Further, the instructor never directed ttention to the targeted form or informed them of any grammar rules while they completed their instructional treatment activities. Participants in the PI group received an explicit explanation of grammar information on processing strategies, and structured input activities. The posttest, which consisted of an interpretation task and a se ntence level production task, was administered one day after the instructional treatments were

PAGE 79

65 completed. P re and posttest scores were submitted to two repeated measures ANOVAs, one for the interpretation task scores and one for the production task score s. The ANOVA on the interpretation task scores revealed an interaction effect between type of instruction and time, a main effect for type of instruction, and a main effect for time Post hoc tests revealed that both the PI group and the structured input only group scored significantly higher from pre to posttest on the interpretation task. For the production task, the repeated measures ANOVA revealed a main effect for type of instruction and a main effect for time, but there was no interaction effect. Post hoc tests revealed that although all groups improved across time on the sentence level production task, only the PI group and the structured input group made significant gains. Therefore, the results reveal that an explicit grammar e xplanation alone was not beneficial for either production or interpretation tasks. actual structured input itself and the form meaning connections being made during input processing that are responsible for the observed effect in the present as well as the component of PI may be superfluous, and subsequent research on the components of PI (Benati, 2004a, 200 4b; Sanz & Morgan Short, 2004; Wong, 2004) appear to support this claim. However, Farley (2004) found that explicit instruction may be a necessary component of PI when the targeted grammatical form is complex, such as the Spanish su bjunctive. Interesting ly, Fern ndez (2008) examined the components of PI with two targeted grammatical forms, the subjunctive in nominal clauses following expressions of

PAGE 80

66 doubt and object pronouns in Spanish. She found that explicit information is helpful for the subjunctive (a complex form) but not for object pronouns (a more transparent grammatical form) Thus, the research results are mixed regarding the efficacy of the explicit information component (grammar explanation and information on processing strategies) of PI. It a ppears that explicit information alone is not beneficial, but it may be necessary to combine structured input activities with explicit information when the targeted grammatical form is complex. The results of VanPatten and Oikkenon, however, may have been distorted due to two factors for which the researchers did not control: (a) t ime on task, and (b) f eedback offered during the treatment. The three groups (explicit explanation only structured input only, and full PI) varied considerably in the length of their treatments. The explicit explanation group received games and activities unrelated to the targeted grammatical form following their treatment because the gram mar explanation did not take as much time as completing structured input activities or as full PI (explicit information and structured input activities). It may be that the amount of time spent on task (e.g. working with the targeted grammatical forms) influenced the findings of this study. Further, participants in the PI group and the str uct ured input group were told i f their answers were correct or incorrect. According to DeKeyser, Salaberry, Robinson, and Harrington (2002), this type of information helps learners figure out the rule system. Thus, the feedback provided to participants was an extraneous variable. Conversely, the group that received explicit information only did not have the opportunity to make mistakes and to receive feedback on their errors, which is a design flaw that was pointed out by Sanz

PAGE 81

67 ( 2004). Future research on P I should attempt to control for time spent on task during instructional treatments Also, future studies should attempt to equalize the feedback (amount and type) that is given to each group In an attempt to extend PI research to a semantic aspectual fea ture of language, Cheng (1995) compared PI to traditional instruction for the acquisition of the Spanish copulas ser and estar with adjectives and past participles, an area of complex grammar that typically causes difficulties for L2 learners of Spanish. VanPatten (198 5 1987) found the following acquisition order for the Spanish copulas ser and estar with stages III V taking place over a lengthy period of time when Spanish is learned in a foreign language context: Stage I: Absence of copula in learner sp eech. Stage II: Selection of ser to perform most copula functions. Stage III: Appearance of estar with progressive. Stage IV: Appearance of estar with locatives. Stage V: Appearance of estar with adjectives of condition. ( 1995) study targeted ser and estar with adjectives, where both verbs are often permissible depending upon whether or not the speaker intends a durative or a punctual aspect. The Spanish copulas ser and estar are redundant markers of aspect when they occur with adjectives, and the y have no inherent semantic value by themselves [ semantic value] [+ redundant]. Thus, the targeted form that Cheng examined ha d a low communicative value. Past PI studies (VanPatten & Cadierno 1993a, 1993b; Cadierno 1995) targeted morphosyntax, specific ally, direct object pronouns and word order in Spanish. Object pronouns have a high communicative value in Spanish because they

PAGE 82

68 have inherent semantic value (they convey meaning) and they lack redundancy [+semantic value] [ redundancy]. Cheng ( 1995) inv estigated whether L2 learners of Spanish would be able to link the use of the copular verbs with the meaning of specific ad jectives. Learners needed to recognize the aspect conveyed in each sentence or utterance (durative or punctual) in order to chose th e correct copula ( ser or estar ) in Spanish. A durative aspect is represented by ser For example, the sentence Mara es guapa or Mara is good looking or inherent by the speaker. However, the sentence Mara est guapa or Mara is looking good today looking than usual today (e.g. she is dressed up for a special occasion). In her study, Cheng ta rgeted the distinctions in meaning that occur when ser or estar is embedded in contexts with adjectives that express either punctual or durative aspects. Thus, h er study explored whether the benefits of PI extent to grammatical features other than morphos ytax. ( 1993) study consisted of 105 undergraduate students of Spanish in their se cond year of language study Similar to previous PI studies, Cheng administered a pretest and used a 60% cutoff score for participation in the st udy. Also similar to past PI studies, Cheng devised two instructional treatment packages, one for the PI group and one for the traditional instruction group. Participants completed the instructional treatment packages, which were balanced for vocabulary items, activity type, number of tokens, etc., over a period of two days. The control group received no

PAGE 83

69 instruction. Assessment tasks included an interpretation task, a sentence level production task, and a guided composition task, which Cheng included si nce she was interested in a feature of past PI studies, and it comprised a series of four drawings with key adjectives and vocabulary items beside them. Participants w ere asked to narrate a story based on the drawings, an activity that forced them to choose between the two Spanish copulas to correctly convey the meaning expressed in each drawing. The assessment tasks were given at three intervals: (a) as a pretest two weeks before the study took place, (b) as an immediate posttest after the instructional treatments were completed, and (c) as a delayed posttest three weeks after the treatments were completed. Cheng ( 1995) submitted the pre and posttest scores for the i nterpretation tasks to a repeated measures ANOVA, with typ e of instruction as the between subjects variable and time as the within subjects variable. The ANOVA revealed significant main effects for type of instruction and time. Post hoc tests revealed th e PI group performed significantly be tter on P osttest 1 than the control group. Interestingly, the post hoc tests also revealed that the PI group had a significant decrease in scores on Posttest 2 compared to Posttest 1. In addition, the PI group did not perform be tter than the control group on P osttest 2. However, the tradition al instruction group performed significantly better tha n the control group on Posttest 2 This was an unexpected finding given that past PI studies (VanPatten & Cadierno, 1993a; 1993b; Cadierno, 1995) did not find an effect for traditional instruction for sentence interpretation tasks.

PAGE 84

70 For the two prod uction tasks, both ANOVAs revealed significant main effects for type of instruction and time, and a significant interaction effect between instruction and time. For the sentence level production task, post hoc tests revealed that the PI group out performed the control group on Posttest 1. On P osttest 2 (the delayed posttest), both the PI group and the traditional instruction group p erformed significantly better than the control group. For the composition task, both the PI group and the traditional instruction group performed significantly better than the control group on both posttests. For the assessment tasks that measured produc ( 1995) study are consistent with other studies that examined PI (Cadierno, 1995; VanPatten & Cadierno 1993a, 1993b; VanPatten & Sanz, 1995). These results support the assertion that PI facilitates L2 production, even though pa rticipants never produced the targeted forms during PI lessons or activities. ( 1995) results regarding interpretation, however, do not support the findings of VanPa tten and Cadierno (1993a, 199b); namely, that PI facilitates interpretation of tar geted grammatical forms but that tr aditional instruction does not. In only participants in the traditional instruction group outperformed the control group on the interpretation task. Since the data did not support e xpectations she reanalyzed the data, only examining the test items that targeted the verb estar Cheng posited that beginning L2 learners of Spanish overgeneralize the use of ser early on in the acquisition process ( 1985, 1987) proposed acquisition order for the Spanish copulas ser and estar ) Thus, Cheng claimed that since beginning level learners of Spanish use ser as a default copula, estar is

PAGE 85

71 the problematic verb form for learners and only this data should be exa mined. When she reanalyzed the data, she found that for the interpretation task the PI group performed significantly better than the cont rol group on Posttest 1; however, on P osttest 2 there were no significant differences between the three groups (PI, tr aditional instruction, and control). For the sentence level production task, both the PI group and the traditional instruction group performed significantly better than t he control group. However, on P osttest 2 only the PI group performed better than the control group. For the composition task, both the PI group and the traditional instruction group performed better than the control group on P osttest 1. T here were no significant differences between the three groups on Postttest 2 The results of P ostte st 1 with the estar only data were identical to those of past PI studies. However, t he results of P osttest 2 for both the interpretation task and the composition task appe ar to indicate that the effects of instruction (for both PI and traditional instruct ion) were not retained three weeks after the instructional treatments took place as there were no statistically significant differences between groups on these meas ures on the delayed posttest. estar only data are question able as researchers typically do not discard half of their data set and reanalyze the remainder when their results do not support their a priori h ypotheses. By doing so, a bias on the part of the researcher in favor of PI is made apparent, which calls int o question the internal validity of the study. Further, based on the results of her study, Cheng ( 1995) claims that PI is helping students make correct form meaning m appings and in restructuring their mental

PAGE 86

72 by the results of either data set, ser and estar or estar only, as both PI and traditional instruction groups were similar i n their performance on all posttests according to the data that Cheng reported Of note, Cheng was a student of VanPatten at the time of her 1995 (dissertation) study. Subsequently, she published the results of her dissertation study in 2002 with a focus exclusively on the estar data. In her subsequent publication, Cheng ( 2002) did not make such strong assertions regarding the benefits PI. Although it is important to examine grammatical forms other than morphosyntactic features in order to determine if PI has an effect on these researchers that investigate PI need to exercise caution when making broad generalizations about its efficacy. More research is needed on PI, especially on features other than morphology and syntax. In addition, studies need to be conducted from a wider base of researchers in the field. Thus far, most PI studies have been conducted by VanPatten, his colleagues, and/or his students. In order to ensure that experimenter bias is not at play, researchers that are not connected to the developer of the instructional method and the theoretical model upon which it is founded need to investigate PI. Authentic versus flawed structured input activities Although s ubsequent research on PI (Benati, 2001; Cadierno, 1995; Cheng, 2002; Farley 2001a; Marsden, 2006; Qin, 2008; VanPatten & Sanz, 1995; VanPatten & Wong, 2004) has supported the claims made by VanPatten and Cadierno (1993a, 1993b) several researchers have refuted their findings. Among them are L. Allen (2000), Collentine (1998 b ), D eKeyser and Sokalski (1996), Erlam (2003), Nagata (1998), and Salaberry (1997). Dekeyser and

PAGE 87

73 Sokalski examined two grammatical forms, direct object pronouns and the conditional among first and secon d year undergraduate students of Spanish. They had three groups, an input processing group, an output processing group, and a control group that only received a ten minute grammar explanation. The researchers attempted to control for explicit information by providing all groups, including the control group, wi th the same grammar explanation. Rather than use the same referential and affective structured input activities as VanPatten and Cadierno (1993a 1993b) for the instructional treatments, DeKeyser and Sokalski developed their own structured input activitie s. For the assessment tasks, they measured text comprehension rather than interpretation (grammar comprehension), and the production tasks consisted of a fill in the blank task and a translation task. DeKeyser and Sokalski found that for direct object pr onouns in Spanish, the input processing group performed significantly better than the control group, but for both production tasks, only the output processing group performed significantly better than the control group. For the Spanish conditional, they f ound that only the output processing group outperformed the con trol group on the comprehension task, a result that is contrary to all past PI studies reviewed in this chapter. For both production tasks, they found that the input processing and output proc essing groups performed significantly better than the control group. DeKeyser and Sokalski claim that their findings refute those of VanPatten and Cadierno (1993a, 1993b). Further, they assert that L2 production and comprehension skills are learned separ ately. A major criticism of DeKeyser and Sokalski (1996) is that their treatment activities were not true structured input activities because they failed to follow the

PAGE 88

74 guidelines for the development of these activities set forth by J.F. Lee and VanPatten (1995, 2003). Guideline two specifies that meaning, or the communicative intent of sentences and utterances, should be the central focus of structured input activities. Thus, learners should not be able to complete structured input activities without comp rehending both the referential meaning of targeted grammatical forms and the propositional content of the input that they receive. In other words, all structured input activities are meaningful even though they are a type of focus on form instructional te chnique. Wong (2004) reviewed the structured input activities that were used in study and found that guideline two was violated because participants did not have to process targeted forms correctly in order to extract meaning from their input. Further, Wong (2004) claims that the structured input activities that were designed by DeKeyser and Sokalski were not PI because they required participants to focus on form and not on meaning in order to complete them. In addition, DeKeyser and Sokalski did not measure f targeted forms. Studies in the PI strand typically measure ation of targeted forms because a major g oal of PI is to help learners make correct form meaning connections T he only way to assess this process i s to examine whether participants are able to identify the referential meaning realized by a specific grammatical feature. Text comprehension, conv ersely, refers to whether learners understand the propositional content of an input text. Learners do not necessarily have to comprehend the meaning of targeted grammatical forms in order to understand the message of an input text ( R. Ellis, 1995). J.F. Lee and VanPatten ( 1995, 2003) claim that the two

PAGE 89

75 constructs (message c omprehension and gramm ar comprehension) overlap. However, there has been scant research to support this claim ( J.F. Lee & Rodr guez, 1997 ). More research studies are needed that examine both text comprehension and input processing. DeKeyser and Sokalski content of the input texts, which doe s not assess whether learners are able to make correct form meaning relationships (input processing). Therefore, their results are not comparable to studi forms (Benati, 2001 2004 a, 2004b 2005; Cadierno, 1995; Cheng, 1995, 2002; Collentine, 1998 a ; Farley, 2001a, 2001b, 2004; VanPatten & Cadierno 1993a, 1993b; VanPatten & Sanz, 1995; VanP atten & Wong, 2004). Other researchers (Erlam, 2003; Nagata, 1998; Salaberry, 1997) had similar findings to DeKeyser and Sokalski; however, these stud ies also measured text comprehension as opposed to interpretation of targeted forms, and their instructio nal treatments were considerably different from those in the PI strand of research since they contained flawed structured input activities. Thus, although L. Allen (2000), DeKeyser and Sokalski (1996), Erlam (2003), Nagata (1998), and Salaberry (1997) cla im that PI is ineffective based on the findings of their research studies, their results are not directly comparable to those in the PI strand of research Collentine (1998 b ) compared processing instruction with traditional (output based) grammar instru ction for the acquisition of the Spanish subjunctive in adjectival clauses when the referent is unknown. Collentine divided 54 s econd semester undergraduate students of Spanish who had no prior instruction on the subjunctive into three groups: a PI group, a traditional instruction group, and a control group. He

PAGE 90

76 operationalized traditional instruction with activities that moved from mechanical to open ended and that required the production of output. Col lentine developed an instrument that contained an in terpretation task and a production task to measure (PI and traditional instruction) performed significantly better than the control group on both types of task (inte rpretation and production), but that neither experimental group performed significantly better than the other. In other words, Collentine found that both PI and traditional (output based) instruction were beneficial for learning the subjunctive in adjecti val clauses in Spanish. His operationalization of PI and his structured input activities, however, were heavily criticized by Farley (2002) and VanPatten (2002) for not being authentic Farley (2002) criticized Collentine (1998 b ) for failing to provi de PI participants with information on processing strategies, wh ich is a key component of PI. Farley claimed that t he PI participants in did not receive any instruction on input processing strategies or how to overcome faulty input proc essing of the subjunctive. s structured input activities because they did not m odel of i nput p rocessing (1993, 1996, 2002, 2004). In addition, VanPatten (2002) also criticize structured input activities because they were with no prior experience with the targeted grammatical form. In order to determine if PI is superior to traditional instruction for the acquisition of the subjunctive in adjectival clauses in Spanish, future research studies would need to carefully follow the guidelines

PAGE 91

77 set forth by J.F. Lee and VanPatten (1995, 2003) for the operationalization of PI and for the creation of structured input activities. L. Allen (2000) also asserts that the findings of her study are contrary to VanPatten and Cadierno (1993a, 1993b), and she refutes the efficacy of PI. She investigated the French causative with the verb faire a structure that causes errors in interpretation and syntax for L2 learners of French whose L1 is English. With this structure, the object of the verb following faire must be placed postverbally, which usually occurs at the end of sentences or utterances and is marked by the pr eposition in French. L2 learners of French tend to interpret the first noun that they encounter as the subject of the verb following faire. Thus, the processing problem that learners encounter with the French causative is very similar to language learn object pronouns in Spanish (the targeted form in VanPatten & Cadierno, 1993a, 1993b), as learners incorrectly rely on the First Noun strategy in both instances. The participants in L. ( 2000) study included 179 four th semester high school students of French who had no prior exposure to the targeted grammatical form. She divided her participants into three groups as follows: (a) a group that received PI, (b) a group that received traditional instruction, and (c) a co ntrol group that received no instruction. She used an interpretation task that was similar to past PI studies and a sentence production task to measure participa L. Allen found that both PI and traditional instruction were sup erior to the control group for the interpretation task, but she found that the traditional instruction group was superior to the PI g roup for the sentence production task. She concluded that traditional instruction is

PAGE 92

78 more effective than PI, and that the results of VanPatten and Cadierno (1993a, 1993b) do not obtain with the French causative structure. The PI treatment materials that L. Allen (2000) developed were heavily criticized by VanPatten and Wong (2004) for the following reasons: (a) n o processing problem or strategy was identified for the participants, (b) t he structured input activities did not require participants to interpret the targeted grammatical form correctly in order to extract meaning, (c) d uring the explicit instruction phase of the st udy, the traditional instruction group received practice activities that required them to process strings of input containing the targeted form while the PI group never received any practice producing the targeted form during their instructional treatment, and (d) e vent probabilities helped participants select the correct answers to the structured input activities A major problem identified by VanPatten and Wong was L. of the structured input act ivities that she employed in her study. These activities were flawed because participants could select the correct answer simply by matching the names in the answer choices to the names in the input sentences. For example, L. Allen (2000) provide d a numb er of questions similar to the following : 1. Tom fait faire les valises Marc. a. Tom packs the bag b. Tom gets Marc to pack the bags (p. 83) In the previous example, the correct answer to each item is evident because each input sentence mention s two people; however, one response mentions two people and the other response mentions only one person. Thus, participants could choose correct responses simply by matching the number of people in the response to the number of people in the

PAGE 93

79 input sentenc e. Authentic structured input activities would require participants to interpret the French causative correctly in order to select the correct response. The following is an example of how L. ( 2000) instructional treatment activity could have been transformed into an authentic structured input activity for the French causative: 1. Jean fait laver la voiture Marc. a. Marc gets Jean to wash the car. b. Jean gets Marc to wash the car. In the previous example, French language learners would be for ced to attend to the agency of the verb following faire in order to interpret the sentence correctly. Thus, with a slight modification, L. ( 2000) treatment activity could have been converted into an authentic structured input activity. VanPatten and Wong (2004) replicated L. Allen (2000) ; however, they adjusted L. The participants in VanPatten and Wong (2004) comprised 77 fourth semester French students from two universities who had no prior instruction on the French causative. Participants were divided into three groups, a PI group, a traditional instruction group, and a control group that received no instruction. Except for the corrections made to the flawed structured input activities and assessment task items, VanPatten and Wong attempted to replicate L. Allen (2000) as closely as possible. (2004) study were dissimilar to L. ( 2000) For the interp retation task, the PI group outperformed the traditional instruction group, and the t raditional instruction group outperformed the control group. On the production task, the PI group performed equally as well as the traditional instruction

PAGE 94

80 group, and bot h of these treatment groups outperformed the control group These results more closely resemble those of VanPatten and Cadierno (1993a, 1993b) and support activities must be closely adhered to if researchers intend to compare PI to ot her instructional methods. When flawed structured input activities are employed researchers are not examining authentic PI; rather they are only comparing input and output based instruction. Therefore, any claims that researchers mak e regarding the efficacy (or lack thereof) of PI are not supportable when the instructional treatment materials contain flawed structured input activit ies Processing instruction and meaningful output based instruction A problem that occurs when comparing PI with tradi tional instruction is the all of the tasks and activities in PI are meaningful, while only some of the tasks and activities in traditional instruction are meaningful. For example, language learners can answer the questions in the mechanical drill activiti es of traditional instruction without comprehending their simultaneously. In order to address the question of whether the differing amounts of attention to meaning in the treatment materials (traditional instruction versus PI) are responsible for the finding that PI is superior to traditional instruction, Farley (2001a) compared PI with output based instruction that was completely meaningful. In other words, the output activities where participants could focus only and form and still supply the correct answer. Farley also matched output based instruction to PI on the explicit explanation

PAGE 95

81 component by providing information on processing strategies and grammar instruction that was nonparadigmatic to the participants who produced output. He named this type of instruction meaning based output instruction (MOI). The only difference between PI and MOI was the type of practice mode, with PI using input based practice activities and MOI using output based practice activities. The participants 29 undergraduate students of Spanish in their fourth semester of language study The targeted form was the Spanish subjunctive in nominal clauses after expressions of doubt. Participants were assigned to either the PI group or the MOI group (there was no control group). As with past PI stud included an interpreta tion task and a sentence level production task. Although Farley controlled for the meaningfulness of the treatment activities, he still found that PI was superior to MOI for the interpretation task, and that PI was equal to MOI for the production task. T hese results mirror the findings of other PI studies that compared PI to traditional instruction (Benati, 2001, 2005; Cadierno, 1995; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004) It appears that the relative difference in meaningful activiti es between PI and traditional instruction is not responsible for the finding that PI is superior to traditional instruction. participants. Benati (2005) compared PI, MOI, and traditional instruction with 77 secondary students of English as a s econd l anguage (ESL) His targeted form was the simple past tense in English, and his assessments included an interpretation task and a sentence level production task. Benati found that PI was superior to both MOI and traditional

PAGE 96

82 instruction for the interpr etation task, and that there were no significant difference s between PI, MOI, and traditional instruction on the s entence level production task. Farley (2001a) also compared PI wi Farley (2001b) replicated Farley (2001a) with the sam e targeted form and with similar instructional treatments and assessment tasks. Farley (2001b) had a larger sample of participants (50 fourth semester undergraduate students of Spanish as opposed to 29), and there were ten instructional activities in each treatment packet rather than eight. Interestingly, his results differed from Farley (2001a). He found that there was no significant d ifference between PI and MOI on either the interpretation task or the sentence level production task. Farley (2001b) spe culated that the different results between his two studies might have been due to the participants having more practice in Farley (2001 b), since they had ten rather than eight instructional activities. Also, the feedback that was given to the MOI groups in both studies resulted in incidental input for the MOI participants because the teacher solicited answers to the treatment activities until the correct answer was given. Therefore, by having ten activities rather than eight, participants in the MOI g roup in Farley (2001b) likely received more incidental input of the targeted form than participants in Farley (2001a). Another factor that may have accelerated track of Spanish language studies, while the participants in Farley (2001a) were on a regular track. However, the greater language ability of the participants in Farley (2001b) should not have affected the two treatment groups differentially (with the

PAGE 97

83 MOI group benefitin g more than the PI group). It appears that the differential amount of ccounted fo r the different results, as MOI participants in Farley (2001b) likely received more inciden tal input than MOI participants in Farley (2001a ) studies that compare d PI with output based instruction (Benati, 2001, 2005; Cadierno, 1995; Farley, 2001a; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004) is the targeted form that he examined. Farley (2004) speculated that PI was not as beneficial for the subjun ctive due to its linguistic complexity. He examined the subjunctive in nominal clauses following expressions of doubt, uncertainty, or denial. In natural speech, this form is always redundant and has little inherent semantic value. Farley manipulated PI d have to match matrix clauses that contained expressions of doubt and denial with subordinate clauses that contained subjunctive forms. It was not possible for the PI participants to rely on the subjunctive form alon e to extract meaning from sentences as these forms rely on matrix clause verbs to express doubt or denial. For example, in the sentence Dudamos que los estudiantes hagan la tarea todas las noches which is rendered We doubt that the students do their homew ork every night the main clause verb is necessary to express doubt because the subordinate clause verb carries no semantic meaning when stripped of the matrix clause. The subordinate clause hagan la tarea contains the subjunctive verb form hagan but thi s form has a low communicative value and does not lend itself well to PI. Subprinciples

PAGE 98

84 m odel of i nput p rocessing (1993, 1996, 2002, 2004) state that meaningful forms that are not redundant [+semantic value] [ redundancy] will be processed before nonmeaningful forms that are redundant [ semantic value] [+redundancy] and that meaningful forms are processed before nonmeaningful forms regardless of redundancy. Pereira (1996) and Woodson (1997) also examined PI with the subjunctive in nominal clauses and found results similar to Farley (2001b). Since the structured input activities in PI are intended to elevate form s that are meaningless may not benefit from PI. The only study that examined a subjunctive form that was meaningful was Collentine (1998 b ), and his study was criticized for not using authentic structured input activities and for not providing PI participa nts with information on processing strategies. Therefore, more research studies are needed that examine meaningful subjunctive forms and that maintain stricter treatment fideli ty to PI, especially with respect to the development of structured input activi ties. Similar to Farley (2001a, 2001b) Morgan Short and Wood Bowden (2006) compared PI and MOI to a control group that received instruction on the targeted form and spent an equal amount of time on task as the two experimental groups for the acquisition o f object pronouns in Spanish. Their sample consisted of 45 first semester u ndergraduate students of Spanish The researchers attempted to control for all variables other than p ractice mode by utilizing computer based instructional treatments They opera tionalized PI by adapting the paper and pencil instructional treatment packet that was used in VanPatten and Cadierno (1993a, 199b) for computer based delivery using

PAGE 99

85 Authorware 5. The MOI and PI groups were both given referential and affective activities; however, only the MOI group was required to produce output while completing these activities. The control group was given reading activities containing the targeted form followed by comprehension questions on the reading passages. All of the treatment m aterials contained the same number of targeted forms, and time spent on task was equal for all three groups. Morgan Short and Wood Bowden found that there was no significant difference between PI and MOI on the interpretation test (both groups outperforme d the control group); however, only the MOI group performed significantly better than the control group on the production task. These results are contradictory to Farley (2001a) and Benati (2005), but they are similar to Farley (2001b). With classroom based s tudies such as Farley (2001a, 2001b ) and Benati (2005), when participants in the MOI groups answered questions orally, their output served as incidental input for their classmates. Similarly, when instructors checked answers to activities with MOI participants aloud in class, more incidental input was provided to these participants. Thus, when an MOI participant made a mistake on an activity item, other MOI participants would hear the targeted forms as input when the correct answer was provided by the teac her or by another student. Consequently, t he output that was produced by MOI participants and the feedback that was give n to these students during the classroom based instructional treatments may have provided a significant amount of incidental i nput for all of the MOI par ticipants in the aforementioned studies It is important t o note that for the studies that compared PI with MOI (Benati, 2005; Farley, 2001a, 2001b), MOI participants were only intended to receive output based practice.

PAGE 100

86 Morgan S hort and Wood Bowden (2006) asserted that their computer based study was superior to the past classroom based studies because feedback was delivered only to individual participants who made mistakes rather than to the entire treatment group. However, MOI participants who made mistakes still received feedback that contained the targeted grammatical form (hence the MOI participants in their study still received incidental input containing the targeted forms). In addition, as Morgan Short and Wood Bowden did were unable to report the number of MOI participants that received incidental input in the form of feedback, or the number of ti mes that participants received feedback (incidental input) during their instructional treatment s It appears that the amount and type of feedback given to participants was an extraneous variable that may have exerted some influence on the findings in Morgan the M OI groups in the aforementioned studies had an advantage over the PI groups because the MOI participants received incidental input containing the targeted forms during their instructional treatments while the PI groups did not receive any output based prac tice. Processing instruction and type of feedback Sanz (2004) asserts that feedback has largely been an uncontrolled variable in PI research, which is also a shortcoming in other research studies that have examined the effects of explicit instruction in SLA Spada & Lightbown, 1993; L. White, 1991). Sanz (2004) examined the effects of explicit and implicit feedback with PI on the acquisition of object pronouns in Spanish with 28 firs t or second year undergraduate students of Spanish. Although participants varied in

PAGE 101

87 their level of language study, all of them scored below 60% on the pretest. Participants in the explicit feedback group received feedback that was immediate and individualized (it was only provided to participants who made mistakes). Explicit feedback was operationalized as feedback that contained explicit information on the nature of the error (e.g. incorrect strategy use) and information on TL rules. Indi vidualized feedback was possible since the instructional treatments were designed for computer based delivery. response was incorrect. Thus, participants in the implicit condition were provided with feedback that informed them whether their answers were correct or incorrect. Interestingly, Sanz did not find any significant differences between the explicit and the implicit feedback groups and both groups pe rformed significantly better across time on interpret ation and production tasks. Sanz posits that the structured input activities provided to the participants in both groups prompted them to make correct form m eaning connections, which she suggests is a more important component of PI than type of feedback offered. The results of this study imply that future PI studies only need to provide participant s with implicit feedback as explicit feedback did not result in improved pe rformance on interpretation and production tasks study is that the sample size was small (only 28 participants) Future research should attempt to examine feedback type with a larger number of participants. Another cri ents of PI (explicit explanation of grammar, information on processing strategies, and

PAGE 102

88 across time was d ue to the structured input component is not supported by her findings. Summary of Findings and Implications for Future Research A review of the research reveals that most PI studies have been undertaken within a small research community that is limited to VanPatten, his students, and his colleagues. Researchers that are not a part of this community ( L. Allen, 2000; Collentine, 1998 b ; DeKeyser & Sokalski 1996; Erlam, 2003; Nagata, 1998; Salaberry, 1997) have been criticized for not maintaining treatment fi delity to PI and for using flawed structured input activities. Further, several studies (DeKeyser & Sokalski, 1996; Erlam, 2003; Nagata, 1998; Salaberry, 1997) also used assessment tasks that did not measure matical forms; rather, these studies which is a different construct. Therefore, due the aforementioned limitations, most research studies that are independent from VanPat ten and his colleagues are not directly comparable to the PI strand of research. More research studies are needed on PI from a wider base of researchers, but special attention needs to be paid to the development of structured input activities and to the t ype of assessment tasks that are used. The present study took these points into consideration in designing all structured input activities and assessment tasks. In addition, the researcher received permission to use two structured input activities for th e subjunctive in adjectival clauses that were developed by Farley (2002), who is a former student and a current colleague of VanPatten. All other structured input activities that were employed in the present study were based on the

PAGE 103

89 format that was provide d by Farley. Regarding the assessment tasks, past studies only measured either interpretation or comprehension; however, the present study examined how the instructional treatments may have affected both constructs, and a correlational analysis was perfor med to determine if there was a relationship between the two. Despite a relatively small research base, the results of studies that have operationalized PI and structured input activities appropriately and that have examined targeted grammatical forms tha t are meaningful (Benati, 2001, 2005; Cadierno, 1995; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004) have yielded notable results regarding the efficacy of PI when it is compared to traditional and other types of output based instruction. Nam ely, the aforementioned studies have found that PI is is remarkabl e given that PI participants in the aforementioned studies never produced any targeted forms during their instructional treatments. Studies that isolated the components of PI have yielded more mixed results; however, every study that examined an explici t grammar explanation alone (Benati, 2004a, 2001; Sanz & Morgan Short, 2004; VanPatten & Oikkenon, 1996; Wong, 2004) found that it was not beneficial for either production or interpret ation tasks, and participants that only received an explicit explanation of grammar did not perform any differently than participants in control groups. VanPatten and Oikkenon claim that structured input is responsible for the beneficial effect of PI, and the results of four subsequent studies support this claim (Benati, 2004 a, 2004b; Sanz & Morgan Short,

PAGE 104

90 2004; Wong, 2004). However, two studies that compared the effects of PI and structured input activities on the acquisition of the subjunctive in nominal clauses following expressions of doubt (Farley, 2004 b ; Fernandez, 2008) found that the explicit explanation component is beneficial when it is combined with structured input activities. In other words, Farley (2004 b ) and Fernandez (2008) found that full PI is superior to structured input activities alone when the targeted fo rm is complex. The targeted form in their studies, however, had a low communicative value, which may have exerted some influence on the results. The present study examined a subjunctive form with a high communicative value to determine if structured inpu t activities alone are as effective as PI when the targeted form is more amenable to PI [ redundancy] [+semantic value]. Given that all studies that examined the explicit grammar explanation component of PI in isolation did not find a beneficial effect fo r it, and that if included it would be impossible to equalize the treatments for feedback and time on task, the explicit grammar explanation component of PI was not examined in isolation in the present study. A review of the studies that compared PI to me aningful output based instruction (Benati, 2005; Farley, 2001a, 2001b) revealed that the feedback that was provided to the output based groups and to the PI groups was not equal, with the output based groups receiving the targeted form as incidental input when teachers or classmates provided correct answers to the treatment activities. Even the meaningful output based study that was computer based (Morgan Short & Wood Bowden, 2006) provided the correct targeted forms to participants following their answers to oral activities. The presen t study only provided implicit feedback to all participants In other words, participants in all

PAGE 105

91 treatment groups were only told if their answers were correct or incorrect. A problem, however, did arise with feedback for t he traditional instruction activity that required oral output. In order not to provide incidental input to these participants, the feedback that participants in the traditional instruction group received was delayed rather than immediate for a single acti vit y that consisted of five items. For the single oral output activity, participants made a voice recording of their answers using an audio drop box that stored their r ecordings on an external site. The researcher accessed the site, listened to the recor dings, and sent participants an email message stating whether their answers were correct or incorrect. The researcher made every effort to provide feedback to participants on the same day that they completed the oral activity. Due to the prohibitive cost s of designing and implementing voice recognition software for the present study, supplying delayed feedback on a single oral output activity was the only way to avoid providing the targeted form as incidental input to the traditional instruction participa nts and to equalize the type of fe edback that was given to all treatment groups. Further, in an effort to help the participants in the traditional instruction group receive some type of immediate feedback for the five items on the oral output activity, th ey were asked to reflect upon their own answers and check true if they believed that an oral response was correct and false if they felt that an oral response was incorrect. Thus far, no study found on PI has examined how participants react to authentic i nput containing the targeted forms following their instructional treatments. All of the PI studies that were reviewed in this chapter used structured input, or input that was manipulated to elevate the communicative value of the targeted form, in their tr eatment

PAGE 106

92 nominal clauses (2001a, 2001b), he placed the subjunctive forms in the sentence initial position in his activities in order to circumvent the Sentence Location Principle, which states that items in the sentence medial position (where the subjunctive is normally located) are processed last. This may have helped learners perceive the subjunctive forms during treatment activities, but it is unlikely that learners w ill ever encounter the subjunctive in nominal clauses following expressions of doubt in the sentence initial position in authentic input. Collentine (2004) called for research studies that examine how participants respond to authentic input once the PI t reatment has concluded. He states, [W] e do not know if learners respond to forms constituting the targeted grammatical phenomenon in normal input conditions once they have left whether l earners processing mechanisms remain altered as a result of the processing instruction intervention; delayed posttests do not reveal authentic input. This should be a key challenge for researchers in the future (Collentine, 2004, p. 179) (2004) call and to extend the scope of PI research, the present study examined whether exposure to any of the instructional treatments (processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structured input without visual input enhancement, and traditional instruction) altered the way learners noticed and processe d targeted forms that were embedded in an authentic

PAGE 107

93 input passage that participants received subsequent to completing their instructional treatments. Finally, the present study investigated PI with distance Spanish language learners, which is a differen t population of students from past studies (all past PI studies used classroom based learners). The studies that were computer based (Morgan Short & Wood Bowden 2006 ; Sanz & Morgan Short, 2004; Sanz, 2004) examined clas sroom based FL learners in a comput er lab. Since PI and structured input activities are input based instructional techniques, they are suitable for online language learning where teachers have greater control over the linguistic input that students receive. Also, with distance language lea rning, students have fewer opportunities to produce the TL and to interact with their teacher and/or their peers in the TL. By examining distance language targeted forms despite the drawback of having no interaction with or feedback from a teacher during their instructional treatments. Input Enhancement Another the foci of the present study was te xtual/visual input enhancement. Past studies that examined visual inp ut enhancement (VIE) have typically operationalized VIE as typographical enhancements, which are achieved through formatting techniques such as bolding, highlighting, capitalizing, and/or changing the font style or size. The literature on the efficacy of VIE has been largely mixed, with some studies demonstrating a positive effect for VIE (Doughty, 1988, 1991; Jordenais, Ota, Stauffer, Boyson & Doughty, 1995; Shook, 1994; Williams, 1998; Wong 2002), some finding only a minimal

PAGE 108

94 effect (Alanen, 1995; Izumi, 2002; Robinson, 1997; J. White, 1998), and still others demonstrating no beneficial effect for VIE (Leow, 1997; 2001; Leow et al ., 2003 ; Jo u rdenais, 1998; Overstreet, 1998, Wong, 2003) Empirical Studies on Input Enhancement design study found that VIE was facilitative for learning present perfect forms in Spanish. He investigated the benefit of VIE with present perfect and relative pronoun forms among 125 first and second year u ndergraduate y had two experimental groups and one con trol group as follows: (1) an experimental group that received input texts with VIE with explicit instructions to pay attenti on to enhanced forms, (2) an experimental group that received input texts with VIE without any such ins tructions, and (3) a control group that received input texts without V IE. Shook found that there was no significant difference between the experimental group that was instructed to pay attention to the enhanced forms and the experimental grou p that was not instructed to do so. Thus, telling the learners what to pay attention to did not affect the learning outcome measures in this study. Shook did find, however, that the experimental groups that received VIE performed significantly better tha n the control group on production and recognition tests. He also found that participants did not perform as well on tests that measured relativ e pronoun usage compared to tests that measured present perfect usage in Spanish. The researcher suggests that the present perfect forms were easier for students to notice because they have a higher communicative value than the relative pronoun forms. In other words, the referential meaning of the present perfect forms (temporal reference)

PAGE 109

95 assisted lear ners with t ext comprehension. Conversely, t he r elative pronoun forms did not need to be processed by learners because they did not contribute to the overall meaning of the input texts. Thus, relative pronouns (a grammatical feature with a low communicative value) d id not benefit from the presence of VIE, but present perfect forms (a grammatical feature with a higher communicative value) did benefit from the presence even when combined with simplified input, for the acquisition of a form with a low communicative value (past participle agreement in relative clauses in French). 94) study had some l imitations, he did not attempt to measure noticing of target ed forms; rather, he only measured acquisition of targeted forms through production and recognition tests. Another is that participants were only exposed to input materials containing VIE for a very short period of time (under one hour). Wong (2002) found that VIE was beneficial in her investigation of input enhancement w ith sentential versus discourse level input. T he researcher theorized that beginning level foreign language learners would benefit more from VIE that was embedded in shorter sentence level input passages. Previous studies that examined VIE tended to embed input enhancement in longer discourse level p assages, which may have caused comprehension difficulties for novice language learners. The targeted form was preposition usage in French. Her participants comprised beginning level undergraduate students of French. P articipants were divided into four g roups. Two groups received

PAGE 110

96 sentence level input and two groups received discourse level input (one group in each level received input containing VIE). ( 2002) found that the two groups that received VIE outperformed the two groups that did not on the post treatment assessment tasks. However, she also found that the groups that received sentential level input performe d better than all other groups (whether they received input containing VIE or not ) Thus, VIE was found to be beneficial for the acqu isition of prepositions wi th beginning level French language learners, but not as beneficial as receiving sentence level input. Some drawbacks of this study were that Wong did not measure any noticing that may have resulted from the presence of VIE, and l ike Shook (1994), her participants were only exposed to visually enhanced texts for under one hour. Even though the main purpose of VIE is to targeted grammatical forms, only four studies have attempted to measure the noticing of target ed forms that takes place as a result of the presence of VIE (Alanen, 1995; Izumi, 2002; Leow, 2001; Leow et al. 2003). The majority of studies conducted thus far have typically only focused on the acquisition of target ed forms. Perhaps this is because noticing is a very difficult construct to operationalize and measure in SLA research. Two studies, Leow (2001) and Leow et al. (2003) exa mined the effect that VIE had on both noticing and comprehension of target ed forms. Interesti ngly, these st udies found that the presence of VIE did not have a beneficial effect on either noticing or compre hension In Leow (2001), the targeted linguistic forms were the present perfect and the present subjunctive in Spanish. The 72 partici pants were beginning level undergraduate

PAGE 111

97 students of Spanish. P articipants were divided into two groups (one experimental group and one control group). The experimental group received input texts with VIE and the control group received the same input texts without VIE. Leow measured noticing through the use of think aloud protocols. His results revealed that there was no significant difference in the amount of noticing that took place between the two groups as measured by the think aloud protocols. Thus, he concluded that VIE does not have a facilitative effect on noticing. However, think noticing in real time, have their limitations. It has been recognized that thinking aloud while trying to complete a task may interfere with ta sk completion and language processing (Izumi, 2002, Johnson, 2001). Of note, Leow only examined the amount of noticing (the number of instances that the target forms were mentioned while participants thought aloud) and not the depth of noticing by assessi Similarly, his comprehension tests did not reveal any significant differences between the experimental and control groups. Thus, Leow concluded that VIE does not increase noticing of target forms or assist learners wi th text comprehension. TL input. The researcher targeted the preterit and imperfect forms in Spanish, and his participants incl uded 50 intermedia te level undergraduate students of Spanish. He investigated VIE in combination with texts that follows: (a) VIE with a familiar text, (b) VIE with an unfamiliar text, ( c) n o enhancements with a familiar text, and (d) n o enhancements with an unfamiliar text. On

PAGE 112

98 the posttest assessment measures for the preterit and imperfect forms, Overstreet found no significant differences between the groups that received VIE and the gr oups that did not. Further, text familiarity did not appear to facilitate the acquisiti on of target ed forms (as measured by the posttests). Surprisingly, Overstreet found that the two groups that received VIE performed significantly less well on comprehe nsion tests than the two groups that did not receive VIE. The researcher suggests that the presence of VIE prompted the learners to focus on form at the expense of meaning. Recent research by S. Lee (2007) on the acquisition of the passiv e voic e in Engli sh by Korean ESL students supports comprehension. S. their ability to comprehend meaning was negatively affected. targeted grammatical forms. The preterit and imperfect forms were completely new to Spanish is fairly complex because verbs in the past tense often change meaning depending upon the context of the surrounding sentences or utterances. (1998) study include the following: (a) the researcher did presence of VIE and (b) participants were only exposed to input materials for under one relat ively short (210 words), the participants in S. ( 2007) study were exposed to a lengthy passage containing over 1,200 words. Thus far, Overstreet (1998) and S. Lee

PAGE 113

99 (2007) are the only researchers to have found that VIE has a negative on text comprehe nsion It appears that text length is not responsible for this finding. Very few studies have examined VIE in combination with another pedagogical device. Alanen (1995) examined VIE in combination with the provision of metalinguistic rules. Her study i nvestigated VIE with 36 students who were learning Finnish as a foreign language and whose L1 was English. The researcher examined language that was based on Finnish. Ala nen divided her participants into four groups as follows: (a) a group that was given the metalinguistic rules for the morphemes followed by input texts with VIE, (b) a group that was given the metalinguistic rules for the morphemes without any input texts, (c) a group that was given input texts with VIE without any metalinguistic rules, and (d) a control group that was given input texts without VIE and without any metalinguistic rules. Noticing was measured by think aloud protocols and production was measu red by production tasks. Alanen (1995) found that participants who received input texts with VIE had significantly more noticing of the inflectional morphemes than those who did not r eceive VIE Also, the groups that read input texts with VIE performed significantly better on production tasks than the control group that read input texts without VIE. However, the group that received only the metalinguistic rules performed significantly better on production tasks than the group that only rece ived texts wi th VIE Thus, VIE was more facilitative for increasing noticing of targeted forms than the provision of metalinguistic rules. However, metalinguistic rules were more beneficial than VIE when participants

PAGE 114

100 were required to use the tar geted forms for product ion. some limitations. The number of participants was relatively small (36), and they were only exposed to the treatment materials for less than one hour. J. White (1998) had d that VIE alone or VIE followed by input flooding is not sufficient for learners to acquire targeted forms (possessive determiners in English by L1 speakers of French). J. White also asserts that VIE is effective for increasing noticing, but once targete d forms are noticed, learners are uncertain about their relevance. This would indicate that some other pedagogical technique in addition to VIE would be needed for learners to acquire the targeted grammatical forms. In an attempt to determine if VIE has a beneficial effect on FL grammar learning, S. Lee and Huang (2008) performed a metaanalysis on twelve published studies (Alanen, 1995; Doughty, 1991; Izumi, 2002; Jourdenais et al 1995; S. Lee, 2007; Leow, 1997; Leow, 2001; Leow et al. 2003 ; Overstreet 1998; Shook, 1994; J. White, 1998; Wong, 2003) and four unpublished studies (Ha, 2005; Jourdenais, 1998; Kubota, 2000; Overstreet, 2002). The following criteria were employed for inclusion in the metaanalysis: (a) a study had to have an experimental or a quasi experimental design with participants who were L2 or FL learners, (b) a study had to examine the effects of VIE on a posttest reading task, (c) a control or comparison group had to be included in (d) a study had to be published in a peer refereed journal or book chapter, or be an unpublished doctoral dissertation, (e ) a study had to report enough data (descriptive statistics) for the effect size to be computed, (f ) a study had to be written in

PAGE 115

101 English, and (g ) a study had to take place between 1981, when Sharwood Smith first proposed input enhancement as a pedagogical technique, and 2007. All of the studies included in the metaanalysis operationalized VIE with simple typographical enhancements. The most commonly used techniques were bolding, underling, or a combination of the two. One study, Alanen (1995), used italicization as the method of input enhancement and Doughty (1991) used color. By examining and combining the effect sizes of all of the studies included in the metaana lysis, S. Lee and Huang found that VIE had a very small positive effect on grammar learning, ( d = .22), and a small but negative effect for text comprehension ( d = 0.26). Of note, while all of the studies examined by S. Lee and Huang (2008) measured par mmar learning, only nine propositional content of the input passages that contained visual enhancements of the targeted forms. It is difficult to draw any definitive conclusions regarding a negativ e effect for VIE on text comprehension when not all of the studies in the metaanalysis examined this construct. S. Lee and Huang assert that the small effect sizes found in their analysis reflects the conflicting results that the aforementioned studies re ported regarding the benefits of VIE for grammar learning. They recommend that more research needs to be conducted on VIE from a wider base of researchers in the field before any definitive Co mputer Based Visual Input Enhancement The main goal of VIE is to increase the visual salience of targeted grammatical forms in order to increase the likelihood that learners will notice them. Thus, the purpose

PAGE 116

102 of employing visual input enhancement techni grammatical forms that are present in their TL input. As the present study will examine VIE with distance FL learners, it is possible that simple typographical enhancements are ention in web based multimedia environments. Since M. Allen (2003) claims that learners ignore stimuli that they perceive as uninteresting in computer based media, the present study proposes to operationalize VIE with word animation, or the animation of t argeted verb forms through movement and the selective use of color. Ri e attention in computer 77). Since motion is an attention drawing device in computer based media, it is an ideal candidate for VIE in the present study. Neurologically, the motion perception system is powerful and less susceptible to disruption than higher cognitive domains such as language, attention, and memory (Jagaroo & Wilkinson, 2008). Further, the Stimulus Movement Effect (Nealis, Harlow, & Suomi, 1977) states that the perceptual system automatically directs attention to motion changes due to a built in bias. Thus V IE that utilizes motion should be a powerful properties or position, and so forth a study on flash animation (Hong, Thong, & Tam, 2004) found that flash attracts displays that are tightly pack ed. The researchers caution, however, that they found no

PAGE 117

103 flashing may decrease the recall of other static items on the screen. To overcome this potential negative effe ct for flash and other types of animation, Sutcliffe and Namoune (2007) suggest that animations should be used sequentially rather than concurrently. Further, they posit that a sequential presentation of animation is preferable as concurrent animations co Thus far, no SLA studies found have attempted to operationalize VIE through word animation. However, research in the area of instructional design supports the use of animation to successfully attr based media (Baek & Layne, 1988; Park & Hopkins, 1993; Rieber 1990). Collentine (1998a), a prominent SLA researcher, advocates the use of structured input and other input enhancement techniques in computer assisted lan guage learning (CALL) tasks, which he claims are particularly effective in web based environments because targeted structures can be made Further, learners are more prone to attend to targeted structures if they have stimulus novelty, which according to Cowan (1995) can be achieved any number of ways with multimedia tools such as graphics, sound, video, and animations. In the present study, VIE through animation of targeted verb forms will provide stimulus novelty, and it will present learners with two layers of information. Lehrer (1993) asserts that computer based tools are superior to text based tools for learning because computers are able to provide learners with multiple layers of data at one time. However, designers of CALL applications and web based materials need to be careful not to overwhelm learners with

PAGE 118

104 too many stimuli at once in order not to overload their processing capabilities. Effective applications employ the prin ciple of selective fidelity, which posits that the only stimuli that should be provided to learners are those that will assist them in forming hypotheses about new knowledge structures, or those that will assist them with modifying hypotheses about existin g knowledge structures (Andrews, Caroll, & Bell, 1995). In the present study, the VIE treatment groups will receive input that combines an image layer (the animation) with a text layer, and it is expected that these two layers of information will not over designers need to constantly search for new techniques to improve different areas of learning. The present study plans does so by updating VIE for multimedia and web based lea rning environments. Summary of Findings and Implications for Future Research It is presently unclear whether VIE is able to facilitate noticing, acquisition, or both as only a limited number of studies have investigated VIE. A careful review of the relev ant literature yielded very mixed results, and two studies ( S. Lee, 2007; Overstreet, indicates that VIE i s more beneficial with sentence level input (Wong, 2002) and m ay be form specific (Shook, 1994), with grammatical forms with a low communicative value receiving little benefit from the presence of VIE. Additionally, VIE may need to be combined with other input enhancement techniques in order for targeted forms to be acquired. More research is needed on VIE, especially studies that measure both noticing and acquisition of targeted forms and studies that combine VIE with other pedagogical

PAGE 119

105 techniques. The present study took these points into consideration and used sen tential level input in all of the treatment materials. Also, the targeted grammatical form is the only subjunctive form with a high communicative value, which should have increased the facilitative effect of VIE. Further, in the present study VIE was com bined with structured input activities to increase the likelihood that participants would make correct form meaning mappings once the targeted forms were noticed. From the low effect size found by S. Lee and Huang (2008), it appears the typographical enha ncements have little effect on grammar learning The present study updated VIE for the web by animating targeted animated subjunctive forms grew larger and smaller ove r a period of seven seconds, after which time they reverted back to the size of the other words in the input sentences. In addition, the animated words also employed the selective use of color to draw attention to subjunctive verb endings, which tend to e ( J.F. Lee, 1987; J.F. Lee & Rodrguez, 199 7) J. White (1998) suggests that VIE should be combined with another instructional technique because she claims that VIE only facilitates n oticing and not learning of study support this claim. The purpose of VIE in the present study was to attract the learners to notice due to its placement in the medial position of sentences. Once noticed, the structured input activities in which the animations were embedded were designed to assist learners with correct input processing. The research reviewed in this section indicates that moveme based

PAGE 120

106 media (Baek & Layne, 1988; Jagaroo & Wilkinson, 2008; Nealis, Harlow, & Suomi, 1977; Park & Hopkins, 1993; Rieber 1990), but flash animation and movement can detract attention from sta tic items on the screen and potentially overwhelm learners if the animations are concurrent (Sutcliffe & Namoune, 2007). Therefore, the present study delivered animations sequentially to avoid these potential negative effects. Output and Language Learning The present study examined traditional inst ruction as a comparison group. Under the traditional instruction paradigm, a heavy emphasis is placed on output practice in the TL. A key difference between the experimental groups (processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, and structured input without visual input enhancement) and the comparison group (traditional instruction) in the present stu dy is the type of instruction that was delivered. The experimental groups received instruction that was input based while the comparison group received instruction that was output based. Past studies that compared PI with traditional instruction as it wa s operationalized in the present study (Benati, 2001; Cadierno, 1995, Cheng, 1995, 2002; VanPatten & Cadierno, 1993a, 1993b, VanPatten & Wong, 2004) found that PI was superior to traditional instr uction for interpretation tasks and that PI was equal to tra ditional instruction for production tasks. VanPatten (2004) claims that the superiority of PI over traditional instruction is due, in part, to the nature of input and output processing in SLA. A prerequisite for PI is the provision of comprehensible inpu t (Krashen, 1982) to the he

PAGE 121

107 asserts that input is the single most important factor for SLA, with all theories of SLA relying on input, in some way, to explain acqu isition. processing ( 1993, 1996, 2002, 2004 ) posits that language acquisition occurs when learners take in and store pairs of form meaning relationships, and structured input activities are desi gned to help learners with this pr ocess. Further, VanPatten (1993, 1996, 2002, 2004) claims linguistic systems if accommodation and restructuring occur. Conversely, when learners produce output, they are only retrievin g information that is already a part of their implicit linguistic systems. VanPatten (2004) also asserts that output is not a direct path to acquisition and that the main role for the production of output is to develop fluency and accuracy in the L2. Swa in (1985, 1993, 1995, 1998), however, asserts that the production of output input, although necessary for SLA to take place, is not sufficient for learne rs to fully de velop native like L2 proficiency Swain (1985) found that long term French immersion students in Canada were able to develop high levels of listening and reading comprehension, but they failed to attain native like production in speaking and writing skill s, even after many years of instruction in the L2. Swain attributed these findings to the nature of the immersion education classes that the students received. Immersion students were exposed to large amounts of comprehensible input during subject matter instruction in the L2; however, they were not required to produce much linguistic output during their French immersion classes. In addition, Swain found that teachers tended not

PAGE 122

108 sage across in the L2. Based on her findings, Swain (1985) asserts that teachers need to push L2 learners to produce TL output (in speech and in writing) in order to assist them in developing grammatical accuracy. She claims that by producing output, le arners are forced to shift to a deeper level of language processing (syntactic rather than semantic), which does not Hypothesis states that the act of producing language (in spe ech or writing), under certain circumstances, contributes to the process of L2 learning. Swain ( 1993, 1995, 1998; Swain & Lapkin, 1995 ) extended t he Output Hypothesis and identified three functions that output serves in SLA: ( a ) the hypothesis testing fun ction, ( b ) the metalinguistic function, and ( c ) the noticing / triggering function. The first function describes the process by which the production of output prompts L2 learners to test out their theories regarding how the TL works. The second function awareness of TL rules and other metalinguistic information. The third function states that the production of output serves as an internal priming device for learners to notice the formal features of the language in their subsequent TL input. The noticing function of Schmidt & Frota, 1986), which states that L2 learners must first notice target language forms in order for input to be conver ted i nto intake for learning. Schmidt (1990, 1993, 1995; Schmidt & Frota, 1986) claims that in order for SLA to take place learners need to

PAGE 123

109 production. Swain and Lapkin (1995) assert that the production of output compels that learners do not know or TL information that they cannot remember in order to communicate a message. Therefore, when learners attempt to produce output, they notice what is missing in their IL knowledge, which prompts them to pay closer attention to their subsequent L2 input for the relevant forms and structures. A criticism of the Output Hypothesis is that there h as been little empirical Lapkin, 1995) was quickly accepted without being supported empirically because it rocess. A handful of studies have examined the Output Hypothesis, and qualitative (Swain & Lapkin, 1995; Swain, 1998) and quantitative research studies (Izumi, 2002; Izumi & Bigelow 2000; Izumi, Bigelow, Fujiwara & Fearnow, 1999) lend some support to the assertion that the production of output may prompt L2 learners to engage in mental processes that affect SLA Empirical Studies Examining the Output Hypothesis Swain and Lapkin (1995) i nvestigated the role of output o n the acquisition process. The purpo se of their study was to determine if adolescent learners of French whose L1 was English would be able to notice their linguistic gaps while producing the L2. If the learners become aware of their linguis tic gaps, the researchers attempted to ascertain wh at types of internal cognitive processes we re triggered by noticing them. They were e specially interested to determine if any type of grammatical or syntactic

PAGE 124

110 analysis was employed by the students while attempting to fill in the gaps in their L2 knowledge The study consisted of 18 French immersion students in grade 8. The students were individually (with a researcher present) asked to write a composition in the L2 on a topic that they had previously covered in class The students were asked to think al oud in either French or English while they composed in French. The output of the two most proficient and the two least proficient students were analyzed for language related episodes. A language in which a learner e i ther spoke about a language problem he/she encountered while writing scent learners do become aware of gaps in their L2 knowledge as they pro duce it. Further, when learners become aware of the gaps, they engage in the type of thought processes that may facilitate SLA (Selinker 1972; Corder 1981; McLaughlin 1987; Larsen Fre eman & Long 1991). The two students with the highest proficiency engaged in over twice as much grammatical analysis during production when compared to the two students with the lowest proficiency. Swain & Lapkin assert that although grammatical analysis is not necessary for comprehension, it is essential for accurate L2 production. Their findings resonate with other researchers ( Hulstij n & Hulstijn, 1984; Hawkins & Trowell, 1992) who suggest that conscious knowledge of rules leads to greater L2 accuracy. While Swain and Lapin provide evidence for the importance of output in L2 learning, they do not state that output is the only source of SLA, and they do not discount the necessity of comprehensible input in L2 classrooms. However, they do posit that out put prompts L2

PAGE 125

111 learners to notice their linguistic deficiencies. Once linguistic gaps are noticed, learners are able to sea rch their internal knowledge for the L2 forms and structures that are needed to solve their linguistic problems. Swain & Lapkin sug gest that language acquisition may take place while learners attempt to fill in the gaps in their L2 knowledge, and that knowledge of L2 grammar facilitates the language learning process Swain (1998) investigated the metalinguistic function of the Output Hypothesis. Her mixed methods study investigated two research questions as follows: ( a ) D oes the modeling of meta talk by teachers influence talk? ( b ) Is there a relationship between meta talk and SLA? Her study defined meta talk as the language talk has important implications for SLA because students gain a deeper awareness of the forms and rules of the L2 when they use meta talk for cognitive purposes. The participa nts consisted of 48 secondary students from two French immersion classrooms. The two classes comprised the two treatment groups. The metalinguistic group (M) recei ved modeling by their teacher and the researcher on how to deploy meta talk when they noticed a gap or a hole in their interlanguage. The comparison group (C) received no such modeling. Four dictogloss activities were given to the two groups. The first three were used for modeling and practice, and the fourth was audio taped for analysis in th e study. The fourth dictogloss focused on the pass compose and the imparfait in French. The students received a mini grammar lesson by the researcher prior to th e treatment, and the M group received modeling of meta tal k following the grammar lesson. A dictogloss passage was read twice, the first time the students were asked to listen only, and the second time they

PAGE 126

112 were asked to take notes for the subsequent reco nstruction of the story Students worked in dyads to reconstruct the passage, and their conversations were audio taped and analyzed for language related episodes (LREs) The average number of LREs by the M group was 14.8 while the C group only averaged 5 .8 LREs. Since the M group demonstrated o ver twice as much meta talk as the C group, the researcher concluded that the modeling of meta talk by the teacher resulted in increased meta talk by the students. To analyze whether meta talk facilitates SLA, th e LREs were divided into 4 categories as follows: ( a ) problem solved correctly, ( b ) problem not solved, ( c ) problem solved incorrectly, and ( d responses (from both groups) fell into the first c ategory. Students were given a posttest to assess their knowledge of the targeted grammatical forms. The researcher matched talk to items on the posttest. She found that when students reache d a correct conclusion, there was a strong t endency to perform accurately on the relevant posttest item. Also, if students inaccurately c onstructed knowledge they had a strong tendency to respond inaccurately on the relevant posttest item, which demonstrates that meta talk influences language lear ning. Based on her findings, Swain ( 1998) asserts that meta talk (or the metalinguistic function of output) facilitates SLA. A criticism of the study is that the researcher counted LREs where the participants did not use the metalinguistic terminology th at was demonstrated by the teacher. Therefore, what is considered to be meta talk employed by the students in this study is highly subjective. Further it is unclear if the meta talk that was modeled by the teacher, the meta talk that

PAGE 127

113 was executed by the abilities to answer grammatical items correctly on the posttest. Hypothesis: Izumi, Bigelow, Fujiwara, a nd Fearnow (1999) and Izumi ( 2002). Izumi et al (1999) compared the noticing function of output to exposure to TL input. Their targeted form was the hypothetical conditional in English among 22 undergraduate ESL students with various L1 backgrounds. Th e researchers divided the participants into two groups, and experimental group (EG) and a control group (CG), with 11 participants in each group. The EG participants were asked to read a text and underline any forms that they would need to reconstruct the text. After reading the passage, they were asked to reconstruct the text from memory. This activity was repeated a second time with the same input passage. The CG followed the same protocol as the EG, except that they were asked to answer true/false co mprehension questions rather than reconstruct the story. A week af ter the first treatment, a post test was administered to both groups to measure the uptake of the targeted forms. The second treatment consisted of the same targeted form as the first (the hypothetical conditional), but the protocol was different. The EG group was asked to write an essay on a specific topic that elicited the targeted form, followed by a reading activity that contained the targeted form. They were also asked to underline th e key words that were necessary to comprehend the reading passage. Following the reading (input) activity, the par ticipants were instructed to re write their essays. The CG was also asked to write an essay, but on an unrelated topic. After completing the ir essays, CG participants were also given an input activity that was followed by true/false

PAGE 128

114 comprehension questions. The researchers did not find greater amounts of noticing ne scores. Also, dur ing the first treatment both groups greatly increased their noticing of the targeted form after the second exposure to the input passage Izumi et al did find, however, that the EG demonstrated a significant increase in their ability to accurately produc e the t argeted forms during the second treatment. T heir findings however, were not able to support the claim that the production of output promotes noticing of targeted grammatical forms in subsequent input Their study, however, was flawed because both groups were exposed to the relevant input numerous times, which resulted in the learning of the targeted form by both the control group (CG) and the experimental group (EG) due to repeated exposure. Izumi and Bigelow (2000) did a follow up analysis of I zumi et al and they found that the priming caused by the comprehension q uestions and the input flood that both groups received diminished any differences between the CG and the EG. Further, the targeted form, the English hypothetical conditional did not prove to be perceptually salient for the participants. The researchers suggest that failure to notice the [+perfect] and [+past participle] form in input, or to produce it in output, does not hinder communication. Hence, the functional expendability of t he targeted form coupled with its formal complexity resulted in diminished noticing by both groups (CG and EG). drawing device) and visual input enhancement (an external attention d rawing device) promote the noticing and subsequent learning of targeted grammatical forms (relative clauses in

PAGE 129

115 English). Izumi compared the noticing function of output with visual input enhancement of texts through typographical formatting techniques. Th ere were four treatment groups and a control group as follows: (a) +Output +Enhanced Input, (b) +Output Enhanced Input, (c) Output +Enhanced Input, (d) Output Enhanced Input, and (e ) a control group that received no instruction. The parti cipants were 47 undergraduate ESL learners from two institutions. Participants in all groups were instructed to read a text, and the +Output participants were asked to reconstruct the text while the Output participants were asked to answer multiple choice extension q uestions about the text. All groups were asked to take notes while reading the L2 input passage on the information that they thought was necessary to either reconstruct (+Output groups) or to comprehend ( Output groups) the text. The targeted grammatical forms were embedded in the L2 reading passage. Izumi an d posttests. After analyzing the no te scores, Izumi was not able to support the claim that the production of output promotes greater noticing of relevant forms in subsequent input. text reconstruction ph ase, it was revealed that the production of output does lead to increased noticing of targeted forms. Conversely, Izumi found that VIE was very scores; however, the pres ence of VIE did not result in greater learning gains as measured by pre to posttest scores. Based on these re sults, Izumi concluded that the production of output leads to a deeper level of language processing, which results in greater learning of

PAGE 130

116 the tar is that the targeted form (relativization in English) was too complex to benefit from VIE. cing, which may have been a better indicator of noticing than note scores. Summary of Findings and Implications for Future Research An examination of the previous studies on the Output Hypothesis (Swain & Lapkin, 1995; Swain, 1998; Izumi, Bigelow, Fujiwara and Fearnow, 1999; Izumi, 2002) and that output may play a direct role in the acquisition process. If producing output encourages learners to process language more d eeply, as the previous studies indicate, then participants who receive trad itional instruction should perform equally as well as part icipants who receive PI and structured input activities on interpretation and production tasks in the present study. Furth er, the studies reviewed in this section also indicate that the production of output appears to prime learners to notice targeted forms in subsequent TL input. Following the instructional treatments in the present study, participants were exposed to an au thentic input text that was embedded with 15 instances of the subjunctive in adjectival clauses. Thus, the production of subjunctive verb forms by the traditional inst ruction group should have helped learners to notice the targeted verb forms that were em bedded in the subsequent au thentic input passage Past studies (Izumi et al 1999; Izumi 2002) were not able to support the noticing function of the Output Hypo thesis by examining learners scores or underline scores; rather, Izumi (2002) examined

PAGE 131

117 partially support the noticing function of the Output Hypothesis. Note scores and underline scores alone do not present a clear picture of learners as they only measure the amount of noticin g that takes pl ace and not whether learners notice targeted forms with a high, medium, or low level of awareness. L eo w (2000) asserts that level o f awareness plays a critical role in form learning The present study takes this point into consi deration an d in addition to measuring the amount of noticing that took place, Distance Foreign Language Learners The present study investigates language learning with distance learners who take courses that deliver ins truction according to the traditional distance learning paradigm. Under the traditional paradigm, the emphasis is on independent learning and self instruction though interaction with the course materials. Thus, the emphasis is on the course materials rat her than on the teacher for the provision of instruction. Under the traditional paradigm, the course materials support the learner and the self instruction process in order to maximize learner autonomy ( C. White, 2003). The teacher provides feedback and answers questions, but there are limited opportunities for interaction between the teacher and the student, espec ially in online classes that have a high volume of students The learning site for distance language learners is typically the home or workpl ace, and learners must create or alter their environment so that it is conducive to learning (Gibson, 1998). Distance language learners must organize and structure their physical study space in order to optimize learning, and Gibson (1998) notes that othe rs

PAGE 132

118 to the learning experience. Further, distance language learners tend to have a greater number of life roles such as that of worker, spouse, and/or parent, and typi cally have a wider range of professional, personal, community, and family responsibilities compared to full time students who attend classes on campus ( C. White, 2003). In order to achieve success in a distance language learning environment, Harrell (1998) identifies seven learner attributes that help learners meet the challenges of learning a language at a distance. They are as follows: The ability to meet deadlines, and to develop effective time management The ability to make the psychological adju stment to learning at home Self ife efficiently and effectively Motivation and discipline The ability to manage the lonelines s of distance language learning The ability to self monitor for personal co ntrol over the learning process The ability to assume perso nal responsibility for learning (Harrell, 1998, p. 180) The attributes listed above lead to learner independence or autonomy, which is particularly important in traditional distance language learning paradigm s. However, not all learners are able to cope with the demands that the traditional paradigm places on them. In addition, distance language learners often enter an online course with a high level of motivation, but motivation tends to decline as factors such as competing commitments, social isolation, absence of the structuring aspects of face to face classes,

PAGE 133

119 and difficulty adjusting to learning in a web based environment exert their influence ( C. White, 2003). Thus, the ability to maintain motivation i s an important factor for successful distance language learning. Harris (1995) found that learners who were able to match course features with their own self supporting strategies were able to create for similar to the environments that teachers create for students in face to face language classes. Similarly, C. White (1999) to cope with distance learning as the two m ost important factors for success as a distance language learner Thus, the present study took into account the importance of developing high quality web based materials that are suitable for courses that follow the traditional distance learning paradigm ( where learning takes place as a result of interaction with the materials rather than from interaction with the teacher ) The web based materials that were created for the present study provided learners with directions that were clear and with screen desi gns that were uncluttered to help maximize self instruction though interaction with the materials. In addition, the experimental schedule of the present study took into account the numerous life roles and wide range of personal and professional responsibi lities that distance language learners typically have, and the experimental schedule allowed as much flexibility as possible for completion of the study related activities.

PAGE 134

120 The Spanish Subjunctive as a Complex Grammatical Feature Modality is a semantic no tion that determines the conditions and contexts in which a proposition is judged (Montrul, 2004). Propositions are evaluated according to whether they are possible, impossible, contingently true or false, or necessary. In Spanish, modality can be expres sed through grammatical devices such as the future and conditional tenses and modal verbs. In the following sentences, modality is expressed in Spanish through the future perfect and the conditional tenses respectively: Marcos todava no ha llegado. Habr perdido su vuelo. He will have missed his flight. Dijo que llegara a las seis. He said that he would arrive at six. Modality can also be expressed through grammatical mood in Spanish. Although modality is a feature of ev ery language, expressing modality by means of grammatical mood is not (Montrul, 2004). In the following sentence, modality is marked with the present subjunctive mood in Spanish: Dudo que (l) venga esta noche. I doubt that he will come tonight. The choi ce of indicative or subjunctive in Spanish is signaled by syntactic and semantic factors (Montrul, 2004) such as the expression of doubt in the matrix clause verb in the previous example. In Spanish the subjunctive mood includes present, past, and future forms Intermediate level Spanish language learners often have difficulty expressing grammatical mood, even after they have had a considerable amount of instruction on it

PAGE 135

121 (Terrell, Baycroft, & Perrone, 1987). In order to master the Spanish subjunctive, Collentine (1995, 2000, 2003) posits that learners must develop both morphological and syntactic abilities. In other words, learners must be able to prod uce the indicative and the subjunctive correctly in obligatory contexts, and they must also be able to produce complex sentences that contain both matrix and subordinate clauses. Terrell and Hooper (1974) and Takagaki (1984) assert that the indicative app ears in all syntactic environments, but that the subjunctive tends to appear only in subordinate clauses. An main and subordinate clauses is a late acquired feature a cross languages. Processability Theory identifies production procedures and their sequence of development by L2 learners. Pienemann (1998, p. 9) proposes five hierarchical procedures that underpin Processability Theory, which are presented in Table 2.1. Pienemann claims that each of the aforementioned procedures is acquired independently, but that the procedures are acquired in a fixed order, with the acquisition of one procedure preceding the next. For example, learners must be able to access and produc e words (procedure 1) before lexical categorization (procedure 2) can take place. Thus, kno wledge of words is a necessary, and logical, prerequisite for categorizing grammatical characteristics such as number, person, and gender. Note that the subordinat e clause procedure is the last feature to be acquired in this model. According to Pienemann (1998), L2 learners must acquire procedures 1 4 before they are ready to produce syntax with main and subordinate clauses, which is necessary for using the subjunc tive in Spanish.

PAGE 136

122 Table 2.1 Processing procedure Structural outcome 1. Word / lemma access 2. Category procedure lexical morphemes 3. Phrasal proce dure phrasal information exchange 4. S procedure inter phrasal information exchange 5. Subordinate clause procedure main and subordinate clause ________________________________________________________________________ Similarly, Collentine (1995 ) posits that one of the primary reasons that L2 learners have difficulty with the subjunctive is their inability to form complex syntax. Other problematic factors include the linguistic complexities of denoting abstract concepts such as unreal or hypothe tical events and states (Collentine, 2003) and the difficulty that learners have in noticing the subjunctive morphological inflections because of their similarity to indicative morphological inflections. J.F. Lee (1987) and J.F. Lee and Rodrguez (1997) f ound that the subjunctive inflections for the present tense tend to The targeted form in the present study is the present subjunctive in adjectival clauses when the referent is uncertain, unknown, or hypothetical to t he speaker. Blake (1985) describes the subjunctive in adjectival clauses as choices that are made based on the following seman tic criteria: [+/ Existential Status of the R eferent]. For example, in

PAGE 137

123 the sentence Busco un hombre que sepa programar computad oras or man who knows how to program computers. The speaker of the sentence uses the subjunctive in the subordinate (adjectival clause) to mark the referent as existential or unknown. In the previous example, if the verb in the subordin ate clause were conjugated in the present indicative mood, Busco al hombre que sabe programar computadoras or knows how to program computers, then it would be understood that the referent is not existential in Spanish (e.g. the speaker of the sentence knows the man who can program computers). Note that this distinction (an existential referent) does not change the morphology of the verb in English as is does in Spanish, which is more precise in expressing [+/ Existential] in adj ectival clauses. The subordinate clause verb knows remains in the simple present tense in both of the previous examples in English. Since this language function (expressing an existential referent through grammatical mood) does not exist in English, lear ners of Spanish whose L1 is English tend to mark verbs that require subjunctive morphology with indicative morphology, which is evidence of the L1 transfer phenomena (Terrell, Baycroft, & Perrone, 1987) Further, when the TL has structural, functional, o r semantic elements that are not presen such as marking existential referents with grammatical mood it is expected that learners will have difficulty mastering those elements (Stockwell, Bow en, & Martin, 1965). A lthough the subjunctive exists in English, it is not common. L2 learners of Spanish whose L1 is English have limited L1 models with which to hypothesize about its use in Spanish (Collentine, 2003). In

PAGE 138

124 addition, Mejias Bikandi (1994) asserts that learners must also understand the pragmatic context of utterances in order to grasp the characterization of mood distribution in Spanish, which is a further complication. Farley (2004) posits that Spanish language learners have problems with the subjunctive due to their use of faulty processing strategies, which can be explained by i nput p rocessing (1993, 1996, 2002, 2004). He asserts that L2 learners of Spanish whose L1 is English have difficulty with the subjunctive in nominal clauses due the Lexical Preference Principle and the Sentence Location Principle. The Lexical Preference Principle states that learners will rely on lexical items rather than on grammatical form to extract meaning when both encode the same semantic information. When the subjunc tive occurs in noun clauses following expressions of doubt or denial, the semantic meaning of the subjunctive form is redundant. In the sentence Dudo que (ella) comprenda el problema or I doubt that she understands the problem doubt is expressed in the m atrix clause by the lexical item dudo or I doubt Thus, L2 learners of Spanish tend to overlook the subjunctive form in the noun clause, which also expresses doubt, because they are able to extract meaning from a lexical item in the matrix clause. With t he subjunctive in adjectival clauses, however, this problem does not occur because the subjunctive morphology is typically the only element of the sentence or utterance that expresses an unknown or hypothetical referent. In the sentence Quiero un trabajo que pague bien or I want a job that pays well the subjunctive form pague is the only element in the sentence that expresses an existential referent. However, language learners may still tend to focus on main clause verbs and other elements in sentences s uch as

PAGE 139

125 vocabulary items due to the Primacy of Meaning Principle, which states that learners process their input for meaning before they process it for form. In other words, novice L2 learners are likely to focus on content words rather than on verb forms such as the subjunctive in order to extract meaning from their input. Farley (2001a, 2001b, 2004) also asserts that the Sentence Location Principle poses problems for the acquisition of the subjunctive because subjunctive forms typically occur in subordin ate clauses, which places them in the sentence medial position. The Sentence Location Principle states that learners first process items in the sentence initial position, followed by items in the sentence final position, and items that occur in the senten ce medial position are processed last. Therefore, items that occur in the middle of sentences, like the subjunctive in subordinate clauses, are processed last by learners. Since subjunct ive inflections are already difficult for learners to notice ( J.F. L ee, 1987; J.F. Lee & Rodrguez, 199 7) their placement in the middle of sentences makes it even less likely that they will be detected by L2 learners of Spanish. In summary, Spanish language learners have difficulty acquiring the subjunctive in adjectival clauses for the following reasons: (a) t he linguistic complexities of the subjunctive (e.g. expressing hypothetical, uncertain, or unknown referents), (b) t he lack of English subjunctive models for L2 learners of Spanish whose L1 is English, (c) t he synta ctic complexity involved with the subjunctive and the late acquisition of the subordinate clause procedure by L2 learners, (d) th e lack of perceptual salience o f subjunctive morphology, and (e ) l to t he aforementioned problems, the Spanish subjunctive is a complex feature for

PAGE 140

126 language learners to acquire. It is useful to investigate complex language features when comparing the effects of various types of instruction because if learners are able acquir e a complex form with a particular in structional technique, then the technique should also be beneficial for the acquisition of simple grammatical forms as well.

PAGE 141

127 Chapter 3 Method Introduction This chapter provides a description of the procedures that wer e used to examine the effects of five web based instructional treatments (processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structured input withou t visual input enhancement, and traditional instruction) for the acquisition of the subjunctive in adjectival clauses by intermediate level distance learners of Spanish. The following research questions were addressed within the context of the present stu dy: 1. Is there a differential performance between treatment groups for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks over time ? 2. Is there a differential performance between treatment groups f or the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks over time? 3. ability to notice targeted forms in subsequent authent ic input as measured by note scores and awareness scores?

PAGE 142

128 4. Following the instructional treatments, is there a differential performance meaning of the targeted grammatical form (input processing) and the message of the authentic input text in which it is embedded as measured by grammar comprehension and text comprehension scores? 5. What is the relationship between text comprehension and input processing when learners encounter th e targeted grammatical form in subsequent authentic input? This chapter describes the research design, sample, and population, and it also provides a detailed description of the materials, instruments, and measures that were employed the present study. In addition, a description of the data collection procedures and a comprehensive description of the statistical tests that were used to answer the research questions are provided. Research Design The study em ployed an experimental design; more specificall y, it utilized a pretest posttest control group design. Although the present study did not have random selection from the population, there was random assignment to groups, which controlled for extraneous variables such as gender, SES, and age. There wer e four treatment groups and a comparison group as follows: p rocessing instruction with visual input enhancement (+PI +VIE), processing instruction without visual input enhancement (+PI VIE), structured input with visual input enhancement (+SI +VIE), stru ctured input without visual input enhancement (+SI VIE) and traditional instruction (TI). The

PAGE 143

129 dependent measures investigated were interpretation test scores (with measurement at Pretest, Posttest 1, and Posttest 2), production test scores (with measurem ent at Pretest, Posttest 1, and Posttest 2), as well as noticing awareness, and comprehension test scores. As the participants in the control group received an alternate treatment rather than no treatment, it was referred to as a comparison group in the p resent study. According to Gall, Gall, and Borg (2007), [T]he pretest posttest control group design effectively controls for eight threats to internal validity originally identified by Campbell and Stanley: history, maturation, testing, instrumentation, statistical regression, differential selection, experimental mortality, and selection maturation interaction. (p. 405) Further, by providing an equal but different treatment to the control group, the additional four threats to internal validity identified by Campbell and Stanley (1963), namely, compensatory rivalry by the control group, experimental treatment diffusion, resentful demoralization of the control group, and experimental treatment diffusion, were controlled (Gall, Gall & Borg, 2007). Thus, th e research design of the present study controlled for the potential threats to the internal validity of an experiment that were identified by Campbell and Stanley. In addition, since participants in the study needed to learn the targeted grammatical form i n order to pass the final exams in their Spanish courses, it would not have been ethical to exclude some participants from any instruction on the targeted form by having a true contro l group. Th e four experimental groups were compared with traditional ins truction as it is currently Spanish En Lnea (online) Spanish language course

PAGE 144

130 materials. A full description of the operationalization of traditional instruction with examples is provided in the Instructional Materials section of this chapter. Figure 3.1 provides a visual depiction of the research design that was employed in the present study. One Between Subje cts Factor Instruments / Measures (+PI +VIE) (+PI VIE) (+SI +VIE) ( SI VIE) (TI) Pretest Interpretation Posttest 1 Posttest 2 Pretest Production Posttest 1 Posttest 2 Noticing Measure (Authentic Input Text) Note Scores Awareness Measure Awareness (Posttreatment Questionnaire) Scores Text / Grammar Comprehension Test Comprehension Scores Figure 3.1 Research D esign Ecological validity was addressed by providing an explicit description of the experimental treatments. In addition, participants were not informed about the nature or expected outcomes of the experiment in order to pr event the Hawt horne effect. Further, there w ould not have been an experimenter or teacher effect because all of the treatments Repeated Measures Within Subjects Factors Repeated Measures Within Subjects Factors

PAGE 145

131 were delivered online. Additionally, the pretest was not expected to react with the experimental treatments since participants had no previous knowledge of the use of the subjunctive in adjectival clauses prior to the experiment. Rather, the pretest served as a screening device to remove participants who already had a basic understanding of how the targeted grammatical form funct ions in Spanish. Similarly, the measurement of the dependent variable (pre and posttests) was not a threat to ecological validity because it incorporated two types of subtests, a Production Subtest and an Interpretation Subtest. The experimental groups had practice with activities that were similar to the activities on the Interpretation Subtest during their instructional treatments, and the comparison group had practice with activities that were similar to the activities on the Production Subtest during their instructional treatments. In the present study, population validity, or the generalizability of the study, was limited to undergraduate second semester students of Spanish in the southeast who take 80 100% of their language coursework online. It is not possible to generalize the findings of the present study to all online Spanish language learners in the United States because the researcher was limited to an experimentally accessible population. Population and Sample The target population cons ists of undergraduate students in a southeastern urban/suburban university setting who take Spanish language classes that deliver all of the course content online. The sample for the present study consisted of studen ts enrolled in two intermediate level S panish II distance courses at a large urban university in the southeast, and studen ts enrolled in one intermediate level Spanish II online course at a

PAGE 146

132 small suburban satellite university in the southeast. Sample participants varied in their level of under graduate studies from freshmen to seniors, as students in most undergraduate majors are typically free to complete their foreign language requirement at any point during the ir course of studies. The large urban univer sity typically enrolls up to 125 stude nts per semester in the Spanish II online course while th e small satellite university enrolls only 25 students per semester in the Spanish II online course. All students in these online courses during spring and summer semesters of 2009 were invited to pa rticipate in the study. Although 190 students signed the informed consent form and enrolled in the research study, the final sa mple consisted of 92 students. Forty four of the initial volunteers were excluded because they scored 60% or higher on the Inte rpretation Subtest and/or the Production Subtest of the Subjunctive Knowledge Test that was delivered as a pretest, suggesting that these students already had prior knowledge of the targeted grammatical form. Fifty two students were excluded because they failed to complete the study. Two students were excluded because Spanish was spoken in their homes. Table 3.1 provides a breakdown of student participants by class In order to ensure that high and low achieving students were evenly distributed between the groups, these students were identified by their test average in their Spanish class, and they were assigned to a treatment group using a stratified random assignment procedure. Low achievers were considered to be students whose test score averages we re lower than 50 on a 100 point scale. A total of 34 participants were identified as low achievers from the three classes as follows: 16 from Class One, 15 from Class Two, and 2 from Class Three. Of these, 7 low achieving participants were randomly assig ned to

PAGE 147

133 Table 3.1 Number of Student Participants by Class Initial Sample Final Sample Class University T ype Semester Class size Participants Par ticipants 1 Large urban Spring 2009 125 87 46 2 Large urban Summer 2009 125 90 41 3 Small suburban Summer 2009 25 13 5 Totals: 275 190 92 ________________________________________________________________________ each of the following groups: structured input without visual input enhancement, processing instruction with visual input en hancement, structured input with visual input enhancement and tr aditional instruction. Six low achieving students were assigned to the processing instruction without visual input enhan cement group. However, two low achieving students did not complete the study (one participant in the structured input with visual input enhancement group dropped out of the study and one participant in the processing instruction with visual input enhancement group dropped the course before completing the study). Thus, the f inal breakdown of the 32 l ow achieving participants that completed the study was as follows: processing instruction without visual input enhancement ( n = 6), structured input without visual input enhancement ( n = 7), processing instruction with visual inpu t enhancement ( n = 6), structured input with visual input enhancement ( n = 6), traditional instruction ( n = 7).

PAGE 148

134 High achieving students were also identified in by their Spanish class test average ho wever; only two participants who were identified as high achievers opted to participate in the study. High achievers were students whose test score average was higher than 90 on a 100 point scale in their Spanish class. One high achieving student was randomly assigned to the structured input without visual in put enhanc ement group, and the other high achieving student was randomly assigned to the traditional instruction group. The number and percent of participants in the final sample assigned to the four experimental groups and to the comparison group are rep orted in Table 3.2 Table 3.2 Number and Percent of Final Sample Assigned to Instructional Gr oup s __________________________________ __ ___________________________________ Participants Instructional Group Number Percent __________________ _____________________________________________________ +PI VIE 19 20.65 +PI +VIE 18 19.57 +SI VIE 19 20.65 +SI +VIE 18 19.57 +TI 18 19.57 ______________ ___________________________________________ ___ __________ Total 92 100.00 _________________________________________________________________

PAGE 149

135 In order to have appropriate statistical power to perform multivariate and univariate infere ntial statistical operations, Stevens (2002) recommends that cell sizes for a repeated measures multivariate approach be at least ( a + 10) where a is the number of levels for repeated measures. Similarly, Cohen (1992) recommends that cell sizes be at leas t 16 when using ANOVA (with five groups) in order to detect a large effect size with statistical power set at .80 and alpha set at .05. Thus, the total N necessary for appropriate statistical power in the present study is 80. In a limited meta analysis o f three studies that examined PI (Cadierno, 1995; VanPatten & Cadierno, 1993; VanPatten & Wong, 2003), Collentine (2004) found a large effect size for PI (about six standard deviations). Thus, a total of 92 participants in th e present study was considered to be adequate to detect an effect for the treatments if one existed. Instruments and Measures The following instruments and measures were employed in the present study: (a) a Pretreatment Questionnaire, (b) a Subjunctive Knowledge Test with an Interpret ation Subtest and a Production Subtest, (c ) a Comprehension Test, (d ) Note sheets, (e ) an Authentic Input Text, and ( f ) a Posttreatment Questionnaire. Pretreatment Questionnaire The Pretreatment Questionnaire contained three parts: demographic informat ion, Spanish language learning experience, and computer experience. The demographic information portion was designed to obtain specific background information from participants, including their age, gender, and native language. In addition to the demogra phic information, the questionnaire also asked participants if they speak Spanish

PAGE 150

136 or another language at least half of the time at home, and if they have daily contact with Spanish outside of class. Participants who indicated that they spoke Spanish at ho me or had extensive contact with Spanish outside of class were excluded from the study. Participants were also asked the number of years/semesters that they studied Spanish in high school or college. In addition to the previous information, participants were also asked to give their opinions regarding using computers and the Internet to learn Spanish. Participants were asked to rate their own computer skills and the ease of using Blackboard Courseware Management System, and they were also asked why they chose to learn a language online. Finally, participants were asked if they would take another language class online. The Pretreatment Questionnaire is available in Appendix B. Subjunctive Knowledge Test A Subjunctive Knowledge Test, which was comprised of an Interpretation Subtest and a Production Subtest, was created for this study. The Subjunctive Knowledge Test had three forms (A, B, and C), which were delivered as a pretest and two posttests in the present study. Figure 3.2 provides a visual displ ay of the four constructs measured on the Subjunctive Knowledge Test. A split block design was used to control for test order: Class 1 received test A as the pretest followed by tests B and C as posttests, while Class 2 received test C as the pretest, fol lowed by tests A and B as the posttests, and Class 3 received test C as the pretest followed by tests B and A as the posttests. As with most previous studies that examined PI, there was a 60% cut off for the pretest (Cadierno, 1995; Cheng, 1995;

PAGE 151

137 Subjun ctive Knowledge Test Interpretation Subtest Production Subtest Interpretation Subjunctive Interpretation Indicative Production Subjunctive Production Indicative Figure 3.2 Constructs Measured on the Subjunct ive Knowledge T est VanPatten & Cadierno, 1993a, 1993b; VanPatten & Sanz; 1995; VanPatten & Oikkenon, 1996). If participants scored 60% or higher on either subtest, they demonstrated that they already possessed a basic understanding of the targeted grammat ical form, and they we re excluded from the present study Forms A, B, and C of the Subjunctive Knowledge Test are presented in Appendix C. Interpretation s ubtest The Interpretation Subtest was created for this study and was designed to measure partici (the subjunctive in adjectival clauses is Spanish) and the referential meaning that this grammatical form encodes. Thus, the Interpretation Subtest is not a traditional comprehension test where learne rs are tested on whether or not they understand the propositional content of the message that they hear or read, as learners do not necessarily have to attend to or comprehend grammatical features in order to interpret messages correctly ( R. Ellis, 1995). comprehension of L2 grammar. The Interpretation Subtest consisted of two parts, an aural component and a written component with ten items each. The aural comp onent comprised a series of 10 utteranc es in Spanish where all of the verbs in the main clauses were in the present

PAGE 152

138 indicative, but the verbs in the adjectival clauses that followed were either in the present indicative or in the present subjunctive. Participants had to indicate if the adjecti val clause referred to an antecedent that was certain and/or known, or to an antecedent that was uncertain and/or unknown by selecting the correct response in Spanish. If participants interpreted the subjunctive correctly, they selected a response indicat ing that the referent was uncertain or unknown. Similarly, if participants interpreted the indicative correctly, they selected a response indicating that the referent was certain or known to the speaker. The following is an example of a question from the aural component of the Interpretation Subtest with an English translation (the English translation was not provided to participants). Response B is correct. Participants heard: Quiero ir a un restaurante que sirva comida francesa. I want to go to a res taurant that serves French cuisine. Participants sel ected one of the following responses: A S The sentence refers to a person, place, or thing that clearly exists or is known. B No. The sentence refers to a person, place, or thing that either does not exist or whose existence is unknown. In the previous example, participants had to correctly interpret the meaning of the verb sirva in Spanish, which is conjugated in the present subjunctive in the aural input sentence, in order to answer the question correctly. The subjunctive verb form in th is example connotes a referent that is unknown or hypothetical to the speaker. The aural input component of the Interp retation Subtest contained 7 items that required interpretation of the subjunctive and 3 items that required interpretation of the indica tive.

PAGE 153

139 Similar to the aural component of the Interpretation Subtest, the written component also consisted of 10 items. Participants were provided with 10 written adjectival clauses in Spanish that contained either a subjunctive or an indicative verb for m. Participants had to determine which main clause was appropriate given the adjectival clause that was provided. If the verb in the adjectival clause was in the subjunctive, then participants had to select a main clause that expressed uncertainty or ind efiniteness. Conversely, if the verb in the adjectival clause was in the indicative, then participants had to select a main clause that expressed certainty or definiteness. Thus, participants had to interpret the referential meaning of the verb form in e ach adjectival clause in order to answer the questions correctly. The following is an example of a question from the written component of the Interpretation Subtest with an English translation (the English translation was not provided to participants on t he Interpretation Test). Response A is correct . . hable espaol. . speaks Spanish A. Mi madre no habla ingls, por eso busco un novio que . boyfriend that . B. Mi madre no habla ingls y tengo un novio que . . . In the previous example participants had to correctly interpret that the verb hable in Spanish (which is conjugated in the pres ent subjunctive) connotes a referent that is unknown or hypothetical to the speaker. T he written input contained 8 items that

PAGE 154

140 required interpret ation of the subjunctive and 2 items that required interpretation of the indicative. Each correct answer on the Interpretation Subtest was worth one point, with a maximum total score of 15 for inter preting the subjunctive and a score of 5 for interpreting the indicative when the aural and written components of the test were combined. For the interpretation of t he subjunctive, a score of 11 15 was considered high, a score of 6 10 was considered average, and a score of 5 or below was considered low. High scores were interpreted to indicate that the participants were able to correctly interpret the subjunctive in adjectival clauses in Spanish. Medium scores were interpreted to indicate that participants were partially able to correctly interpret the subjunctive in adjectival clauses in Spanish. Low scores were interpreted to indicate that participants were unable to correctly interpret the subjunctive in adjectival clauses in Spanish. For the interpretation of the indicative, a score of 4 5 was considered high and was interpreted to indicate that participants were able to correctly interpret the indicative in adj ectival clauses in Spanish without overgeneralizing the subjunctive forms. A score of 2 3 was considered average and was interpreted to indicate that participants were partially able to correctly interpret the indicative in adjectival clauses in Spanish w ithout overgeneralizing the subjunctive forms. A score of 0 1 was considered low and was interpreted to indicate that participants were not able to correctly interpret the indicative in adjectival clauses in Spanish and may have overgeneralized the subjun ctive forms.

PAGE 155

141 Each of the three versions of the Interpretation Subtest had an identical format; however, the individual test items varied slightly. All three versions of the Interpretation Subtest were balanced for number and use of subjunctive verb for ms, with each test containing an equal number of verbs in the indicative and in the subjunctive as well as an equal number of regular and irregular verbs in Spanish. In addition, the lexical items used in all three versions of the Interpretation Subtest c ontained high frequency lexical items that participants had already been exposed to during previous Spanish language coursework. The Interpretation Subtest was delivered to participants online, and the test was timed to ensure that participants did not ha ve time to consult outside resources such as their textbooks or the Internet. Durin g piloting with the 18 advanced level students in their fourth or fifth semester of Spanish language study who were already familiar with the test content, it was establish ed that participants would need between 10 to 15 minutes to complete the Interpretation Subtest. Once the tests were completed online, they were printed by the researcher. The Interpretation Subtest was graded by the computer and checked by the researche r to ensure that there were no mistakes. Finally, in designing the Interpretation Subtest, R. creation of interpretation tasks were followed which are listed below: 1. Learners should be required to process the target struct ure, not to produce it 2. An interpretation activity consists of a stimulus to which learners must make some kind of response

PAGE 156

142 3. The stimulus can take the form of spoken or written input 4. The response can take various forms (e.g., indicate true false check a box, select the correct picture, draw a diagram, perform an action) but in each case the response will be either completely nonverbal or minimally verbal (p. 98) Production s ubtest The Production Subtest was created for this study and was designed to m easure how well participants were able to accurately produce subjunctive verb forms in adjectival clauses in Spanish when there was a non referential antecedent. The Production Subtest also measured whether participants overgeneralized subjunctive forms w hen the antecedent was certain or known, which required the production of indicative verb forms. There were a total of 20 items on the Production Subtest and there were three components as follows: 1. Fill in the blank Sentence Completions (5 items) 2. Mini d ialogue Sentence Completions (10 items) 3. Dehydrated sentences (5 items) For the fill in the blank and the mini dialogue sentence c ompletions, participan ts had to write the correct subjunctive or indicativ e verb form. The items in the f ill in the blank comp onent of the Production Subtest were discrete point questions, which are typically easier for novice language learners to answer than test items that are part of a connected discourse in the target language. The following is an example of an item from the f ill in the blank component of the Production Subtest with an English translation (the English translation was not provided to participants). Busco a alguin que _______________ (querer) compartir un apartamento conmigo

PAGE 157

143 ______________ (wants) to share an apartment with me. *The correct answer is quiera which is conjugated in the present subjunctive because the referent is unknown to the speaker of the sentence. The test items in the m ini dialogue component of the Pro duction Subtest were designed to be slightly more difficult for participants because these items were contextualized in short dialogues between two Spanish speakers, which more closely resembles how learners would encounter the targeted grammatical forms i n authentic input. The following is an e xample of a test item from the m ini dialogue component of the Production Subtest with an English translation (the English translation was not provided to participants). Juan: Hay un banco por aqu que 1. __________ _____ (estar) abierto? Is there a bank around here that _______________ (to be) open? Paco: No, no hay ningn banco aqu que 2. _______________ (abrir) a las seis de la maana. (to open) bef ore six in the morning. *The correct answer to number one is est because the referent is unknown to the speaker, and the correct answer for number two is abra because negative expressions that are followed by an adjectival clause in Spanish always take t he subjunctive. For the dehydrated s entences component of the Production Subtest, participants had to take elements of sentences that were devoid of most function words and that only

PAGE 158

144 contained infinitive verb forms to create complete sentences in Spanish. Participants had to decide if verbs in the main and subordinate clauses required conjugations in the present subjunctive or in the present indicative. Below is an example of an item from the dehydrated s entences component of the Production Subtest with an English translation (the English translation was not provided to participants). Write a complete sentence in Spanish describing your ideal house and life. Use the elements that are provided to construct each sentence. Yo / buscar/ casa/ que / tener / ocho dormitorios I / to look for / house / that / to have / eight bedrooms All 5 items in this component of the Production Subtest required the subjunctive in the adjectival clause because the referent (the ideal house, job, or life) was always hypotheti cal. Participants were not given instructions to use any particular tense or mood in Spanish; thus, the dehydrated s entences component of the Production Subtest measured whether or not participants were able to recognize the need to use the subjunctive to express a hypothetical antecedent, and if they were able to produce the appropriate subjunctive form in order to do so. Each correct answer on the Production Subtest was worth one point, with a maximum total score of 15 for pr oducing the subjunctive an d 5 for producing the indicative when the three components of the test were combined If the answer contained misspellings or a lack of agreement in person or number but the participant made an effort to produce the correct indicative or subjunctive verb ending, or if the participant produced a vowel switch or stem change toward the subjunctive form, then .5 was

PAGE 159

145 awarded, which was in keeping with past stud ies in the PI strand. For the dehydrated s entences component of the Production Subtest, only the verb in the subordinate clause was assessed according to the rubric described above, and all other elements of the sentence that participants produced such as the main clause verb and vocabulary items were not scored. For measuring the product ion of the subju nctive, a score of 11 15 was considered high, a score of 6 10 was considered average, and a score of 5 or below was considered low. High scores were interpreted to indicate that the participants were able to correctly produce the subjunctive in adjectival clauses in Spanish. Medium scores were interpreted to indicate that participants are partially able to correctly produce the subjunctive in adjectival clauses in Spanish. Low scores were interpreted to indicate that participants were unable to correctly produce the subjunctive in adjectival clauses in Spanish. The highest total score for measuring the production of the indicative was 5. A score of 4 5 was considered high and was interpreted o indicated that participants were able to correctly produce t he indicative in adjectival clauses in Spanish without overgeneralizing the subjunctive forms. A score of 3 was considered average and was interpreted to indicate that participants were partially able to correctly produce the indicative in adjectival clau ses in Spanish without overgeneralizing the subjunctive forms. A score of 0 2 was considered low and was interpreted to indicate that participants were not able to correctly produce the indicative in adjectival clauses in Spanish and may have overgenerali zed the subjunctive forms.

PAGE 160

146 Each of the three versions of the Production Subtest had an identical format; however, the individual test items varied slightly. All three versions of the Production Subtest were balanced for number and use of subjunctive ve rb forms, with each test containing an equal number of verbs in the indicative and in the subjunctive as well as an equal number of regular and irregular verbs in Spanish. In addition, the lexical items used in all three versions of the Production Subtes t contained high frequency lexical items that participants had already been exposed to during previous Spanish language coursework. The Production Subtest was delivered to participants online, and the test was timed to ensure that participants did not ha ve time to consult outside resources such as their textbooks or the Internet. Durin g piloting with the 18 advanced level students in their fourth or fifth semester of Spanish language study who were already familiar with the test content, it was establish ed that participants would need between 10 to 15 minutes to complete the Production Subtest. Once the tests were completed online, they were printed by the researcher. The Production Subtest was scored by two raters, who were provided with an answer key and a gradi ng rubric for each test. Inter rater reliability was computed, weighted Kappa = 0.97 Comprehension Test The Comprehension Test was created for this study, and it was designed to measure two constructs: (a) text comprehension, which refers to c omprehension of the propositional content of the input passage, and (b) grammar comprehension, which refers to comprehension of the referential meaning of the targeted verb forms The

PAGE 161

147 Comprehension Test comprised the following two components: (a) t he text c omprehension component and (b) t he grammar c omprehension component The text comprehension component of the Comprehension Test contained five multiple choice items that were passage dependent and that tested both the main ideas and the details of the pas sage. The following is an example of an item from the text comprehension component of the Comprehension Test: When the author of the ad states QUE ACEPTEN MASCOTAS ES What must be allowed ? a children b. pets c. collectibles The maximum score on the text comprehension portion of the Comprehension Test was 5. A score of 4 5 was cons idered high, and it was interpreted to indicate that participants understood the propositional conte nt of the authentic input passage. A score of 3 was conside red average, and it was interpreted to indicate that participants were partially able to comprehend the propositional content of the input passage. A score of 0 2 was considered low, and it was i nterpreted to indicate that participants did not understand enough of the propositional content of the authentic input passage to extract an accurate message. The grammar comprehension component of the Comprehension Test comprised 2 multiple choice and 2 short answer questions. The multiple choice questions measured whether participants were able to determine the grammatical mood of the conjugated verb (present subjunctive or present indicative), and the short answer questions determined whether or not p articipants comprehended the referential meaning of the subjunctive

PAGE 162

148 forms (an unknown or hypothetical antecedent). The following is an example of a multiple choice and short answer test item from the grammar comprehension component of the Comprehension Te st: In the following except from a Spanish want ad: Busco una casa . que est en buen estado The author of the ad says that he or she is looking for a house that is in good condition. What form of the verb estar is used? a. present indicative b. present subjunctive Why does the author of the ad use this form of the verb estar ? In other words, what meaning does this form of the verb estar express when conjugated this way? _______________________________ Each multiple choice answer was worth one point, and each short answer question was worth two points. The maximum score on the grammar comprehension component of the Comprehension Test was 6. A score of 5 6 was considered high and was interpreted to indicate that participants comprehended th e referential meaning of the targeted verb forms. A score of 3 4 was con sidered average and was interpreted to indicate that participants were partially able to comprehend the referential meaning of the targeted verb forms. A score of 0 2 was considered low and was interpreted to indicate that participants did not comprehend the referential meaning of the targeted verb forms. The Comprehension Test was delivered via Blackboard, and it was timed to ensure that participants only had enough time to read and answer each question without seeking assistance from their texts, notes, others, or the web. Further, each item was delivered one at a time, and participants were prohibited from backtracking on this exam. The Comprehension Test was piloted with 18 adva nced Spanish language students in their fourth or fifth semester of language study, and it w as determined that students who

PAGE 163

149 already know the subjunctive would need 10 15 minutes to complete the Comprehension Test. Once completed online, the multiple choic e items were scored by the computer and checked for accuracy by the researcher. After participants completed their Comprehension Tests online, they were then printed by the researcher and the short answer items were scored by two raters. The raters were provided with an answer k ey and a grading rubric. Inter rater relia bility was computed, Kappa = 0.92 Test in the present study. Wolf (1993) performed a comprehensive review of the literature on language comprehension testing and devised the following guidelines for the formulation of individual comprehension te st items: 1. That all items be passage dependent 2. That items test information from different levels of the passage, that is, main ideas as well as details 3. That all distracters be plausible 4. That items paraphrase information in the passage so that learners cannot match words and phrases from the item to the passage 5. That test takers not be allowed to refer to the passage whi le performing the comprehension tasks, thereby discouraging surface reading of the passage ( p. 327) All of the guidelines listed above where adhered to except for item number five. The Comprehension Test contained excerpts from an authentic input passag e as a reference to assist participants with answering questions that were related to specific targeted subjunctive verb forms in context. However, as recommended by Wolf, the participants

PAGE 164

150 were not able to refer back to the entire input passage when compl eting the Comprehension Test. The Comprehension Test is available in Appendix D. Validity and Reliability of the Tests The tests that were created for the present study (the Interpretation Subtest, the Production Subtest, and the Reading Comprehension Te st) were checked for reliability and validity. To ensure that interpretations of the test scores were valid, evidence from the test content, evidence from the response process, and evidence from the internal structure of each test was collected as describ ed below. Evidence of test content A panel of foreign language teaching experts who are nativ e speakers of Spanish and whose university teaching experience ranged from three to twenty five years examined the three tests that were employed in the presen t study to construct that it was intended to measure The experts were given objective statements for the instructional treatments, and they logically analyzed whether the tests were consistent with the instru ctional objectives of the treatments in the present study. In addition, they also examined the individual test i tems to evaluate whether the items measured what they were purport ed to measure All three experts agreed that the content of each test as wel l as the individual test items measured what they were designed to measure (interpretation, production, or comprehension of the subjunctive in adjectival clauses in Spanish). The experts also found that the tests were consistent with the instructional obj ectives of the treatments and that they were appropriate for the level of the learner (second semester students of Spanish).

PAGE 165

151 Response p rocess Evidence from the response process was collected through think aloud protocols. A small group of three undergr aduate students in their fourth or fifth semester of Spanish language study were asked to think aloud while taking the tests that were created for the present study (Interpretation Subte st Production Subtest, Comprehension Test). The fourth and fifth sem ester students were already familiar with the targeted grammatical form and had no problems responding correctly to the test items. Participants were audio recorded as they thought aloud, and their statements were transcribed and examined for reflections on the verbs in the adjectival clauses and their referential meaning. For the Interpretation Subtest, participants reflected on the subjunctive verb forms and their referential meaning in order to answer the questions correctly, as was expected. For ex ample, one student reflected on the verb incluir and paid particular attention to the verb ending in order to determine if the verb was in the subjunctive or in the indicative mood, which told him whether the antecedent was certain / known, or uncertain / unknown to the speaker of the sentence: Number one is incluya viajes a paises extranjeros . um and the choices are . tengo un trabajo que or busco un trabajo que and incluir is an ir incluya so I would us e busco un trabajo que because it is unknown if it exists. It is clear from the previous example that the student reflected on the subjunctive verb form in the adjectival clause and the referential meaning that it encodes in order to answer the question c orrectly.

PAGE 166

152 Another student demonstrated a similar pattern for the interpretation of the indicative. For example Busco a una mujer que vende bocadillos . the rule is that if the person is tive is used . . if she exists . the subjunctive is used . um . it is not used if you is vende which is in the present {indicative} so I would choose s . the person exists. In the previous example, the student demonstrated that she understood the grammatical rule for using the subjunctive and the indicative in adjectival clauses. Further, she also reflected on the mood of the verb in the adjectival clause (in this case the verb was in the indicative) in order to interpret the sentence correctly. Similarly, for the Production Subtest, participants reflected on the formation of the subjunctive verb forms and where they were needed in order to answer the test items correctly. The following is an example of a participant reflecting on why he needed to produce a subjunctive verb form to answer the question correctly: Busco una persona que querer compar tir un apartamento conmigo . mmm . so the verb would be . mmm I would say quiera because querer is an er verb and it needs to be in the subjunctive because they T he student above showed the same type of refle ctive thought process when determining whether the verb in the adjectival clause should be produced in the indicative: Hay un apartamento en mi barrio que tener dos dormitorios this person tiene

PAGE 167

153 Thu s, the think alo ud data that were collected while participants took the Interpretation and Production Subtests support the claim that these tests measure the interpretation and production of the subjunctive and/or indicative in adjectival clauses in Spanish. For the Comp rehension Test, participants reflected on specific vocabulary items following example: No est alejado de la ciudad . . mmm . alejado means . but is says without water problems or without trash problems . However, in order to answer grammar comprehension questions, participants reflected on verbs in the adjectival clause, as was expected. For example: Why does the author of the ad says busco un apartamento que est en buen estado . why is the verb estar conjugated this way, in other words, what meaning does it express . well . it expresses that they are looking for a house . house exists . The participant in the previous example reflected on the verb est which is conjugated in the present subjunctive, and its referenti al meaning in order to answer the question correctly. Thus, the evidence from the response process that was collected while participants thought aloud as they took the Comprehension Test supports the claim that this test measures both text comprehension a nd grammar comprehension. Internal s tructure Evidence from the internal structure of the Interpretation and Production Subtests and the Comprehension Test was collected during the piloting, which took place with 18 advanced Spanish language learners in t heir fourth or fifth semester of

PAGE 168

154 language study and with 13 intermediate level Spanish language learners at the end of their second semester of language study. During piloting, the participants took all three versions of the Interpretation Subtest and the Production Subtest as well as the Comprehension Test, which only had one version. To ensure that test items measuring the same construct hung well together, item to total correlations were calculated for each construct that these tests measured. For the Interpretation Subtest, item to total correlations were checked for items that were intended to measure interpretation of the subj unctive and items that were intended to measure interpretation of the indicative. Similarly, for the Production Subtest, ite m to total correlations were checked for items that were intended to measure production of the subjunctive and items that were intended to measure production of the indicative. Finally, for the Comprehension Test, item to total correlations were checked f or items that were designed to measure text compre hension and items that were designed to measure grammar comprehension. After the first round of piloting with the advanced Spanish language students, two items on the Interpretation Subtest that measured i nterpretation of the indicative and two items on the Production Subtest that measured production of the subjunctive were removed because their item to total correlations were significantly lower than the other items that measured these same constructs. Th ese items were reworded and/or problematic vocabulary items were removed and replaced. Similarly, the wording was changed on two items that measured grammar comprehension on the Comprehension Test after the first round of piloting. During the second roun d of piloting, which included the 13 i ntermediate level students at the end of their second semester of language study,

PAGE 169

155 the test items measuring each construct hung well together as revealed by the item to total correlations for each test. After the secon d round of piloting, the item to total correlation for each test item was examined by the researcher, and no single item was found to be significantly lower than the other test items measuring the same construct for any of the tests that were created for t he present study. Reliability e vidence In order to determine if all three forms of the Interpretation S ubtest were equivalent, form s A, B, and C of the Interpretation Subtest were piloted with 18 advanced Spanish language learners in their fourth or fifth semester of language study and with 13 intermediate Spanish language learners at the end of their seco nd semester of language study. Scores from the three administrations of the Interpretation Subtest were correlated to yield a coefficient of equiva lence All of the correlation coefficients that were c omputed reflected a strong positive relationship between the three versions of the Interpretation Subtest. The correlation between tests A and B was r = .78, p < .0001, the correlation between tests A and C was r = .78, p < .0001, and the correlation between tests B and C was r = .95, p < .0001. The correlation between tests B and C may have been higher due to a practice effect. In other words, participants may have become familiar with the instructions a nd format of tests B and C through exposure to T est A, which was administered first. Familiarity with the format and instructions may Similarly, in order to establish that all three versio ns of the Production S ubtest were equivalent, forms A, B, and C were piloted with 18 advanced Spanish language learners in their fourth or fifth semester of language study and with 13 intermediate

PAGE 170

156 Spanish language learners at the end of their seco nd semest er of language study. Scores from the three administrations of the Production Subtest were correlated to yield a coefficient of equivalence All of the correlation coefficients that were computed reflected a strong positive relationship between the three versions of the Production Subtest. The correlation between tests A and B was r = .81, p < .0001, the correlation between tests A and C was r = .89, p < .0001, and the correlation between tests B and C was r = .90, p < .0001. Means and standard deviatio ns of scores on the three forms of the Interpretation and Production Subtest s are reported in Table 3.3. An examination of Table 3.3 reveals that all of the mean scores on the three forms of the respective tests were similar, which provides support for eq uivalence of the three forms. In order to provide evidence of the reliability of the Comprehension Test, the test was administered on two separate occasions to the same individuals, an intact class of 18 advanced students of Spanish in their fourth or fif th se mester of language study and an intact class of 16 intermediate students of Spanish at the end of their second semester of language study. There was a wait time of two days between the two administrations of the Comprehension Test. Participants were instructed not to ask questions or look up information on the targeted grammatical form or its use between the test administrations. The means and standard deviations for both administrations of the Comprehension Test are reported in Table 3.4. The cor relation between scores from the two administrations was calculated to yield a stability estimate of reliability (test retest reliability), which was r = .86, p < .0001 for text comprehension, and r = .93, p < .0001 for grammar comprehension. An examinati on of Table 3.4 reveals that the

PAGE 171

157 Comprehension Test yielded scores with similar means and standard deviations, which provides evidence for the test retest reliability of the Comprehension Test. Table 3.3 Means and Standa rd Deviations of Scores on Three For ms of the Interpretation and Production Subtest s Form of Test Mean SD Interpretation Subtest Form A 1 7.42 3.24 Interpretation Sub test Form B 17.58 3.33 Interpretation Sub test Form C 17.55 3.37 Production Sub test Form A 17.06 3.86 Production Sub test Form B 17.10 4.13 Production Sub test Form C 16.97 3.66 Note. N = 31. Internal consistency reliability As further evidence of the reliability of the tests that were created for the present study (Interpretatio n Subtest, Production Subtest, and computed for each construct that the tests were intended to measure. Estimates of l exceeded .70, which is the minimum acceptable value recommended by Nunnally (1978).

PAGE 172

158 Table 3.4 Means and Standard Deviations of Scores on the Comprehension Tes t Across Two Times of Testing Test c omponent Mean SD Text C omprehension Time 1 4. 44 0.78 Text C omprehension Time 2 4.44 0.86 Grammar C omprehension Time 1 2.50 1.56 Grammar C omprehension Time 2 2.52 1.58 Note N = 34. Reliability estimates were .78, .78, and .77 for the construct interpretation of the sub junctive on the Interpretation Subtest, forms A, B, and C respectively. The reliability estimates were .86, .87, and .83 for the construct production of the subjunctive on the Production Subtest, forms A, B, and C respectively. For the construct interpre tation of the indicative, reliability estimates were .88, .80, and .82 for the Interpretation Subtest, forms A, B, and C respectively. Finally, for the construct production of the indicative the reliability estimates were .72, .82, and .73 for the Product ion Subtest, versions A, B, and C respectively. For the Comprehension Test, two constructs were measured: (a) text comprehension, and (b) grammar comprehension. Reliability estimates were .71 for text comprehension and .77 for grammar comprehension.

PAGE 173

159 No te Sheets One to three days after completing their instructional treatments, participants were asked to read the Au then tic Input Text F ive want ads for renting houses or apartments in Spanish speaking countries comprised t he A uthentic Input Text. Ther e were 15 instances of the subjunctive in the adjectival clause in the Authentic Input Text, with seven different subjunctive verb forms. They were as follows: acepten, alquilen desee, est, sea, sean, and tenga Tenga i s a very common verb in adjectiva l clauses in Spanish, and it appeared six times in the Authentic Input Text. Est also commonly occurs in the adjectival clause in Spanish and it appeared three times in the Authentic Input Text. All other subjunctive forms appeared only one or two tim es in the Authentic Input Text. As participants read the text, they were asked to take notes on what they noticed and perceived to be important whi le reading. After each want ad, there was a text box for participants to record their observations. The d irections for the activity were as follows: As you read this passage, please record any word or words that you feel are important for comprehending the text. Please do not write down every single word, just the vocabulary items and/or verb forms that are necessary for you to understand the text. After participants read the passage and recorded their notes in the text boxes, their results were stored on Blackboard Courseware Management System. The researcher printed Sheets and counted t he number of instances of subjunctive forms that were noted by each participant. A research assistant who is a native speaker of Spanish double Sheet to ensure that the tally was correct.

PAGE 174

160 Each instance of a subjunctive ve rb form that was noted was worth one point, and no points were deducted for misspellings. Thus, p articipants received one point for each subjunctive form that they noted. Since there were a total of 15 subjunctive verb forms embedded in the Authentic Inp ut Text the maximum note score was 15. There was an increased likelihood that participants would note subjunctive forms that were flooded in the input passage ( tenga and est ), rather than those that appeared only once ( dese e and acepten ). The Authentic Input Text reflected how the subjunctive in adjectival clauses is actually used by native speakers of Spanish ; therefore, the number of targeted verb forms that repeated thems elves could not be controlled. As some participants may not have written down s ubjunctive forms that appeared more than once in the passage, the scores was adjusted to account for forms that repeated themselves in the passage. A note score of 6 15 was considered high and was interpreted to indica te that participants noticed subjunctive verb forms in the adjectival clause. A note score of 3 5 was considered average and was interpreted to indicate that participants were able to partially notice subjunctive verb forms in adjectival clauses. A note score of 0 2 was considered low and was interpreted to indicate that participants failed to notice subjunctive verb forms in adjectival clauses in the Authentic Input Text. An example of a Note Sheet is provided in Appendix E. Authentic Input Text In ord er to find authentic examples of the targeted grammatical form in its natural context as native speakers of Spanish use it, the researcher consulted two web sites that post classified ads for free. These websites were www.MundoAnuncio.com and

PAGE 175

161 www.adoos.com.mx The researcher typed busco una casa que or house that and busco un apartamento que or in the search boxes of bo th web sites. Nine ads were retrieved from the web sites on December 10, 2008, and the researcher selected the five ads that contained the most subjunctive forms in the adjectival clause to include in the Authentic Input Text for the present study. The a ds were edited for spelling errors, but not for vocabulary, content, or punctuation. For example, two of the ads were written in all capital letters by their authors. The researcher left two of the ads in all capitals as they were originally written in o rder to maintain their authenticity. Table 3.5 provides a frequency count of the seven different subjunctive verb forms that were pres ent in the Authentic Input Text. The Authentic Input Text is a reflection of how learners are likely to encounter the ta rgeted grammatical form in colloquial usage by native speakers of Spanish. The Authentic Input Text is presented in Appendix F. Posttreatment Questionnaire The Posttreatment Questionnaire was designed as a retrospective measure of s of the subjunctive in adjectival clauses in Spanish as it appears in authentic input. Participants completed the Posttreatment Questionnaire immediately after reading the Authentic Input Text and completing the Comprehension test. Participants were ask ed if they could articulate a particular grammatical form or structure that was present in the Authentic Input Text, and if they were able to do so they were also asked to give an example of such a form or structure in Spanish. It is important to note tha t participants in the structured in put groups (+SI +VIE) and (+SI VIE) did not receive

PAGE 176

162 Table 3.5 Frequency Count of Subjunctive Verbs in the Authentic Input Text ___________________________________________ Verb Frequency ______________ _____________________________ Acepten 1 Alquilen 2 Desee 1 Est 3 Sea 1 Sean 1 Tenga 6 __________ ________________________________ __ Total 15 _____________________________________ _______ any explicit grammar explanation of the targeted form; thus, the Posttreatment Questionnaire was able to detect whether these participants were able to learn metalinguistic information about the Spanish subjunctive inductively. If a participant mentioned the presence of the subjunctive in t he Authentic Input Text on the Posttreatment Questionnaire, then the participant was awarded .5. Further, if the participant was also able to provide a target language example of a subjunctive verb form that was present in the Authentic Input Text, then a nother .5 was awarded. Thus, if a participant expressed that he or she noticed the subjunctive mood and was able to

PAGE 177

163 provide an example of it, then the participant received one point and demonstrated ; Rosa & Leow, 2004). If a participant was also able to state the morphological rule for using the subjunctive in adjectival clauses in Spanish (e.g. when the referent is unknown, uncertain, or hypothetical), then he or she demonstrated awareness at the level of understanding (Rosa points that could be earned on the Posttreatment Questionnaire was 3. A score of 0 .5 indicated that participants had a low level of awaren ess of the subjunctive in adjectival clauses as it appeared in authentic input. A score of 1 was considered medium, and indicated that participants noticed the subjunctive in adjectival clauses in subsequent authentic input and were also able to p rovide a n example of it, which wa s considered awareness at the level of noticing. A score of 2 3 was considered high, and indicated that participants understood the referential meanin g that the verb encodes, which wa s considered awareness at the level of unders tanding. The Posttreatment Questionnaire was printed by the researcher a nd scored by two raters. Inter r ater reliability was computed, weighted Kappa = 0.97 Further, the Posttreatment Questionnaire also asked participants how they felt about the instruc tional treatment package that they completed, and they were asked to compare the grammar instruction and materials that they received in the study with the instruction and materials that they normally receive in their online Spanish classes. Participants were asked to answer the questions as honestly as possible, and they were told that the answers they provided would be kept confidential and anonymous.

PAGE 178

164 Participants were also asked what aspects of their instructional treatments were the most and least hel pful for learning Spanish grammar online. The Posttreatment Questionnaire is presented in Appendix G. Variables The primary independent variable in the study was type of instruction, with five levels (processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structured input without visual input enhancement, and traditional instruction). The within subjects variables were type of task with two levels (int erpretation and production) and time with three levels (Pretest, Posttest 1, and Posttest). Interpretation and production of the subjunctive were measured by scores for interpreting and producing the subjunctive on three forms of the Subjunctive Knowledge test, which was created for the present study and delivered as a pretest and two posttests. The following dependent variables were examined within the context of the present study: comprehension, noticing, and awareness. Comprehension was measured by t ext comprehension scores and grammar comprehension scores on a Comprehension Test that was created for this study. Noticing was measured by note scores, which were scores that participants received from the notes that they took while reading an authentic input passage in Spanish that contained 15 subjunctive forms in the adjectival clause (an online measure) Questionnaire level of awareness of the targeted

PAGE 179

1 65 gra mmatical form as it appeared in authentic input that was received subsequent to the instructional treatments (an off line measure). Instructional Materials Five web based instructional treatment packages were created for the present study that were deli vered via Blackboard Courseware Management System. A separate web based instructional package was developed for each treatment group (processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structured input without visual input enhancement, and traditional instruction). Each web based package reflected a different technique for teaching the Spanish subjunctive in adjectival clauses to online language lea rners. All five treatment packages were balanced for vocabulary and number of activities. In addition, the number of tokens of the targeted verb forms, either produced or interpreted, was identical. The type of feedback given to the participants in each group was also equivalent. All of the group s received implicit feedback. In other words, participants were only told if their answers were correct or incorrect. Correct answers were not provided if participants answered incorrectly in order to avoid pr oviding them with incidental input of the targeted verb forms. A full description of each instructional treatment package is provided in the following section. Multimedia User Interface Design The instructional treatment packages were delivered online us ing a combination of media (text, audio, and pictures), and all of the treatment packages were balanced for the amount of text, audio, and the number of pictures that they contained. Graphical

PAGE 180

166 elements were identical in all five web based treatment packag es. For example, the same colors, screen size, font style and size, navigation bars/buttons, and backgrounds were used in each treatment package. The only variation was for the groups that received visual input enhancement (a between subjects variable in the present study). The processing instruction with visual input enhancement (+PI +VIE) group and the structured input with visual input enhancement (+SI +VIE) group received computerized visual input enhancement of the targeted verb forms, which was ope rationalized as word animation in the present study. In a further attempt to balance the online instructional treatment packages and to increase their efficacy for web based delivery, all of the instructional materials that were developed for the present the areas of computer science, graphics design, instructional design, and psychology, are as follows: 1. Use the medium that best communicates information 2. Use multimedia in a supportive, not a decorative, way 3. Present multimedia synchronously 4. Use elaborative media 5. Make the user interface interactive 6. Use educational multimedia with nave and lower aptitude learn ers 7. Present educational multimedia to motivated learners

PAGE 181

167 8. To avoid developmental effects, use educational multimedia with adults and older children 9. 10. Encourage learners to actively process information (pp. 3 11 332) in a multimedia environment. Research in the field of instructional design indicates that text is better than sound for communicating verbal information when the information must be retained over long periods of time (Chan, Travers, & Van Mondfrans, 1965; Menne & Menne, 1972; Severin, 1967; Sewell & Moore, 1980). Thus, text rather than audio was used as the medium of delivery for the explicit information portion o f the following instructional treatments: processing instruction with visual input enhancement (+PI +VIE), processing instruction without visual input enhancement (+PI VIE), and traditional instruction (+TI). The structured input groups did not receive a n explicit grammar explanation. Further, Najjar (1998) indicates that multimedia should be used in a supportive rather than a decorative way. He also claims that multimedia tools are more effective when used synchronously. The results of several researc h studies indicate that pictures should be used in support of verbal information, and that pictures and illustrations are more effective than texts for helping learners to recall information (Lieberman & Culpepper, 1965; Nelson, Reed, & Walling, 1976; Paiv io & Csapo, 1969, 1973). Najjar

PAGE 182

168 materials and instructional activities are typica lly designed to provide learners with practice activities in four skill areas: reading, writing, speaking, and listening (Omaggio Hadley, 2000). Traditional instruction, as it was operationalized in the present study, contained two aural activities that e licited oral or written language production from participants. The aural activity that required a written response incorporated audio with illustrations to help participants comprehend the aural input. Similarly, the experimental groups also had two act ivities that provided participants with aural TL input, one of which incorporated illustrations to assist participants with comprehension. The illustrations employed in each of the five instructional treatment packages were supportive of the auditory verb al content that was input. All of the treatment packages contained five identical illustrations. Further, the illustrations and audio were presented simultaneously to hel p learners use dual coding (verbal and pictorial) more effectively (Clark & Paivio, 1991; Paivio, 1971, 1986, 1991). Research on dual coding theory supports the claim that information that is processed through both pictorial and verbal channels is more be neficial than information that is processed through a single channel (Barrow & Westley, 1959; Levin, Bender & Lesgold, 1976; Mayer & Anderson, 1991; Nugent, 1982; Paivio, 1975; Paivio & Csapo, 1973; Pezdek, Lehrer, & Simon, 1984; Stoneman & Brody, 1983; We tstone & Friedlandler, 1974).

PAGE 183

169 Another principle that Najjar (1998) sets forth for effective multimedia user interface may allow learners to control, manipulate, and exp Further, interactivity has been documented to have a beneficial effect on multimedia learning (Bosco, 1986; Fletcher, 1989, 1990; Stafford, 1990; Verano, 1987). In the present study, the participants in each instructional tre atment group (processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structured input without visual input enhancement, and traditional instruction) wer e able to control the speed of their instructional activities. Participants were able to navigate both forwards and backwards through the treatment materials at their own pace. Further, the participants in all of the treatments groups received immediate feedback on their written answers (correct versus incorrect), and participants in the traditional instruction group received delayed feedback on their oral answers (correct versus incorrect). Due to the prohibitive costs of designing and implementing spee ch recognition software, the traditional instruction group received delayed feedback (one to three days later) for the single activity that required oral output. This single activity only contained five questions; thus, the difference in feedback type bet ween the groups (immediate versus delayed) was minimized in the present study. Participants in the traditional instruction group recorded their answers for the oral activity using an audio drop box. A voice recorder was embedded on the web page that cont ained the oral activity, and participants were provided with instructions on how to use it. If they were using a desk top computer, an

PAGE 184

170 external microphone was necessary; however, if participants were using a lap top computer with a built in microphone th en the audio recorder utilized the microphone and an external microphone was not necessary. The answers to the oral activity were stored on an external server, which the researcher accessed to listen to and In addition to interactivity, Najjar (1998) proposes two principles for education multimedia user interface design that involve the characteristics of the learner. He asserts that educational multimedia is more effective for beginning level learners and f or learners with low prior knowledge because of its elaborative nature. Further, he posits that multimedia may help beginning level learners because they are not sure where to focus their attention. Another characteristic of novice learners is that they often have trouble distinguishing which information is the most imp ortant The learners in the present study did not have prior knowledge of the use of the subjunctive in adjectival clauses; therefore, they fit the ideal learner characteristics set forth by Najjar for using educational multimedia. In addition, Najjar states that learners who have intrinsic motivation appear to learn better, and he claims that certain instructional designs, such as using an informal style of speech (McConnell, 1978), can i study employed an informal style of speech (the subject pro noun t versus usted in Spanish) and made references to famous people in the traditional and structured input

PAGE 185

171 additional advantage to the use of educational multimedia in the current study is that all participants chose to take their Spanish language coursework online, which may indicate that they are motivated to learn language through multimedia user interfaces. Another characteristic to which Najjar refers is the age o f the learner. He asserts that educational multimedia is more effective with older children and adults than with younger children because younger children process information at the perceptual rather than at the semantic level. Also, younger learners dev elop the ability to process auditory information before visual information (Carterette & Jones, 1967; Stevenson & Siegel, 1969), but as young learners mature, they improve in their ability to process and retain information at the semantic level (Craik & Lo ckhart, 1972; Craik & Tulving, 1975). To avoid these developmental effects, Najjar recommends using educational multimedia with older learners. The participants in the present study were all adults; thus, developmental effects were not a concern. For th attention to the relevant information through multimedia to improve learning. This hypothesis in the field of SLA. In the present study, animation of targeted verb forms (1988) assert that the contrast between an animated figure and a static background increases the pe rceptual salience of the animated object for the learner, and Rieber (1990) asserts that animation serves as an attention

PAGE 186

172 enhanceme nt (VIE) is a between subjects variable in the present study, only the groups that received VIE (processing instruction with visual input enhancement and structured input with visual input enhancement) were given instructional treatments that contained wor d animation. Najjar also recommends encouraging learners to actively process information, which is another principle of multimedia user interface design that pertains to the pears to improve when the learning task encourages the learner to actively process the process information reveals that this condition leads to improved learning performan ce (Auble & Franks, 1978; Bock, 1978; Hunt & Elliot, 1980, Kolers, 1979; Sherman, 1976). This principle is consistent with PI; more specifically, the structured input activities that comprise PI are designed in such a way as to force learners to process t argeted forms correctly in order to extract meaning from sentences or utterances. All of the treatment materials in the experimental groups (processing instruction with visual input enhancement, processing instruction without visual input enhancement, str uctured input with visual input enhancement, and structured input without visual input enhancement) contained structured input activities that forced learners to process the targeted grammatical form in order to answer questions correctly. The participant s in the comparison group (traditional instruction), however, were not forced to process the targeted grammatical form through the in structional activities. Rather than process forms, these participants were forced to produce the targeted grammatical form

PAGE 187

173 Alpha/Usability Testing of Instructional Materials Once the instructional materials were developed for the web, alpha (or usability) testing was conducted to test the user interface of each instructional package and to determine the overall usability of the web based instructional treatment packages. According to Nielsen (1993) usability has five main components: Learnability, Efficiency, Memorability, Errors, and Satisfaction. In order to determine the usability of each web based treatment package, a set of usability measures was developed to defined as the amount of time it takes users to reach a specific level of proficiency. To assess the learnability of t he web based instructional materials, novice users were given five tasks that ranged from easy to difficult to complete. The amount of time that a user took to complete each task on his or her first attempt to use the materials was recorded. Five second semester university level Spanish language learners who use computers, the Internet, and Blackboard CMS frequently were asked to take part in the usability study (five per instructional treatment). Niels e n and Landauer (1993) assert that five users are ab le to detect 85 percent of the usability problems that are present, and testing more users only results in diminishing returns. The researcher worked with one student at a time in The results of the learnability study revealed that students did not experience any difficulties initially completing the five tasks for each instructional treatment package. The easiest task in each of the instructional treatment packages, opening the a ctivity on Blackboard with a password, only took users between 19 to 24 seconds to complete. The

PAGE 188

174 most difficult task, using the audio drop box, which only applied to Grammar Package 1, took students between 70 to 75 seconds to complete. Table 3.6 provide s an overview of the learnability assessment for each of the five instructional treatment packages. Further, an examination of the standard deviations in Table 3.6 reveals that none of the tasks had a standard deviation of greater than 5 seconds, which su pports the conclusion that learners were quickly able to reach a high level of proficiency using the web based instructional materials. ( Nielsen 1993). Since the instructional materials were new to all of the participants in the present study and the treatment packages were designed for a single one time use, it was not necessary to measure efficiency. Similarly, memorability, which is the ability to remember how to use the in structional materials based on previous learning experiences with them ( Nielsen 1993), was not a concern of the present study since participants only used the treatment materials once. Errors are defined as actions that do not accomplish a desired purpos e, and they are measured by counting the number of incorrect actions that users make while attempting to perform a specified task (Nielsen, 1993). Another way to view errors is the number of deviant clicks that users make while they perform a task. In or der to compute an error score for each treatment package, the number of clicks to perform five separate tasks was computed and compared with the actual number of clicks that it takes users to perform these tasks. The error scores for five tasks in each of the web based instructional treatment packages are displayed in Table 3.7.

PAGE 189

175 Table 3.6 Learnability Assessment for Each Task by Instructional Treatment Package Mean SD Min. Max. Package (seconds) (seconds) (seconds) (seconds) Package 1 (+TI) 1 Open the Grammar Package with password 21.30 2.14 18.99 24.43 2 Open the link for the grammar explanation 35.13 1.44 33.80 37.14 3 Open the link for 4.1 and sel ect a response 66.06 4.40 59.19 70.51 4 Open the link for 6.8 and type a response 29.54 1.54 27.12 31.07 5 Use the audio drop box to record an oral 72.25 2.09 70.04 74.59 respon se to activity 10.1 Package 2 (+PI VIE) 1 Open the Grammar Package with password 19.82 1.40 18.31 21.97 2 Open the link for processing strategies 45.93 2.22 42.77 48.34 3 Open the link for 2.6 and select a res ponse 48.57 2.10 46.28 51.53 4 Open the link for 5.4 and type a response 43.55 3.96 38.58 49.36 5 Read the instructions for activity 8 and 70.38 1.97 68.17 73.36 put the sentences in order Package 3 (+SI VIE) 1 Open the Grammar Package with password 19.91 0.96 18.84 21.12 2 Click on response for activity 1.3 37.68 4.21 33.30 44.58 3 Open the link for 2.4 and select a response 54.10 4.96 46.47 59.71 4 Open the link f or 5.5 and select in a response 52.48 2.31 49.36 55.71 5 Read the instructions for activity 8 and 66.04 2.52 62.36 68.76 put the sentences in order

PAGE 190

176 Mean SD Min. Max. Package (seconds) (seconds) (seconds) (seconds) Package 4 (+PI +VIE) 1 Open the Gr ammar Package with password 21.14 1.60 19.85 23.80 2 Open the link for processing strategies 40.23 3.71 34.65 44.82 3 Open the li nk for 2.6 and select a response 46.75 2.48 43.43 49.72 4 Open the link for activity 6.4 and respond 49.03 2.83 45.94 52.48 5 Read the instructions for activity 10.5 44.57 2.49 41.38 47.69 and type a response (word animation) Package 5 (+SI +VIE) 1 Open the Grammar Package with password 19.79 1.02 18.81 21.06 2 Click on response for activity 1.3 45.71 4.99 41.05 52.18 3 Open up the link for 2.4 and select a response 48.78 2.71 45. 86 52.02 4. Complete 7.1 and rank the sentence 50.40 3.12 47.11 54.72 5. Read the instructions for activity 10.5 43.97 4.34 37.39 49.09 and type a response (word animation) Note There were 5 users per instructional treat ment package. An examination of Table 3.7 reveals that the error rate for each task was low. Thus, students did not make many incorrect actions, as measured by deviant clicks, while completing five tasks in each of the web based instructional treatment p ackages. The students who took part in the usability study were representative of the actual participants in the research study in that they were accustomed to using Blackboard CMS on a daily basis to access course materials and complete assignments.

PAGE 191

177 Tabl e 3.7 Error Score for Each Task by Instructional Treatment Package Package Mean clicks Expected clicks Package 1 (+TI) 1 Open the Grammar Package with password 4.0 4.0 2 Open the link for the grammar explanation 6.2 6.0 3 Open the link for 4.1 and select a response 5.2 5.0 4 Open the link for 6.8 and type a response 5.0 5.0 5 Use the audio drop box to record an oral 8.6 8.0 response to activity 10.1 Package 2 (+PI VIE) 1 Ope n the Grammar Package with password 4.2 4.0 2 Open the link for processing strategies 6.0 6.0 3 Open the link for 2.6 and select a response 5.2 5.0 4 Open the link for 5.4 and type a response 5.2 5.0 5 Read the instructions for activity 8 and 8.2 7.0 put the sentences in order Package 3 (+SI VIE) 1 Open the Grammar Package with password 4.0 4.0 2 Click on response for activity 1.3 3.0 3.0 3 Open the link for 2.4 and select a response 5.2 5.0 4 Open the link for 5.5 and select in a response 5.0 5.0 5 Read the instructions for activity 8 and 7.8 7.0 put the sentences in order

PAGE 192

178 Package Mean clicks Expected clicks Package 4 (+PI +VIE) 1 Open the Grammar Package with password 4.4 4.0 2 Open the link for processing strategies 6.2 6.0 3 Open the link for 2.6 and select a response 5.0 5.0 4 Open the link for activity 6.4 and type in a response 5.4 5.0 5 Read the instructions for activity 10.5 and 5.2 5.0 type a response (word animation) Package 5 (+SI +VIE) 1 Open the Grammar Package with password 4.0 4.0 2 Click on response for activity 1.3 3.2 3.0 3 Open up the link for 2.4 and select a response 5.0 5.0 4 Complete 7.1 and r ank the sentence 5.4 5.0 5 Read the instructions for activity 10.5 and 5.2 5.0 type a response (word animation) Note There were 5 users per instructional treatment package. Subjective satisfaction, or how well users liked the web based mat erials, was determined by asking participants in the pilot tests for their subjective opinion about the materials. Nielsen (1993) recommends averaging the replies of multiple users to obtain pleasant user interface likelihood that learners will complete the web based instructional materials. Ten participants who took part in the pilot tests completed the Satis faction Survey The

PAGE 193

179 Satisfaction Survey asked participants to rank five statements on a scale of 1 5 with 1 being strongly agree and 5 being strongly disagree. The results of the satisfaction survey indicate that participants enjoyed completing the web based activities in Packages 2 4 (processing instruction without visual input enhancement, structured input without visual input enhancement, processing instruction with visual input enhancement, and structured input with visual input enhancement respectiv ely). However, participants in the pilot study indicated that they did not enjoy the output based activities that comprised Package 1 (traditional instruction) because they were too similar to the instructional materials that they complete on Quia, the we bsite that hosts the online workbook and lab manual activities for the face to face Spanish Package 1 (traditional instruction) were told that the grammar activities in P ackage 1 were based on their regular course materials. Since students who piloted the materials perceived their regular course materials as being difficult and frustrating, this negative opinion likely influenced how they perceived the web based materials that they were was made not to tell study participants in the traditional instruction group that their materials were based upon their regular course materials; rat her, all participants were told that they would be receiving a novel instructional technique for teaching complex Spanish grammar. The results of the satisfaction survey are presented in Table 3.8.

PAGE 194

180 Table 3.8 Results of the Satisfaction Survey by Instruct ional Treatment Package Package Average score Package 1 (+TI) The directions were clear and easy to follow 1.0 I learned something from completing the activity package 4.0 I preferred these activities to my regular classroom activities 4.5 It was easy to complete the web based grammar activities 3.5 I enjoyed learning Spanish grammar using the materials 4.0 Package 2 (+PI VIE) The directions were clear and easy to follow 1.0 I learned something from compl eting the activity package 1.0 I preferred these activities to my regular classroom activities 1.0 It was easy to complete the web based grammar activities 1.5 I enjoyed learning Spanish grammar using the materials 2.0 Package 3 (+SI VIE) The dire ctions were clear and easy to follow 1.5 I learned something from completing the activity package 1.0 I preferred these activities to my regular classroom activities 1.0 It was easy to complete the web based grammar activities 1.5 I enjoyed learn ing Spanish grammar using the materials 1.5

PAGE 195

181 Package Average score Package 4 (+PI +VIE) The directions were clear and easy to follow 1.5 I learned something from completing the activity pack age 1.5 I preferred these activities to my regular classroom activities 1.0 It was easy to complete the web based grammar activities 1.5 I enjoyed learning Spanish grammar using the materials 2.0 Package 5 (+SI +VIE) The directions were clear and ea sy to follow 1.5 I learned something from completing the activity package 1.0 I preferred these activities to my regular classroom activities 2.0 It was easy to complete the web based grammar activities 1.0 I enjoyed learning Spanish grammar using the materials 1.5 Note There were 2 users per instructional treatment package. The satisfaction survey also revealed that participants who com pleted instructional treatment P ackages 2 4 believed that the directions were clear and easy to follow. Par ticipants who received Packages 2 4 also indicated that they learned something from the instructional materials, that the materials were easy to complete online, and that they preferred this type of instruction to their r egular classroom instruction. Furt her, p articipants who comple ted the Satisfaction Survey had the opportunity to provide additional comments about their experiences completing the web

PAGE 196

182 based instructional treatment packages. One participant who com pleted instructional treatment P ackage 4 ( to use. Felt like I was comprehending instead of memorizing grammar and regurgitating pleted instructional treatment P ackage 3 (structured input without vi believe it should be used in regular classrooms! Very helpful in building my knowledge In summary, no problems with us ability emerged during alpha testing of the web based instructional materials Participants who took part in the alpha tests had no trouble using Blackboard CMS to access the materials, and similarly they had no difficulty opening the links to the grammar explanations, audio files, and/or word animation files depending upon the instructional treatment package. In addition, although participants that worked with Package 1 (traditional instruction) at first seemed hesitant to use the audio drop box since th is technology was completely new to them, they had no trouble following the instructions and recording their responses using the audio recorder th at was embedded on a web page that was delivered via Blackboard. Further, it was expected that the materials in P ackage 1 would not necessarily be pleasing to students since these materials represented traditional instruction, which places a heavy emphasis on output based production and mechanical drill activities which learners perceive as tedious Conversely, the students who com pleted instructional treatment P ackages 2 4 found the materials to be pleasing and easy to use.

PAGE 197

183 Beta and Pilot Testing of the Web Based Instructional Materials Beta tests were conducted to test the content of the treatment packages, a nd pilot tests were conducted to test the flow of the web based instructional packages. For the beta test, two representative second semester Spanish language students and one representative subject matter expert (SME), a Spanish instructor who is a nativ e speaker of Spanish, were invited to give their opinions on the content of the five web based instructional treatment packages. The researcher conducted the beta test with each participant individually, and participants were stopped after they completed each screen (in each of the five treatments) in order for the researcher to ask their opinions about the content. The researcher took notes as the participants made comments. After the beta tests were conducted for each participant, the comments were con solidated and examined for trends. The beta test with the SME uncovered two grammar/spelling mistakes in the Spanish language content that had to be corrected. The beta test s with the student participants, however, revealed many more areas that needed a ttention and revision. The most serious problem was that students did not remember verbs and vocabulary words to which they were exposed in previous Spanish language coursework. Their lack of comprehension of the meaning of basic verbs and vocabulary ite ms negatively impacted their ability to answer questions regarding the targeted verb forms in the instructional treatment materials. The following Spanish verbs and vocabulary items were problematic for participants in the first round of beta tests:

PAGE 198

184 Verb s: buscar, cargar, cambiar, costar, contribuir, descansar, encontrar, ensear, ofrecer, prestar Vocabulary items: alguien, bocadillos, esquina, extranjero, nadie, piscina, peso, pluma, vecinos, vista After the first round of beta tests, the researcher cr eated a vocabulary practice activity that students had to complete with 80% accuracy or higher prior to beginning their grammar packages. Further, the researcher revised the web based instructional materials to provide English translations of the most pro blematic vocabulary items and verbs that are listed above. The Vocabulary Practice Activity is available in Appendix I. Another problem area that was revealed in the first round of the beta tests was the wording in the explicit grammar explanation that w as provided to participants in the traditional instruction and processing instruction groups. Several of the participants in the beta tests had trouble understanding the terminology opposite ending when referring to the formation of the present subjunctiv e mood. This wording was removed from the instructional materials, and the terminology used in the explicit grammar explanations was simplified. The last major problem uncovered by the beta tests was with the flow of the instructional activities. Some o f the participants did not notice when the instructions changed from one activity to another. The researcher revised the materials adding the following statement in between activities: NOTE: THIS ACTIVITY HAS A NEW SET OF DIRECTIONS! The previous statemen t alerted students to pay attention to the new set of instructions.

PAGE 199

185 Once the instructional materials were revised, another round of beta tests was conducted with five university students at the end of their second semester of Spanish language study. One participant was assigned to each of the instructional treatment packages for the second round of beta tests. The addition of the vocabulary practice activity, the inclusion of English translations for problematic vocabulary items and verb forms, the revis ion of the wording in the explicit grammar explanations, and the additional directions prompting students to pay attention to the directions when the activities changed all appeared to be effective. The second round of beta tests did not reveal any signif icant problems with the web based instructional materials. The pilot tests were conducted after the alpha and beta tests were completed with university students a t the end of their second semester of Spanish language study who use computers and the Internet frequently. The participants in the pilot test were representative of the actual participants in the study. Two participants were assigned to each treatment p ackage for the pilot tests. The participants were not interrupted as they completed their web based instructional packages in order for the researcher to determine the amount of time it takes for a learner to complete a given treatment. The time that eac h participant took was recorded on Blackboard Table 3.9 provides a description of the time that participants took to complete the instructional treatment packages. Most participants were able to complete their grammar package within one and one half hour s. Only one participant who completed Package 4 (processing instruction with visual

PAGE 200

186 Table 3.9 Time Needed to Complete the Instructional Treatment Packages Package Time on t ask Package 1 (+TI) Participant 1 1 hour, 21 minutes, 25 secon ds Participant 2 1 hour, 2 minutes, 32 seconds Average 1 hour, 12 minutes, 59 seconds Package 2 (+PI VIE) Participant 3 1 hour, 23 minutes, 38 seconds Participant 4 1 hour, 12 minutes, 39 seconds Average 1 hour, 18 minutes, 9 seconds Package 3 (+SI VIE) Participant 5 1 hour, 0 minutes, 19 seconds Participant 6 1 hour, 21 minutes, 50 seconds Average 1 hour, 11 minutes, 5 seconds Package 4 (+PI +VIE) Participant 7 1 hour, 45 minutes, 44 seconds Partic ipant 8 1 hour, 0 minutes, 21 seconds Average 1 hour, 23 minutes, 3 seconds Package 5 (+SI +VIE) Participant 9 1 hour, 7 minutes, 25 seconds Participant 10 1 hour, 18 minutes, 22 seconds Average 1 hour, 12 minutes, 54 seconds N ote. N = 10 participants in the pilot test

PAGE 201

187 input enhancement) took over that amount of time. The researcher observed this participant taking copious amounts of notes while he completed his grammar activity package, which slowed down his progress conside rably. Interestingly, of the ten participants in the pilot study, only one took notes as he went through the materials. Further, the pilot study revealed that the inclusion of an explicit grammar expla nation in P acka ge 1 (traditional instruction), P ackag e 2 (processing instruction without visual input enhancement), and P ackage 4 (processing instruction with visual input enhancement) did not appear to increase the amount of time that participants would need to complete the materials as one participant who completed P ackage 1 and one participant who completed P ackage 4 were able to do so in slightly over one hour. It appeared that individual differences between participants played a role in the amount of time that it took to complete an instructional treat ment package. The researcher checked the work of all of the participants in the pilot study to ensure that they did not approach the materials in a cursory way. All of the participants in the pilot study scored 80% or higher, which indicated that they di d put forth effort while completing the web based instructional materials during the pilot test. It was determined from the pilot test that participants in the research study would need at least one hour to complete their web based instructional treatment packages, but that some students who work more slowly would need more time. Thus, participants in the research study were given up to two hours to complete their instructional treatment packages online.

PAGE 202

188 In add ition, the ten participants who took part in the piloting of the instructional treatment materials were a lso asked to pilot a reading a ctivity (note taking activity with the Authentic Input Text), the Comprehension Test, and the Posttreatment Questionnaire online. Participants took between 10 to 20 minutes to complete each of the aforementioned activities online. Given these results, participants in the research study were allowed up to 30 minutes to complete each of these activities online. Traditional Instruction The traditional instruction t reatment activities were based on the activities found En Lnea 2.0 Spanish distance learning course as well as the accompanyin g loose leaf companion text The traditional instruction treatment package began with an explicit exp lanation of the targeted grammatical form and its rules of use in Spanish. In addition, several target language examples followed the grammar explanation. With traditional instruction, the full paradigm of present subjunctive forms was presented to parti cipants, who were required to produce all of the subjunctive forms immediately following the grammar explanation through output based practice activities. The Traditional Instruction Explicit Grammar Explanation is presented in Appendices J and K Follow ing the grammar explanation, the traditional instruction treatment package contained ten output based practice activities. The activity types were as follows: mechanical drill, transformational drill, meaningful drill, and open ended communicative. The m echanical and transformational drills had only one possible correct answer, and learners did not have to comprehend the meaning of the input sentences in these activities in order to respond correctly. With the meaningful drill activities, learners had to attach

PAGE 203

189 meaning to both the stimulus and the response, but the intended meaning of the response was already known before the question was asked. It is important to note that authentic communication does not take place with meaningful drills. Further, the re was only one possible correct answer for the meaningful drill activities. Conversely, the open ended communicative practice activities required learners to comprehend both the stimulus and the response, and the intended meaning of the learner was not k nown in advance. See Sequential Ordering. Of the ten activities in the traditional instruction web based treatment package, there were two mechanical drill activities, two transformational drill activities, four meaningful drill activities, and two open ended communicative activities. Thus, forty percent of the activities focused on form and not meaning, and sixty percent of the activities focused on meaning and form, w hich is consistent with the amount and type of activities that are employed to practice a new grammatical structure in the En Lnea course materials. Past studies that compared PI with traditional i nstruction (Benati, 2001, 2005; Cadierno, 1995; Cheng, 199 5, 2002; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004) operationalized traditional instruction with fifty percent of the activities focusing on form only, and fifty percent focusing on form and meaning. VanPatten (2002) asserts that traditio nal instruction as it has been operationalized in the PI strand of research under his direction is ubiquitous in both classrooms and texts in the United States.

PAGE 204

190 Two of the activities in the traditional instruction web based treatment package were aural. One of the aural activities required a written response and the other required an oral response from participants. The aural activities included one transformational drill activity and one open ended communicative activity. The aural transformational d Instruction Treatment Package is presented in Appendix L Processing Instruction Processing instruction was operationalized according to the guidelines set forth by J.F. Lee and VanPatten (2003), which includes an explicit explanation of grammar that is nonparadigmatic. In other words, only one grammar point should be presented at a time ore, while traditional instruction presented the entire paradigm of the present subjunctive verb forms, the PI materials in the present study only focused on the third person singular and plural forms. The Processing Instruction Explicit Grammar Explanati on is presented in Appendices M and N After the PI participants read an explicit grammar explanation, they received information on processing strategies to help them master the subjunctive in adjectival clauses. The information on processing strategies included the faulty processing strategies that Spanish language learners are likely to engage in when reading input sentences that contain the subjunctive in adjectival clauses. In addition, participants were presented with more optimal strategies for pro cessing the subjunctive. When learning the subjunctive in adjectival clauses in Spanish, learners are likely to have difficulty with the

PAGE 205

191 m odel of i nput p rocessing: Princip le 1. The Primacy of Meaning Principle Learners process input for meaning before they process it for form. Principle 1f. The Sentence Location Principle Learners tend to process items in sentence initial position before those in final positio n and t hose in medial position (p. 14) meaning of their linguistic input before they focus on grammatical form. The subjunctive is a form that is particularly difficult for stude nts to notice because the present subjunctive endings are very similar to the present indicative verb endings. For example, the third person singular indicative form of the verb hablar (to speak) is habl a while the third person singular subjunctive form i s habl e This subtle difference in form (a vowel switch), which is often overlooked by L2 learners of Spanish, denotes an entirely different grammatical mood. Participants in the PI group had the Primacy of Meaning principle explained to them, and they w ere given alternate strategies to overcome this faulty processing strategy. For example, they were directed to pay particular attention to the verb endings in order to determine the grammatical mood of the TL input that they receive. In addition to the P rimacy of Meaning principle, the present study also focused on the Sentence Location principle, which states that learners process information in the sentence initial and sentence final position before they process information in the sentence medial positi on. The targeted grammatical form in the present study always

PAGE 206

192 occurs in the sentence medial position in natural speech. When the subjunctive occurs in an adjectival clause, it is the subordinate clause of a sentence or utterance, which causes the subjunc tive form to appear in the middle of the sentence. For example, in the following Spanish sentence, the subjunctive form pueda occurs in the sentence medial position: Busco a alguien que pueda limpiar la casa. e. Participants were made aware of their tendency to overlook items in the sentence medial position. They were also directed to pay attention to the verb form in the middle of sentences in order to extract meaning (whether the referent is unknown / hypot hetical or known / certain), which is a more optimal processing strategy. The Information on Processing Strategies is presented in Appendix O Structured input activities are the final component of PI. J.F. Lee and VanPatten (2003) describe the guidelin es for developing structured input activities in detail, and Wong (2004) stresses that the guidelines must be followed explicitly in order to create authentic structured input activities. Research on the components of PI (Benati, 200 4b ; Fernandez, 2008; V anPatten & Oikkenon, 1996) indicates that structured input is the most important feature of PI. Wong (2004) also cautions that not every activity that is input based is automatically a structured input activity. A review of the literature on PI revealed that flawed structured input activities yielded results that were not comparable to the PI strand ( L. Allen, 2000; DeKeyser & Sokalski, 1996; Erlam, 2003; Nagata 1998; Salaberry, 1997) The present study maintained strict treatment fidelity to PI by caref ully

PAGE 207

193 following the published guidelines for creating structured input activities ( J.F. Lee & VanPatten, 1995, 2003). J.F. Lee and VanPatten (2003) present the following six guidelines for developing structured input activities: 1. Present one thing at a time 2. Keep meaning in focus 3. Move from sentences to connected discourse 4. Use both oral and written input 5. Have the learner do something with the input 6. (p. 154) Because L2 learners are limited capacity proce ssors (McLaughlin, 1987; McLaughlin, Rossman, & McLeod, 1983), J.F. Lee and VanPatten (2003) assert that input must be delivered to learners in an efficient way. They claim that by presenting learners with one grammatical form or function at a time, they are more likely to notice and process the targeted feature. Guideline two is a general recommendation to keep the communicative intent of sentences and utterances as the central focus of structured input activities. Wong (2004) proposes that learners m ust understand the propositional content of the input that they receive in order to successfully complete structured input tasks and activities. Further, she cautions that L2 learners should not be able to complete structured input tasks and activities wi thout comprehending the referential meaning of their input. This guideline directly opposes the mechanical drill activities that are prevalent in the audiolingual, PPP, and traditional output based instructional methods. With mechanical drills, L2 learne rs

PAGE 208

194 need not comprehend the referential meaning of their linguistic input. Conversely, with J.F. Lee and VanPatten (2003) recommend that structured input activities begin with short, sentential level input, and then gradually move towards longer passages of processing capabilities, which would likely result in them skipping over the ta rgeted grammatical form in favor of processing content words. By presenting L2 learners initially with sentential level input (which is easier to process), they are more likely to pay attention to the targeted linguistic feature. The fourth guideline rec ommends that L2 learners should receive both written and spoken input. J.F. Lee and VanPatten (2003) suggest that structured input activities can be either written, spoken, or a combination of the two. They state that the principal reason for providing l earners with both types of input is to make adjustments for individual differences in language learning as some learners benefit more from visual cues while others prefer to learn by listening. Perhaps one of the most innovative features of structured inp ut activities pertains to guideline five. J.F. Lee and VanPatten (2003) assert that L2 learners should not be passive recipients of TL input, which is not sufficient for acquisition to take place. Rather, L2 learners should become actively involved with their input to increase the likelihood that they will process the targeted grammatical form(s). The authors suggest agreeing/disagreeing, checking off things that app J.F.

PAGE 209

195 Lee and VanPatten, 2003, p. 158). Although learners actively attend to their input during structured input activities, it is important to note that they do not produce (either in speaking or writing) the targeted g rammatical form. Learners may produce output during however, their production involves alternative features rather than the target ed grammatical form. Figure 3.3 pr oves a visual depiction of the major types of structured input activities. Although structured input activities prohibit students from producing the targeted grammatical form, VanPatten (2004) is not opposed to learners producing other types of output. H grammatical forms upon their first exposure to them, which is the initial step in the acquisition process. Thus, access and production procedures do not figure into this inst ructional technique. Supplying Information Binary Options Surveys Ordering/Ranking Matching Selecting Altern atives Figure 3.3 Major Types of Structured Input A ctivities (from J.F. Lee & VanPatten, 2003) Structured Input Activities

PAGE 210

196 f input processing (1993, 1996, 2002, 2004) need to be carefully examined before developing structured input activities. Faulty processing strategies need to be identified, and activities that push learners toward more optimal processing strategies should be created (Wong, 2004). For example, for the Lexical Preference Principle (which states that learners will tend to rely on lexical items rather than on grammatical form when both encode the same meaning), all structured input activities that are created should attempt to remove redundant lexical items in order to prompt learners to glean the communicative intent of the sentence or utterance from the targeted grammatical form rather than from lexical items within the sentence or utterance. The previous p rocessing strategy and all of the others were taken into account when developing the structured input activities for the present study. Further, J.F. Lee and VanPatten (2003) describe two types of structured input activities, referential and affective. I n referential activities, learners must extract the meaning of the sentence or utterance from the targeted grammatical form. Also, with referential activities there is a right or a wrong answer, which allows the teacher or researcher to determine whether or not learners are attending to the targeted grammatical form for the meaning that it encodes. The guidelines also suggest having learners begin with two or three referential structured input activities and following these with affective structured input activities. Affective activities are those in which learners communicate a belief, an opinion, or an affective response as they engage in processing real world information ( J.F. Lee & VanPatten, 2003). The present study employed a total of ten

PAGE 211

197 structure d input activities, five referential and five affective. Two of the referential activities were aural, one of which contained five illustrations to facilitate comprehension. The Processing Instruction Treatment Package is shown in Appendix P Structured Input The web based instructional treatment package for the structured input groups did not contain any explicit grammar explanation or any information on processing strategies; otherwise, it was identical to the Processing Instruction Treatment Package. The treatment package for structured input only contained ten structured input activities (five referential and five affective), which were the same activities in the PI package. Two of the referential activities were delive red orally, and one of them co ntained illustrations to assist comprehension. The Structured Input Tre atment Package is presented in A ppe ndix Q. Visual Input Enhancement Visual input enhancement was operationalized as word animation of the targeted grammatical forms in the present stu dy. Two of the groups received visual input enhancement: (a) processing instruction with visual input enhancement (+PI +VIE), and (b) structured input with visual input enhancement (+SI +VIE). The treatment packages for these two groups contained ten ide ntical structured input activities: five referential and five affective activities. The five affective activities for these two groups contained visually enhanced input. In the present study, computerized visual input enhancement was operationalized as m ovement of the targeted verb forms, which grew larger and then

PAGE 212

198 worked online. The referential activities were not visually enhanced because participants had to cho ose a correct response (subjunctive or indicative verb form) in these activities, and visual enhancement of subjunctive forms might have prompted participants to select the enhanced forms even when they were not correct. With the affective structured inpu t activities, answers were not right or wrong. The purpose of the affective structured input activities was for learners to communicate their beliefs, opinions, or other types of affective responses as they processed real world information ( J.F. Lee & Van Patten, 2003). Thus, visual enhancement of the targeted grammatical forms through movement should have increased their perceptual salience and helped participants to notice the subjunctive forms in the input sentences of the affective structured input act ivities. Once targeted forms were noticed, the structured input activities in which the animations were embedded were designed to help learners correctly process these forms. Thus, computerized visual input enhancement was a good fit for the affective str uctured input activities. The same type of animation was used in all five affective structured input activities in order to maintain continuity for the two groups that received input enhancement. Although M. Allen (2003) suggests that stimulus novelty is short lived, if the same study, the results could have been confounded. This is especially true given that some types of animation, such as flash animation, could have a negative effect as learners must suppress the distraction of flash before they can process the other information that is on the screen (Hong, Thong, & Tam, 2004).

PAGE 213

199 Each animated verb form had a further enhancement on the verb ending, which is the most difficult feature for Spanish language learners to notice ( J.F. Lee, 1987; J.F. Lee & Rodrguez, 199 7) Each subjunctive ending appeared in a different color than the rest of the word for increased visual salience. Finally, the animations in the fi ve affective structured input activities were presented to learners sequentially rather than simultaneously because the simultaneous presentation of animated objects could have overwhelmed participants and detracted from the static information that was on the page (Sutcliffe & Namoune, 2007). Therefore, each sentence that contained an animated verb form was delivered to participants one at a time. Sentences that contained word animation were presented as external links in the treatment packages that conta ined visual input enhancement. When participants clicked on the link, the sentence that contained Computerized Visual Input Enhancement is presented in Appendix R In addition the Processing Instruction with Visual Input Enhancement Treatment Package is presented in Appendix S and the Structured Input with Visual Input Enhancement Treatment Package is presented in Appendix T Data Collection Procedures Duri ng Day1 of the exp eriment, students from three separate online classes were asked to participate in the study and to sign an informed consent form. At that time, it was explained that completing all of the study activities would take approximately three and one half hours of their time. Students were informed that they did not have to participate in the study and that their course grades would not be affected if they chose

PAGE 214

200 not to participate. Also, students who signed an informed consent form and agreed to participate in the study were informed that they could drop out of the study at any time without any penalties to their course grade. Students from Class 1 and Class 2 who participated in the study and completed all of the study activities received ten bonus points adde d to their lowest test grade and three bonus points added to their final average in their Spanish class. Students in Class 3 received two bonus points added to their final average and also had the opportunity to enter a raffle for cash prizes ($200 first prize, $100 second prize, and $50 third prize). Students who completed some of the study activities, but who dropped out of the study before finishing it, received ten bonus points added to their lowest test score. Students who did not wish to participat e but who wanted to earn the bonus points were given the option of completing an alternative pencil and paper instructional activity package with reading and writing activities that focused on the targeted grammatical form. The optional mat erials also req uired three and one half participate in the study opted for the alternative activity. Once the informed consent documents were signed, participants were asked to fil l out the Pretreatment Questionnaire. In order to maintain confidentiality, participants were assigned an identification number that was used on all of their study activities from that point on. In addition, informed consent forms were stored in a locked filing cabinet. Participants completed all other study activities online. The Informed Consent Form is presented in Appendix A.

PAGE 215

201 Part icipants were required to take the Subjunctive Knowledge Test as pretest, which was comprised of an Interpretation Subte st and a Production Subtest, the same day or the day after signing the informed consent form. After completing the pretest, participants were randomly assigned to groups. However, low achieving students were identified based on their test average in thei r Spanish class, and these participants were assigned to groups based on a stratified random assignment procedure. Low achievers were students whose test average was below 50% in their Spanish class. High achieving students were also identified in each c lass, however; only two participants that were identified as high achievers opted to participate in the study. High achievers were students whose test average was higher than 90% in their Spanish class. The two high achieving participants were randomly a ssigned to groups in the present study. Further, participants who scored 60% or higher on either the Interpretation Subtest or the P roduction Subtest of the Subjunctive Knowledge Test that was administered as a pretest were excluded from the study as they already demonstrated a basic understanding of the use of the subjunctive in adjectival clauses in Spanish. The c ut off level at pretest was in keeping with past studies in the PI strand. Participants who scored 60% or higher may have been exposed to the subjunctive if they took Spanish for more than two years in high school. The re were three forms of the Subjunctive Knowledge Test, which were delivered as a pretest and two posttests in the present study. Table 3.10 provides a visual display of the orde r in which the tests were presented to the three classes that participated in the study.

PAGE 216

202 Table 3.10 Order of Test Delivery by Class Class Pretest Posttest 1 Posttest 2 Class 1 Version A Version B Version C Class 2 Version B Version C Version A Class 3 Version C Version A Version B The instructional treatment packages were available online through Blackboard Courseware Ma nagement System (CMS) two to three days after participants co mpleted the pretest. Th e instructional treatment packages were available for a period five days on Blackboard CMS As this was a web based study, participants were able complete all of the study activities online any time of the day or night that was convenient for their schedu les. In addition, Blackbo ard CMS recorded the time that each participant took to complete the study activities. Blackboard also provided a date stamp for each activity. Participants were asked to spend a minimum of one hour on the ir instructional treatm ent package s but they had up to two hours if needed. The amount of time needed to complete the instructional treatment packages was determined based on the results of pilot testing. The researcher collected information on how much time each participant spent on the instructional treatments. In addition, the researcher checked each cursory way. For example, if it was evident that a participant did not follow the

PAGE 217

203 dire ctions for the activities, exhibited random clicking, or spent significantly less than the minimum amount of time that was required, then he or she was excluded from the study. Participants were told that they had to follow the instructions carefully, spe nd at least the minimum amount of time required, and answer all of the questions completely (without leaving blanks) in order to receive the extra credit in their course. The extra credit offered was a strong incentive for participants, and very few appro ache d the tasks in a cursory way. Only t wo students were excluded from the study for approaching their instructional treatment packages in a superficial way. These students spent under 15 minutes on their grammar packages and performed poorly on all of a lmost all of the items, which was likely the result of random clicking. Participants received the following in structions for completing the study activities : After you complete your pretest, the researcher will send you a password for your grammar activ ity package. Please note the deadlines for completing this activity package on Blackboard. You will receive daily emails reminding you to begin your activity package until it is completed. You will need to allow up to a two hour block of time to complete your grammar activity package, which is followed by a short test. Once you open your activity package, please complete all of the activities at one time. If you encounter an emergency and you must discontinue the activity package, please contact the rese archer as soon as possible at vrussell@mail.usf.edu for further instructions. Once your activity package Participants were asked to complete their activity packages on t he same day that they opened them. If a participant encountered an emergency and had to stop the instructional treatment, he or she was asked to contact the researcher as soon as possible. If the participant had been working on the treatment activities f or less than 30 minutes, then he

PAGE 218

204 or she was allowed to resume the treatment from the beginning on another day of the experiment. If, however, the participant had been working on the treatment activities for more than 30 minutes before stopping the treatme nt, then he or she was no longer able to participate in the study. In order to keep the treatments homogeneous across groups, it was important for all participants to complete their instructional treatments in one day. Only one participant had to stop wo rking on her instructional treatment before finishing it due to an emergency, and since she only spent 15 minutes working before having to stop, she was permitted to restart her grammar package the next day. Particip ants were required to complete P osttest 1 immediately following their instructional treatments. Posttest 1 was timed, and participants had up to 30 minutes to complete this test. The amo unt of time needed to complete P osttest 1 was determined during p ilot testing. Once completed, participant P osttest 1 were stored on Blackboard and printed by the researcher. One to three days after completing thei r instructional treatment packages and P osttest 1, participants completed three online activities as follows: (a) the Authentic I npu t Reading Activity, (b) t he Comprehension Test, and (c) t he Posttreatment Questionnaire. Participants were required to complete the Comprehension Test immediately following the Authentic Input Reading Activity. Similarly, they were also required to compl ete the Posttreatment Questionnaire immediately following the Comprehension Test. All three activities were timed on Blackboard, and they were also time and date stamped upon completion. The researcher checked to ensure that participants completed the ac tivities in the appropriate order.

PAGE 219

205 While striving to control time on task between the treatment groups, the data collection procedures employed in the present study attempted to allow participants some flexibility when completing the activities online. D istance learners are accustomed to working asynchronously, during the days and times that suit their individual schedules. Thus, the study protocol did not require participa nts to spend more than two and one half hours at any given time on the study relat ed activities, and participants were able to choose when they opened the activities provided that they were within the appropriate date range (e.g. the instructional treatment packages could be opened anywhere from two to six days following the completion of the pretest). Participants received an email message with the following instructions after they submitted their instructional treatment activities: Thank you for completing yo ur online activity package and P osttest 1! You will now have up to three day s to complete the following activities: 1) A Reading Activity where you will take some notes while you read a passage in Spanish (this should take only 10 15 minutes), 2) A Reading Comprehension Test (to be completed immediately after the reading activity; this should take about 15 20 minutes to complete), and 3) A Posttreatment Questionnaire (to be completed immediately after the Reading Comprehension Test; this will also take about 10 15 minutes to complete). You may begin these activities starting tomorr ow. If you have any questions or concerns, please contact the researcher at vrussell@mail.usf.edu Thank you for giving your best effort to comply with these instructions! The Authentic Input Reading Text, th e Comprehension Test, and the Posttreatment Questionnaire were delivered via Blackboar d CMS After participants completed and submitted the s e items, their responses were stored on Blackboard and

PAGE 220

206 printed by the researcher. Further, the researcher checked over these activities to ensure that participants did not leave any items blank and that they followed the instructions. Finally, participants were asked to take P osttest 2 about two weeks after they completed their instructional treatment packages. T he researcher made the delayed posttest available fourteen days after the first participant in each class completed the instructional treatment package; however, the majority of participants waited until the las t day that the instructional treatment packag es were available on Blackboard to complete them. Similarly, most participants also waited until the last day that P osttest 2 was available on Blackboard to complete it. Thus, most participants took P osttest 2 thirteen da ys after P osttest1. Posttest 2 w as also delivered via Blackboard, and the researcher printed it once it was completed and submitted by each participant. In addition, all directions that were emailed to participants were also posted as permanent announcements on Blackboard. The experime ntal sch edule is presented in Figure 3.4 Data Analysis All data was analyzed using SAS 9.1 for Windows software. Data were screened for outliers prior to running any statistical tests. In addition, the researcher checked to ensure that the underlyin g assumptions for each statistical test that was employed in the present study. The procedures that were used to analyze each research question are described below. In order to answer the first two research questions, which are listed below, data were su bjected to two repeated measures Analyses of Variance (ANOVA s ) with one between subjects factor (type of instruction) and one w ithin subjects factor ( time). Type

PAGE 221

207 Figure 3.4 Experimental S chedule Day 1: Face to Face Day 4 8: Online Posttest 1 was completed immediately after the instructional treatments +PI VIE / Po sttest 1 Day 5 11: Onlin e 1 3 d ays a fter the instructional t reatments Pretreatment Questionnaire +PI +VIE / Posttest 1 +SI VIE / Posttest 1 +SI +VIE / Posttest 1 +TI / Posttest 1 Reading Activity (note sheets) Comprehension Test Posttreatment Questionnaire Posttest 2 Day 18 21: Online 10 17 d ays after the i nstructional t reatments Pretest Day 1 or 2: Online Informed Consent

PAGE 222

208 of instruction had five levels: processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structur ed input without visual input enhancement, and traditional instruction. The within subjects fa ctor, time, had three levels (Pretest, Posttest 1, and Posttest 2). A separate analysis was performed for each type of task, interpretation or production. 1. Is t here a differential performance between treatment groups for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks over time ? 2. Is there a differential performance between treatment groups for the a cquisition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks over time? Each repeated measures ANOVA examined the within subjects effects for time, the between subjects effect for t ype of i nstruction, and the pos sible interaction effect between time and type of instruction. In addition, if significant main or interaction effects were revealed by the repeated measures ANOVAs, post hoc procedures were performed to determine which specific treatments differed from e ach other. If any of the ANOVAs were found to be significant, follow up Tukey tests were performed to examine all pairwise comparisons in order to determine which groups had significant differences. The researcher reported the following descriptive stati stics: group means, standard deviations, skewness values, and kurtosis values for scores on the Interpretation and Production Subtests by group. In addition, the researcher reported the F and p values, the

PAGE 223

209 sums of squares, the mean squares, and the degree s of freedom Any statistically significant interaction effects were graphed, and any significant differences by group revealed by follow up Tukey tests were reported. Research Q uest ion T hree, which is listed below, to notice t he targeted grammatical form when it was embedded in an authentic input text that was received subsequent to the instructional treatments. 3. ability to notice targeted forms in sub sequent authentic input as measured by note scores and awareness scores? To address the question of how well do participants notice the targeted form that is embedded in an authentic input text, their note scores, which measured the amount of noticing th at took place, and their awareness scores, which measured the depth of noticing that took place, were submitted to a multivariate analysis of variance (MANOVA) with alpha set at .05 to determine if there were significant differences by treatment group. Th e dependent variables were note scores and awareness scores. The independent variable was type of instruction with five levels: processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input with visual input enhancement, structured input without visual input enhancement, and traditional instruction. If significant effects were found o n the MANOVA, follow up ANOVAs were performed on each dependent variable with alpha set at .05. If the F va lue was found to be significant with the follow up ANOVAs, post hoc Tukey tests were performed to examine all pairwise comparisons in order to

PAGE 224

210 determine which groups had significant differences. The researcher reported the group means and standard deviati ons. Also, ANOVA summary tables were constructed with the F and p values, the sums of squares, the mean squares, and the degrees of freedom. In addition, any significant group differences revealed by the post hoc Tukey tests were reported by the research er. Research Question F our, which is listed below, lity to comprehend the targeted grammatical form when it was embedded in an authentic input text that was received subsequent to the instructional treatments. 4. Following the instructi onal treatments, is there a differential performance between the targeted grammatical form (input processing) and the message of the authentic input text in which it is embe dded as measured by grammar comprehension and text comprehension scores? passage that they received subsequent to completing their instructional treatments. The purpose of the Comprehension Test was to determine if the treatments, which utilized structured or manipulated input, were able to improve how participants processed the targeted grammatical form when they encountered it in its natural TL context (authentic input). The Comprehension Test yielded two scores per participant, one score for comprehension of the propositional content of the text (text comprehension) and another score for comprehension of the referential meaning of the targeted grammatical form (input proc essing). Text and grammar comprehension scores were submitted to a

PAGE 225

211 MANOVA with alpha set at .05 to determine if there were significant differences by treatment group. If statistically significant differences were found with the MANOVA, follow up ANOVAs w ere performed on each dependent variable In addition, if the F value was found to be significant with the follow up ANOVAs, then post hoc Tukey tests were performed to examine all pairwise comparisons in order to determine which groups differed from each other The researcher reported the following descriptive statistics: group means, standard deviations, skewness values, and kurtosis values for text and grammar comprehension scores by group. In addition, for the overall MANOVA test, the researcher repo p value associated with it. For the follow up ANOVAs, the researcher reported the following for each independent variable, the F and p values, the sums of squares, the mean squares, and the degrees of freedom. Further, any sig nificant group differences revealed by the post hoc Tukey tests were reported. Research Question 5, addressed the relationship between text comprehension and input processing. 5. What is the relationship between text comprehension and input processing when learners encounter the targeted form in subsequent authentic input? In order to answer this question, a Pearson Product Moment Correlation Coefficient was computed using on the two scores that the Comprehension tests yielded: the grammar comprehension sco re and the text comprehension score. The Pearson r determined the magnitude of the relationship between message comprehension and grammar comprehension for each instructional treatment group (processing instruction with visual input enhancement, processin g instruction without visual input enhancement, structured

PAGE 226

212 input with visual input enhancement, structured input without visual input enhancement, and traditional instruction). The Pearson r value was reported for each group.

PAGE 227

213 Chapter 4 Results Introducti on This chapter presents the results of the major analyses conducted on the data that was obtained from participants before, during, and after the instructional treatments. Major analyses were conducted on pre and posttest scores, note scores, awareness responses on the Pretreatment and Posttreatment Questionnaires are also provided. The chapter is divided as follows: ( a Pretreatment Questionnaire, ( b ) analyses of pre and posttest scores from the Interpretation and Production Subtests, ( c ) analyses of note and awareness scores, ( d ) analyses of grammar and text comprehension scores, and ( e ses on the Posttreatment Questionnaire. Summary of the Pretreatment Questionnaire The Pretreatment Questionnaire was divided into three major sections as follows: demographic information, language background information, and perceptions on learning a lang

PAGE 228

214 Table 4.1 Participant Background Information from the Pretreatment Questionnaire +PI VIE +PI+VIE +SI VIE +SI+VIE +TI n = 19 n = 18 n = 19 n = 18 n = 18 Gender d istribution Male (8) Female (11) Male (3) Female (15) Male (4) Female (15) Male (6) Female (12) Male (8) Female (10) Age range Mean SD 20 45 27.58 7.74 19 42 25.78 6.57 19 40 25.79 6.40 19 45 26.72 7.93 19 39 23.72 5.51 First lang uage English (18) Singhalese (1) English (18) English (19) English (18) English (18) Home language English (18) Singhalese (1) English (18) English (19) English (18) English (18) Number of semesters of college Spanish Four (0) Three (1) Two (18) Four (1) Three (1) Two (16) Four (0) Three (0) Two (19) Four (0) Three (1) Two (17) Four (0) Three (0) Two (18) Number of years of high school Spanish Four (2) Three (1) Two (5) One (6) None (5) Four (1) Three (5) Two (8) One (2) None (2) Fo ur (2) Three (4) Two (7) One (2) None (4) Four (1) Three (2) Two (6) One (2) None (7) Four (0) Three (2) Two (10) One (1) None (5) Why elected to study a language online Convenience (16) Enjoy computers (0) Other (3) Convenience (16) Enjoy computer s (0) Other (2) Convenience (17) Enjoy computers (1) Other (1) Convenience (18) Enjoy computers (0) Other (0) Convenience (17) Enjoy computers (1) Other (0) Computer skills High (12) Fair (7) Poor (0) High (9) Fair (9) Poor (0) High (8) Fair (11) P oor (0) High (6) Fair (11) Poor (1) High (14) Fair (4) Poor (0) Ease of using blackboard CMS Easy (19) Moderate (0) Difficult (0) Easy (18) Moderate (0) Difficult (0) Easy (18) Moderate (1) Difficult (0) Easy (18) Moderate (0) Difficult (0) Easy (1 5) Moderate (1) Difficult (2) Would take another language class online Yes (7) No (11) Maybe (1) Yes (8) No (10) Maybe (0) Yes (11) No (8) Maybe (0) Yes (12) No (5) Maybe (1) Yes (11) No (7) Maybe (0) Note N = 92.

PAGE 229

215 As is shown in the p articipant sample (n = 92), there were 29 males (31.52% ) a nd 63 females (68.48% ) Participants ranged in age from 19 to 45, with a mean age of 25.94 and a standard deviation of 6.86. All of the participants spoke English at home, but one participant al so spoke Singhalese at home. A total of 88 participants (95.65%) were at the end of their second semester of Spanish at the university level, and 4 participants (4.45%) indicated that they had completed more than two semesters of Spanish at the university level. A closer inspection of the Pretreatment Questionnaire revea led that all four of these participants failed eith er their first or second semest er of Spanish at the university level, which had to be repeated. None of the participants had advanced be yond their second semester of Spanish language study at the university level. With respect to the number of years of high school Spanish, 23 participants (25%) indicated that they had never taken Spanish in high school, 13 (14.13%) indicated that they had taken one year of Spanish in high school, 36 (39.13%) indicated that they took two years of high school Spanish, and 20 (21.74%) indicated that they took three or four years of high school Spanish. It is important to note that at both the large urban uni versity and the small suburban university Spanish language students who completed more than two years of high school Spanish were required to take a placement exam before enrolling in a Spanish course. If students were placed in second semester Spanish o r lower, they did not demonstrate a sufficient understanding of the subjunctive mood in Spanish to warrant placement in a

PAGE 230

216 higher level of Spanish, even though it is likely that they had exposure to the subjunctive mood in high school. Regarding why stu dents elected to take Spanish online, 84 participants (91.31%) listed convenience as the main reason that they chose to study Spanish online. Only 2 participants (2.17%) stated that they elected to take Spanish online because they enjoyed using computers and the Internet. However, 6 participants (6.52%) stated that they took Spanish online for other reasons, which included having a disability. Participants were aske d to rate their computer skills and the responses broke down as follows: 49 participants ( 53.26%) rated their computer skills as high, 42 (45.65%) rated their computer skills as fair, and 1 participant (1.09%) rated her computer skills as poor. Participants were also asked to rate the ease of using Blackboard Courseware Management System to ac cess and complete assignments, and they responded as follows: 88 participants (95.65%) responded that Blackboard was easy to use, 2 participants (2.17%) responded that it was moderate to use, and 2 participants (2.17%) responded that Blackboard was difficu lt to use to access and complete assignments. Finally, participants were asked whether they would take another language class online. In response, 49 participants (53.26%) indicated that they would take another language class online, while 41 (44.57%) in dicated that they would not take another language class online. In a third category, 2 participants (2.17%) were undecided and stated that they might take another Questionna ire by treatment group is provided in Table 4.1.

PAGE 231

217 Time on Task Before examining the inferential statistical procedures presented in this chapter, descriptive statistics are provided for the amount of time that participants spent completing their instructio nal treatment packages online. There were five types of instruction examined in the present study: processing instruction with visual input enhancement, processing instruction without visual input enhancement, structured input without visual input enhance ment, structured input with visual input enhancement, and traditional instruction. The participants assigned to the first four instructional types compr ised the experimental groups while the participants assigned to the last instructional type served as a comparison group. The amount of time that participants spent completing their instructional treatment packages was recorded on Blackboard. For the 92 participants in the present study, the mean time it took to complete an instructional treatment package online was 75.15 minutes with a standard deviation of 19.62. The minimum amount of time spent was 33 minutes, and the maximum amount of time spent was 126 minutes. Table 4.2 provides the descriptive statistics on the amount of time that participants spe nt completing their instructional treatments online by group. Participants in the traditional instruction group (TI) spent the most time (M = 85.38 min.) completing their instructional treatment packages, approximately 10 minutes more than the mean time fo r all participants (75.15 minutes), while participants in the structured input with visual input enhancement group (+SI +VIE) spent the least amount of time (M = 69.56 min) completing their instructional treatment packages,

PAGE 232

218 Table 4.2 Time Spent on Task by Instructional Group Time in Minutes Group n Mean SD Min Max + PI VIE 19 76.68 17.46 59 126 + PI +VIE 18 73.66 20 .00 42 120 + SI VIE 19 70.79 18.6 0 33 110 + SI +VIE 18 69.56 12.42 52 106 + TI 18 85.38 25. 46 40 120 Overall 92 75.15 19.62 33 126 approximately 5 minutes less than the mean time for all participants. In order to determine if the mean difference in time spent completing the instructional treatments by group was statistically signifi cant, the amount of time participants in each group spent on task was submitted to a one way analysis of variance (ANOVA) with the amount of time as the dependent variable and type of instruction as the independent variable. The results revealed that ther e were no statistically significant differences between the groups for time spent on task, F (4, 87) = 1.96, p > .05. Analyses of Pre and Posttests A repeated measures ANOVA with one between subjects factor, type of instruction, and one within subjects f actor, time with three levels (Pretest, Posttest 1,

PAGE 233

219 P osttest 2), was conducted for each of the four constructs that the pre and posttests assessed; namely, interpretation of the subjunctive, production of the subjunctive, interpretation of the indicative, and production of the indicative to determine if there was equivalence between groups on each of these measures The four analyses are presented below. Establishing Pretreatment Equivalence of Groups Before conducting the analyses to examine the effects of the instructional treatments by group over time, scores from the interpretation of the subj unctive component of the Subjunctive Knowledge Test and scores from the production of the subjunctive component of the Subjunctive Knowledge Test were submitted to two one way ANO VAs to determine if there were group differences prior to the experiment on Spanish. The ANOVA that examined pretest scores for the interpretation of the subjunctive by group revealed no significant group differences prior to the experiment, F (4, 87) = 0.73, p > .05. Similarly, the ANOVA that examined pretest scores for the production of the subjunctive by group also revealed no significant group dif ferences at pretest, F (4, 87) = 0.24, p > .05. Analysis of Scores for Interpretation of the Subjunctive post and delayed posttest scores on the Inter pretation Subtest of the Subjunctive Knowledge Test were analyzed using two repeated measures ANOVAs with one between subjects factor (type of instruction) and one within subjects factor (time) The within subjects factor had three levels: Pretest, Posttest 1, and Posttest

PAGE 234

220 2. One repeated measures ANOVA analyzed ation of the interpretation of the subjunctive in adjectival clauses and 5 of the items measured cative. The latter analysis was included in the present study to examine the possibility of learner overextension of the targeted grammatical form. The highest score possible for interpretation of the subjunctive was 15. Since the P retest was used as a screening device, only participants who scored 8 (53.33%) or below for the interpretation of the subjunctive component of the Interpretation Subtest were invited to participate in the study. Thus, participants were excluded from the study if they scored 9 (60%) or higher. The 60% cutoff level was employed in order for the results of the present study to be comparable with past research the subjunctive component of the Interpretation Subtest are included in Table 4.3. An examination of the table of means reveals that mean scores for all groups on the Pretest appear to be similar. The traditional instruction group had the highest mean score on the Pretest (7.2 2) while the processing instruction with visual input enhancement group had the lowest mean score on the Pretest (6.50). The differences in mean scores for Posttest 1, however, appear to be further apart. The processing instruction with visual input enha ncement group had the highest mean score (12.50) while the structured input

PAGE 235

221 without visual input enhancement group had the lowest mean score (9.89). On Posttest 2, the processing instruction with visual input enhancement group had the highest mean Table 4.3 Descriptive Statistics for Scores on the Interpretation Subtest by Group at Pretest Posttest 1, and Posttest 2 for Interpretation of the Subjunctive Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + PI VIE 19 M 6.53 11.26 10.16 SD 1.9 0 2.35 3.04 sk 0.97 0.5 0 0.4 0 ku 0.44 0.26 0.29 + PI +VIE 18 M 6.5 0 12.5 0 10.28 SD 1.5 0 1.95 2.89 sk 0.64 0.81 0.48 ku 0.24 0.48 0.13 + SI VIE 19 M 6.79 9.89 9.53 SD 1.23 2.9 0 3.13 sk 0.76 0.41 0.01 ku 0.24 1 .00 0.19 + SI +VIE 18 M 6.83 10.28 9.89 SD 1.47 3.04 2.89 sk 1.19 0.61 0.57 ku 1.06 0.09 0.84

PAGE 236

222 Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + TI 18 M 7.22 10.56 8.89 SD 1.06 2.94 3.31 sk 1.16 0.06 0.04 ku 0.16 1.47 0.62 Overall 92 M 6.77 10.89 9.75 SD 1.45 2.77 3.03 sk 1.04 0.38 0.1 0 ku 0.3 0 0.75 0.53 score (10.28) while the traditional instruction group had the lowest mean score (8.8 9). To determine if there were significant differences in group means over time interpretation test scores were submitted to a re peated measures ANOVA with one between subjects factor (type of instruction) and one within subjects factor (time of testing), which had three levels: Pretest, Posttest 1 and Posttest 2. Before proceedin g with the statistical test, univariate normality a nd sphericity assumptions underlying factorial ANOVA with repeated measures factors and between subjects factors were checked; namely, independence, random sampling, univari ate normality, and sphericity. T he repeated measures test allows for data to be co llected from participants at multiple points in time under the within subjects variable as subject differences are removed from the error term, which leaves error components independent from treatment group to treatment group.

PAGE 237

223 Univariate rather than multi variate normality was assessed because the present analysis took a univariate approach, examining the results of the ANOVA test rather than the results generated from the MANOVA test. In order to assess univariate normality, the distributions of interpret ation test scores were examined to assess skewness and kurtosis at each level of time by group. For the P retest, skewness values ranged from 1.19 to .64 and kurtosis values ranged from .44 to 1.06. For P osttest 1, skewness values ranged from 81 to .41 and kurtosis values ranged from 1.47 to .48. The distributions of interpretation test scores for P osttest 2 by group w ere fairly normally distributed with skewness values ranging from .48 to .57 and kurtosis values ranging from .84 to .19. I n addition to the examination o f skewness and kurtosis values the Shapiro Wilk test for normality was performed on each dependent variable by group The Shapiro Wilk tests revealed that the assumption of normality was met for the distributions of interpr etation test s cores for all of the groups on Posttest 1 and P osttest 2, as p values were all in excess of .05. H owever, the Shapiro Wilk test revealed that the assumption of normality was violated for the distribution s o f interpretation test scores on the P retest for all four experimental groups and for the comparison group The non normal distribution of scores was likely due to the fact that participants w ho scored higher than 8 on the P retest were excluded from the study. Thus, the P retest did not refl ect any scores higher than 8 even though the highest score possible for the interpretation of th e subjunctive component of the P retest was 15. In addition, t he data were checked for outliers by examining box plots for each dependent variable by instructio nal group. The box plots

PAGE 238

224 revealed that there were no significant outliers in the data set. Although the assumption of univariate normality was partially violated, the ANOVA test is fairly ro bust to violations of normality. Given the robustness of the te st to violations of normality, it seemed reasonable to proceed with the analysis. The final assumption that was checked was sphericity. In order to assess the assumption of sphericity, the Greenhouse Geisser estimator, which is a fairly conservative test was used. Sphericity assumes that the difference variables have equal scores at three points in time (Pretest, Posttest 1, and Posttest 2) the estimate for sphericit y could have ranged from .5 to 1, with 1 being an ideal estimate of sphericity. The Greenhouse Geisser estimate was = .98 which is nearly an ide al estimate of sphericity. Since the Greenhouse Geisser estim ator is a conservative test that tends to underestimate the sphericity parameter, it is likely that the actual va lue was even slightly higher. Thus, t he results o f the Gree nhouse Geisser test verified that the a ssumption of sphericity was met in the present study. After the assumptions were assessed, the data were subjected to a repeated measures analys is of variance (ANOVA) with one between subjects factor (type of instruction) and one within subjects factor (time) to determine if there were significant differences in i nterpretation test scores across time (from pre to p osttests). The results of the analysis are reported in Table 4.4.

PAGE 239

225 Table 4.4 Analysis of Va riance of Interpretation Test Scores by Instructional Treatment Group and Time for Interpretation of the Subjunctive Source df SS MS F Between Subjects (Ss) 91 932.10 Group (A) 4 36.20 9.05 0.88 S/A 87 89 5.90 10.30 Within Subjects 184 1 625.00 Time (B) 2 832.73 416.36 100.26** A x B 8 69.71 8.71 2.10* SB/A 174 722.56 4.15 Total 275 2 557.10 No te N = 92. p < .05 ** p < .0001 An examination of Table 4.4 reveals a significant Group x Time interaction effect, F (8, 174) = 2.10, p < .05. The Greenhouse Geisser test also showed a significant effect ( p < .05). There was also a significant main effect for time, F (2, 174) = 100.26, p < .0001, but there was not a significant main effect for type of instruction, F (4, 87) = 0.88, p > .05. The effect size for the Group x Time interaction effect was computed, = .09, which was a small effect size. The effect size for the main effect for time was also

PAGE 240

226 computed, = .54, which was a very large effect size. A graph of the interaction effect is shown in Figure 4.1. The interaction is disordinal. Figure 4 .1 Interaction Plot for Instruction Type and Time for Interpretation of the S ubjunctive In order to determine which groups had significant differences in mean interpretation test scores over time, post hoc comparisons of mean interpretation test scores us instruction with visual input enhancement group had a significantly higher mean interpretation test score than the structured input without visual input enhancement group at Posttest 1 c ompared to Pretest (p < .05). However, the processing instruction with visual input enhancement group did not have a significantly higher mean interpretati on test score than the processing instruction without visual input enhancement group, the structured input with visual input

PAGE 241

227 enhancement group, or the traditional i nstruction group at P osttest 1 compared to Pretest ificant group differences from Posttest 1 to Posttest 2 or from Pretest to P osttest 2. Ste vens (2002) asserts that the Tukey test is appropriate in repeated measures designs and that alpha is controlled for the set of tests if the sphericity assumption is met and there are equal or nearly equal group sizes. The repeated measures ANOVA also re vealed a highly significant main effect for time. A graph of the significant main effect for time is presented in Figure 4.2. Figure 4.2 Graph of the Main Effect for Time for the Mean Interpretation Test S core at Pretest, Posttest 1, and Posttest 2 for Interpretation of the S ubjunctive P ost hoc contrast tests were performed to determine i f changes in the mean interpretation test score were significant at each point in time. The contrast tests r evealed that the mean interpretation test score was signifi cantly higher at Posttest 1 compared to Pretest, F (1, 87) = 182.86, p < .0001. T he mean interpretation test score was also

PAGE 242

228 significantly lower at Posttest 2 compared to Posttest 1, F (1, 87) = 17.25, p < .000 1. However, the mean interpretation test scor e was still significantly higher at Posttest 2 compared to Pretest, F (1, 87) = 87.30, p < .0001. In order to control for the Type I error rate, the Bonferroni adjustment was applied, with alpha set at .0167 for the set of post hoc contrast tests. Analysi s of Scores for Production of the Subjunctive Partic re and p osttest scores on the P roduction Subtest of the Subjunctive Knowledge Test were analyzed using two repeated measures ANOVAs with one between subjects factor (type of instruction) and on e within subjects factor (time of testing ) Production Subtest consisted of 20 items, 15 of the item production of the subjunctive in adjectival clauses and 5 of the items measured ve, while a separate analysis the present study to examine the possibility of learner overextension of the targeted grammatical form. The highest score possible f or production of the subjunctive was 15. Since the P retest was used as a screening device, only participants who scored 8 (53.33%) or below for the production of the subjunctive component of the Production Subtest were invited to participate in the study. Similar to the Interpretation Subtest, participants who scored 9 (60%) or above on the production of the subj unctive

PAGE 243

229 component of the Production Subtest were excluded from the study, which is in line with past studies in the PI strand. Table 4.5 present s the scores on the production of the subjunctive component of the Production S ubtest Table 4.5 Descriptive Statistics for Scores on the Production Subtest by Group at Pretest, Posttest 1, and Posttest 2 for Pro duction of the Subjunctive Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + PI VIE 19 M 1.95 10.21 7.66 SD 2.2 0 2.96 3.61 sk 1.24 1.45 0.26 ku 0.77 3.11 1.51 + PI +VIE 18 M 2.72 9.28 8.14 SD 2.4 0 4.17 3.71 sk 0.41 0.71 0.44 ku 1.53 0.62 0.31 + SI VIE 19 M 2.47 7.37 7.16 SD 3.44 3.72 3.94 sk 0.9 0 0.07 0.14 ku 1.11 1.18 0.91 + SI +VIE 18 M 2.25 7.86 7.78 SD 2.69 2.92 3.3 0 sk 1.21 0.17 0.53 ku 0.49 0.88 0.55

PAGE 244

230 Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + TI 18 M 2.58 10.08 8.47 SD 2.43 3.76 4.35 sk 0.78 0.14 0.16 ku 0.14 1.46 1.52 Overall 92 M 2.39 8.96 7.83 SD 2.63 3.65 3.74 sk 0.9 0 0.35 0.22 ku 0.48 0.8 0 1.01 A visual examination of the table of means reveals that mea n scores for all groups on the P retest appear to be similar. The processing instruction with visual input enhancement group had the highest mean score on the P retest (2.72), while the processing instruction without visual input enhancemen t had the lowest mean score at P retest (1.95). However the mean scores on P osttest 1 appear to be significantly higher than the mean scores earned on the P retest for all groups. The processing instruction without visual input enhancement group had the highest mean score (10.21), and the structured input wit hout visual input enhancement group had t he lowest mean score (7.37) on P osttest 1 for produ ction of the subjunctive. For P osttest 2, the scores for all of the groups appear to be similar. The traditional instruction group had the highest mean score on P osttest 2 (8.47), while the structured input without visual input enhancement

PAGE 245

231 grou p had the lowest mean score on P osttest 2 for production of the subjunctive (7.16). In order to determine if mean scores differed significantly by group over time, participa Subtest were submitted to a r epeated measures ANOVA with one between subjects factor (type of instruction) and one within subjects factor (time of testing) Before the data were subject ed to the statistical test, univariate normality and sphericity assumptions for factorial ANOVA with repeated measures factors and between s ubjects factors were assessed. In order to assess univariate normality, the distributions of interpretation test scores were examined to check the skewness and kurtosis values at each level of time by group. For production test s cores on the P retest, most of the distributions appear to be approximately normally distributed, with most skewness and kurtosis value s less than 1. For Posttest 1, skewness values ranged from 1.45 to .07 and kurtosis values ranged from 1.46 to 3.11. For Posttest 2, the skewness values ranged from .53 to .14 and the kurtosis values ranged from 1.51 to .31. Further, the data were also checked for outliers by examining box plots for each dependent variable by instructional treatment group. The box plots revealed that there were no significant outliers. In addition to the examination of skewness and kurtosis values, the Shapi ro Wilk test for normality was performed on each dependent variable by group. The Shapiro Wilk tests revealed that the assumption of normality was met for the distributions of production test s cores for all of the groups on P osttest 2 ( p > .05). T he Shap iro Wilk test also revealed that the assumption of normality was met for the distributions of production

PAGE 246

232 test scores on Posttest 1 for all groups except for the processing instruction without visual input enhancement group (p < .05). F or the distribution of production test scores on the P retest, the Shapiro Wilk test revealed that the assumption of normality was violated by all f our of the experimental groups ( p > .05 ) However, the Shapiro Wilk test revealed that production test scores were normally distr ibuted for the tradit ional instruction group on the Pretest The non normal distribution of scores for the f our experimental groups on the P retest was likely due to the fact that participants who scored h igher than 8 on the P retest were excluded from the s tudy. Thus, even though the highest score possible for the production of the subjunctive component of the Production Subtest was 15, the P retest did not reflect a ny scores higher than 8. Another assumption of the repeated measures ANOVA test is spherici ty. In order to assess the assumption of sphericity, the Greenhouse Geisser estimator was used. Sphericity assumes that the difference variables have equal variances and that they do not covary The Greenhouse Geisser estimate was = .89 Since this v alue was close to 1, which is an ideal estimate, the assumption of sphericity was met in the present study. It is also likely that the actual estimate for sphericity was slightly higher since the Greenhouse Geisser estimate for the sphericity parameter is somewhat conservative. After assessing all of the assumptions of the statistical test, it seemed reasonable to proceed with the analysis as the repeated measures ANOVA is robust to violations of normality The results of the repeated measures analysi s on production test scores at Pretest, Posttest 1, and Posttest 2 are presented in Table 4.6.

PAGE 247

233 Table 4.6 Analysis of Variance of Production Test Scores by Instructional Treatment Group and Time for Production of the Subjunctive Source df SS MS F Between S ubjects (Ss) 91 1,597.76 Group (A) 4 71.20 17.80 1.01 S/A 87 1,526.56 17.55 Within S ubjects 184 3,783.02 Time (B) 2 2,267.91 1,133.96 137.31* A x B 8 78.11 9.76 1.18 SB/A 174 1,437.00 8.26 Total 275 5,380.78 Note N = 92. p < .0001 The results revealed that t here was no t a significant Group x Time interaction effect, F (8, 174) = 1.18, p > .05; however, there was a highly significant main effect for time, F (2, 174) = 137.31, p < .0001 The Greenhouse Geisser test also showed a significant effect for time, p < .0001. Further, the main effect for type of instruction was not significant F (4, 87) = 1.01, p > .05. The effect size was computed for the significant

PAGE 248

234 main effect for time, = .61, which was a very large effect size. A graph of the s ignificant main effect for t ime is presented in Figure 4.3 Figure 4.3 Graph of the Main Effect for T ime for the M ean Production Test S core at Pretest, Posttest 1, and Posttest 2 for Production of the S ubjunctive As a follow up to the significant main effect for tim e, post hoc contrast tests wer e performed to determine if changes in the mean production test score at each point in time were s tatistically significant. The contrast tests revealed that the mean production test score was significantly higher at P osttest 1 compared to P retest, F (1, 87) = 212.11, p < .0001. In addition, there was a stat istically significant decrease in the mean production test score from Posttest 1 to Posttest 2, F (1, 87) = 10.63, p < .001. However the mea n production test score was si gnificantly higher at Posttest 2 compared to P retest, F (1, 87) = 136.05, p < .0001. In ord er to control for the Type I error rate, the B onferroni adjustment was applied, with alpha s et at .0167 for the set of post hoc contrast tests.

PAGE 249

235 Analysis of Scores for Interpretation of the Indicative and p osttest scores on the Inter pretation Subtest of the Subjunctive Knowledge Test were analyzed using two repeated measures ANOVAs with one between subjects factor (type of instruction) and one wi thin subjects factor (time of testing) Interpretation Subtest consisted of 20 items, 15 of interpretation of the subjunctive in adjectival clauses and 5 of the items measured ion of the indicative, which was included in the present study to examine the possibility of learner overextension of the targeted grammatical form. Participants in the present study already knew how to form and use the indicative mood in Spanish; however the instructional treatments focused on subjunctive versus indicative contrasts. Therefore, it was important to determine if the knowledge of the use of the indicative mood in Spanish. The interpretation of the indicative component of the Interpretation Subtest was not used as a screening device for exclusion from participation in the study; rather, scores measuring the interpretation of the indicative at Pretest serv ed as a baseline measure of highest score possible for interpretation of the indicative was 5. Table 4.7 provides t he

PAGE 250

236 es on the interpretation of the indicative component of the Interpretation S ubtest Table 4.7 Descriptive Statistics for Scores on the Interpretation Subtest by Group at Pretest, Posttest 1, and Posttest 2 for Interpretation of the Indicative Inst ructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + PI VIE 19 M 2.95 3.84 3.47 SD 0.85 1.01 1.07 sk 0.11 0.36 0.08 ku 1.62 0.91 1.16 + PI +VIE 18 M 3.39 4.22 3.44 SD 0.98 0. 88 1.2 0 sk 0.50 1.07 0.31 ku 1.03 0.87 0.66 + SI VIE 19 M 3.00 3.63 3.79 SD 0.82 1.12 0.98 sk 0.00 0.24 0.33 ku 1.48 1.24 0.73 + SI +VIE 18 M 3.50 3.33 3.56 SD 0.99 1.24 1.15 sk 0.00 0.93 0.55 ku 0.84 1.91 0.13

PAGE 251

237 Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + TI 18 M 3.17 3.72 3.56 SD 1.04 1.07 1.25 sk 0.02 0.01 0.04 ku 0.08 1.37 1.68 Ov erall 92 M 3.20 3.75 3.57 SD 0.94 1.09 1.11 sk 0.00 0.59 0.24 ku 0.5 0 0.09 0.93 A visual examination of Table 4.7 reveals that mean interpretation scores for all five groups appear to be similar. On the P r etest, the structured input with visual input enhancement group had the highest mean interpretation test score (3.50), while the processing instruction without visual input enhancement had the lowest mean score (2.95). On P osttest 1, the processing instru ction with visual input enhancement group had the highest mean score (4.22), while the structured input with visual input enhancement group had the lowest mean score (3.33). On P osttest 2, the structured input without visual input enhancement group had th e highest mean score (3.79), while the processing instruction with visual input enhancement group had the lowest mean score (3.44). In order to determine if there were significant group differences over time, the

PAGE 252

238 data were subjected to a repeated measures ANOVA. However, before conducting the ANOVA, the assumptions underlying the statistical test were assessed. In order to assess the assumption of univariate normality, skewness and kurtosis values were examined for each dependent variable by group. For the distributions of scores on the Pretest, skewness values ranged from .50 to .11 and kurtosis values ranged from 1.62 to 1.03. For the distributions of scores on Posttest 1, skewness values ranged from 1.07 to .01 and kurtosis valu es ranged from 1. 37 to 1.91, and f or the distributions of scores on Posttest 2, skewness values ranged from .55 to .08 and kurtosis values ranged from 1.68 to .13. Shapiro Wilk tests were performed on each variable by group. The Shapiro Wilk tests revealed that t he dis tributions of scores on the P retest were not normal for the following groups: processing instruction without visual input enhancement, structured input without visual input enhancement, and processing instruction with visual input enhancement. Similarly, interpretation test scores at P osttest 1 were found to deviate from normality for the following groups: traditional instruction, structured input without visual input enhancement, and processing instruction with visual input enhancement. Finally, for the distribution s of scores at P osttest 2, the Shapiro Wilk tests revealed that the f ollowing distributions were non normal: traditional instruction and processing instruction without visual input enhancement. As several of the distributions were found to devi ate from normality box plots were examined for outliers and no significant outliers were identified in the data set As the ANOVA test is robust to violations of normality, the decision was made to proceed with the analysis.

PAGE 253

239 The assumption of sphericity was checked with the Greenhouse Geisser test, which estimated sphericity at = .97. As this is almost a perfect estimate the assum ption of sphericity was met for the present analysis. Data were subjected to a repeated measures ANOVA wi th one between subjects factor (type of instruction) and one within subjects factor (time of testing ). The results of the analysis are presented in Table 4.8. Table 4.8 Analysis of Variance of Interpretation Test Scores by Instructional Treatment Group and Time for Inte rpretation of the Indicative Source df SS MS F Between S ubjects (Ss) 91 128.33 Group (A) 4 2.34 0.58 0.40 S/A 87 125.99 1.45 Within S ubjects 184 186.38 Time (B) 2 14.37 7.18 7.76 A x B 8 10.88 1.36 1.47 SB/A 174 161.13 0.93 Total 275 314.71 Note N = 92. p < .001

PAGE 254

240 The results did not reveal a significant Group x T ime interaction effect F (8, 174) = 1.36, p > .05. The main effec t for time using the Greenhouse Geisser test was significant, F (2, 174) = 7.77, p < .001. The main effect for type of instruction was not significant, F (4, 87) = 0.40, p > .05. The effec t size for the significant main effect for time was computed, = .08, which was a small effect siz e. A graph of the significant m ain effect for time is presented in Figure 4.4 Figure 4.4 Graph of the Main Effect for Time for the Mean I nterp retation Test S core at Pretest, Posttest 1, and Posttest 2 for Interpretation of the I ndicative As a follow up to the significant main effect for time, post hoc contrast tests were performed to dete rmine if changes in the mean interpretation test score were statis tically significant at each point in time P ost hoc contrast tests revealed that the mean interpretation test score was significant ly higher at Posttest 1 compared to P retest, F (1, 87) = 14.77, p < .001. Similarly, the mean interpretation test score was also significant ly

PAGE 255

241 higher at Posttest 2 compared to Pretest, F (1, 87) = 7.87, p < .01. However, there was no statistically significant difference in the mean interpretation test score from Posttest 1 to Posttest 2 F (1, 87) = 1.50, p > .05 In order t o control for the Type I error rate, the B onferroni adjustment was applied, with alpha set a t .0167 for the set of contrast tests. Analysis o f Scores for Production of the Indicative Part and p osttest scores on the P roduction Subtest of the Subjunctive Knowledge Test were analyzed using two repeated measures ANOVAs with one between subjects factor (type of instruction) and one within subjects factor (time of testing) subju Production Subtest consisted of indi production of the indicative, which was included in the present study to examine the possibility of learner overextension of the targeted grammatical form. Participant s in the present study already knew how to form and use the indicative mood in Spanish; however, the instructional treatments focused on subjunctive versus indicative contrasts. Therefore, it was important to determine if the instructional treatments eith er positively Spanish. The production of the indicative component of the Production Subtest was not used as a screening device for exclusion from participation in th e study; rather, scores measuring the production of the indicative at P retest served as a baseline measure of

PAGE 256

242 highest score possible for production of the indi cativ e was 5. Table 4.9 presents the of the indicative component of the Production Subtest Table 4.9 Descriptive Statistics for Scores on the Production Subtest by Group at Pretest, Posttest 1, and Posttest 2 for Production of the Indicative Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + PI VIE 19 M 2.42 3.05 2.79 SD 1.44 1.12 1.31 sk 0.22 0.01 0.13 ku 0.86 1.06 0.02 + PI +VIE 18 M 2.8 0 3.25 3.14 SD 1.43 1.49 1.17 sk 0.63 1 .00 0.28 ku 0.1 0 0.22 0.08 + SI VIE 19 M 2.45 3.53 2.39 SD 1.52 1.36 1.7 0 sk 0.03 0.43 0.3 0 ku 0.86 1.06 0.83

PAGE 257

243 Instructional G roup Time of T esting n Pretest Posttest 1 Posttest 2 + SI +VIE 18 M 2.83 2.89 2.86 SD 1.36 1.54 1.49 sk 0.22 0.43 0.56 ku 1.16 0.36 0.2 0 + TI 18 M 2.33 2.89 2.61 SD 1.47 1.84 1.37 sk 0.08 0.28 0.03 ku 0.96 1.54 1.27 Overall 92 M 2.57 3.13 2.76 SD 1.43 1.47 1.41 sk 0.07 0.49 0.18 ku 0.89 0.71 0.68 A visual examination of Table 4.9 reveals that mean scor es for production of the indicative appear to be similar for all five groups at Pretest, Posttest 1, and P osttest 2. The structured input with visual input enhancement group had the highest mean score on the P retest (2.83), while the traditional instructi on grou p had the lowest mean score at Pretest (2.33). On P osttest 1, the mean scores for all of the groups appear to be somewhat higher than the mean scores for production of the indicative at Pretest The processing instruction with visual input enhance ment group had the highest mean score on P osttest 1 (3.25), while the traditional instruction group and the structured input with visual input

PAGE 258

244 enhancement groups had the lowest mean scores on Posttest 1 (2.89). On P osttest 2, the processing instruction wi th visual input enhancement group had the highest mean score (3.14), while the structured input without visual input enhancement group had the lowest mean score for production of the indicative (2.39). In order to determine if mean scores differed signifi indicative component of the Production Subtest were submitted to a r epeated measures ANOVA with one between subjects factor (type of instruction) and one within subjects factor (time of testing ). Before conducting the ANOVA test, univariate normality and sphericity assumptions were assessed An examinatio n of the skewness and kur tosis values fr om Table 4.9 reveal that the distributions of scores on the P retest were approximately norm ally distributed, with skewness values ranging from .63 to .22 and kurtosis values ranging from 1.16 to .10 The distributions of scores on P osttest 1 had skewness values that ranged from 1.0 to .01 and kurtosis values that ranged from 1.54 to .22. The distributions of scores on Posttest 2 had skewness values that ranged from .56 to .30 and kurtosis values that ranged from 1.27 to .02 As a further assessment of normality, the Shapiro Wilk test was performed on each dependent variable by group. T hese tests revealed that there were no violations of normality for the distribution of mean production test scores on the Pretest and on P osttest 1 for all of the groups ( p > .05 ) However, the distribution s of mean production test scores were not norma ll y distributed on P osttest 2 for the traditional instruction group, the structured input without visual input enhancement group, and the processing

PAGE 259

245 instruction with visual input enhancement group Although there was a violation of the assumption of normali ty, the decision was made to proceed with the analysis since the ANOVA test is fairly robust to violations of this assumption. The final assumption that was checked was sphericity. The Greenhouse Geisser test was used to assess this assumption, and the estimate for sphericity was = 97. This was nearly a perfect estimate of sphericity. After assessing the assumption s the data were submitted to a r epeated measures ANOVA with one between subjects factor (type of instruction) and one within subjects fa ctor (time of testing ). The results of the analysis are presented in Table 4.10. The ANOVA did not reveal a significant Group x Time interaction effect, F (8, 174) = 0.77, p > .05. There was a significant main effect for time using the Greenhouse Ge isser test, F (2, 174) = 04.92, p < .01; however, there was not a significant main effect for type of instruction, F (4, 87) = 0.45, p > .05. The effect size for the main effect for time was computed, = .05, which was a small effect size. A graph of the significant main effect for time is presented in Figure 4.5. As a follow up to the significant main effect for time, post hoc contrast tests were performed to determine if changes in the mean product ion test score at each point in time were statistically significant. In order to control the Type I error rate, the Bonferroni adjustment was applied, with alpha set at .0167 for the set of post hoc contrast tests. The contrast tests revealed that the me an production test score was significantly higher at Posttest 1 compared to Pretest, F (1, 87) = 8.79, p < .01; however, there was not a

PAGE 260

246 Table 4.10 Analysis of Variance of Production Test Scores by Instructional Treatment Group and Time for Production of the Indicative Source df SS MS F Between S ubjects (Ss) 91 297.58 Group ( A) 4 5.98 1.49 0.45 S/A 87 291.60 3.35 Within S ubjects 184 280.61 Time (B) 2 14.52 7.26 4.92* A x B 8 9.15 1.14 0.77 SB/A 174 256.94 1.48 Total 275 578.19 Note N = 92. p < .05 significant difference in the me an p roduction test score at P osttest 2 compared to Posttest 1 F (1, 87) = 4.99, p = 03. Nor was there a significant difference in the mean production test score at P osttest 2 compared to Pretest F (1, 87) = 1.03, p > .05.

PAGE 261

247 Figure 4.5 Graph of the Main Effect for Time for the Mean Production Test Score at Pretest, Posttest 1, and Posttest 2 for Production of the I ndicative Analysis of Note and Awareness Scores The results of the data obtained from the notes that participants took while they read an Au thentic Input Passage online and the results of the data obtained from a of the targeted forms that were embedded in the Authentic Input Passage were submitted to a M ultivariate Analysis of Variance (MANOVA) with type of ins truction as the independent variable and note scores and awareness scores as the dependent variables. There were a total of 15 instances of the targeted verb forms embedded in the Authentic Input t ext; thus, the maximum note score possible was 15, which would indicate that a participant noticed all of the subjunctive forms in the passage. The Posttreatment

PAGE 262

248 form as it appeared in authentic input, and it required participants to provide metalinguistic information about the use of the subjunctive in adjectival clauses. The highest awareness score possible was 3, with a score of 2 3 indicating a high level of awareness A score of 1 indicated that a participant had a medium level of awareness, and a score of 0 .5 indicated a low level of awareness. As noticing and awareness are separate but related constructs, note scores were used to assess the amount of participan and awareness scores by group are provided in Table 4.11. Table 4.11 Descriptive Statistics on Note and A wareness Scores by Group Instructional G roup T ype of Measure n Note Awareness + PI VIE 19 M 9.42 1.63 SD 2.81 1.08 sk 0 .00 0.21 ku 1.13 1.71 + PI +VIE 18 M 9.61 2.14 SD 3.82 1.05 sk 0.47 1.09 ku 0.08 0.17

PAGE 263

249 Instructional G roup T ype of Measure n Note Awareness 19 +SI VIE M 8.42 1.03 Sk 3.76 1.03 sk 0.75 1.22 ku 0.12 0.11 + SI +VIE 18 M 7.72 1.06 SD 3.88 0.95 sk 0.19 1.08 ku 0.17 0.24 + TI 18 M 7.83 1.17 SD 3.11 1.04 sk 0.24 0.77 ku 0.73 0.88 Overall 92 M 8.61 1.4 0 SD 3.51 1.1 0 sk 0.31 0.38 ku 0.2 0 1.45 An examination of Table 4.11 reveals that mean note scores were the highest for the processing instruction with visu a l input enhancement group (9.61) and they were the lowest for the structured input with visual input enhancemen t group (7.72). However, there does not appear to be a large difference between the mean note scores of any of the

PAGE 264

250 groups. For awareness scores, the processing instruction with visual input enhancement group had the highest mean awareness score (2.14), w hile the structured input without visual input enhancement had the lowest mean awareness score (1.03). In order to determine if the differences were statistically significant, the data were subjected to a MANOVA with one independent variable (type of inst ruction) and two dependent variables (note scores and awareness scores). Before submitting the d ata to the statistical test, multivariate normality and homogeneity of covariance assumptions were asses sed The research design ensured that the assumption o f independence was met, as the Authentic Input text and Posttreatment Questionnaire were individually administered to participants online. In addition, there was random assignment of participants to groups In order to evaluate normality, both univariate and mult ivariate normality were examined. Univariate normality was assessed by checking the skewness and kurtosis values of the distributions of note and awareness scores by group. The distributions of note scores for all of the groups appear to be fa irly normally distributed, with skewness values ranging from .75 to .24 and kurtosis values ranging from 1.13 to .73. The distributions of awareness scores had skewness values that ranged from 1.09 to 1.22 and kurtosis values that ranged from 1.71 to .24. As some of the skewness and kurtosis values were higher than 1, the Shapiro Wilk test for normality was performed on each dependent variable by group. The p values for all of the tests were higher than .05, which indicates that the assumption of uni variate normality was not violated.

PAGE 265

251 In order to assess multivariate normality, the data were first checked for multivariate outliers by calculating the effect size for note and awareness scores using D = 7.95. Two multi variate outliers were identified; however, the outliers were not outside the range of possible scores as indicated by the F test to check for multivariate outliers, F (2, 89) = 4.31, p < .05. Since the scores for the multivariate outliers were possible, t he analysis was run with the scores included. Multivariate skewness was checked and found to be in the range expected for samples from a multivariate normal distribution, (4) = 4.44, p > .05. The chi square value was not significant, which indicates that multivariate skewness was not violated. Similarly, multivariate kurtosis was checked and converted to a z score, which fell within the normal distribution, indicating that multivariate kurtosis was not violated. Thus, the examination of multivariate skewness and kurtosis values revealed that the assumption of multivariate normality was met. The correlation between the two dependent variables was checked to examine the strength of the relationship. The relationship between note scores and awareness scores was linear and positive ( r = .27). In addition, the standard deviations for both dependent variables for each group were examined and found to be similar. Finally, in order to verify that the assumption of homogeneity of covariance matrices was met, the data were square value was not statistically significant, (12) = 6.59, p > .05 Thus, it is reasonable to conclude that the assumption of homogeneity of covariance matrices was not violated.

PAGE 266

252 In addition, the degree of association was quantified by calculating . The obtained value was 0.17, which indicates that appro ximately 17% of the generalized variance in note and awareness scores was accounted for by type of instructional method. The proportion accounted for in the population was estimated to be somewhat less, c = .07. Once it was verified that the assumpti ons were met, the data were submitted to a MANOVA test, with note and awareness scores as the dependent variables and type of instruction as the independent variable. The test yielded a statistically significant difference in group centroids, = 0.83, p < .05. The effect size for the MANOVA was calculated, = 17, which was a medium effect size. Since the MANOVA was significant, follow up ANOVA tests were performed on each of the dependent variables to determine on which of the variables th e groups differed. The follow up ANOVA with note scores as the dependent variable revealed that there was no statistically significant difference for type of instruction, F (4, 87) = 1.15, p > .05. However, the follow up ANOVA with awareness scores as th e dependent variable revealed that there was a statistically significant difference for type of instruction F (4, 87) = 3.98, p < .01. The magnitude of the treatment effect was computed, R = 0.15, which w a s a small treatment effect Table 4. 12 presents the results of the significant ANOVA.

PAGE 267

253 Table 4.12 ANOVA Summary Table for Awareness Scores by Instructional Treatment Group Source df SS MS F Type of i nstruction (A) 4 16.61 4.15 3.98* (S/A) 87 93.01 1.07 ______ __________________________________________________________________ Total 91 109.62 Note N = 92. p < .01 Because significant differences were found on the one way ANOVA with awareness scores as the dependent variable, a post hoc Tukey test was performed controlling alpha at the .05 level for the set of tests to determine which groups had the mean awareness score for the processing instruction with vis ual input enhancement group was sign ificantly higher than mean awareness score s for the traditional instruction group, the structured input without visual input enhancement group, and the structured input with visual input enhancement group ( p < .05). However, the Tukey test did not reveal any significant differences in mean awareness scores between the two processing instruction groups (processing instruction with visual input enhancement and processing instruction without visual input enhancement).

PAGE 268

254 Analysis of Text and Grammar Comprehensi on Scores Two scores were obtained from a comprehension test that participants completed after reading an authentic input passage in Spanish that contained 15 instances of the ty to comprehend the message of the passage and also their ability to comprehend the referential meaning of the targeted grammatical form, also known as input processing. Thus, the comprehension test yielded two scores, a text comprehension score and a gr ammar comprehension score. The maximum score for text comprehension was 5, and the maximum score for grammar comprehension was 6. The descriptive statistics for text and grammar comprehension scores by group are presented in Table 4.13. Table 4.13 Descri ptive Statistics on Text and Grammar Comprehension Scores by Group Instructional G roup T ype of Measure n Text Grammar + PI VIE 19 M 4.05 4.21 SD 0.85 2.09 sk 0.72 0.51 ku 0.37 1.59 + PI +VIE 18 M 4.11 4.5 0 SD 0.76 1.92 sk 0.19 0.7 0 ku 1.12 1.37

PAGE 269

255 Instructional G roup T ype of M easure n Text Grammar + SI VIE 19 M 4.32 2.47 SD 0.82 1.54 sk 0.68 1.33 ku 1.13 1.17 + SI +VIE 18 M 4.44 2.94 SD 0.86 1.76 sk 1.07 0.82 ku 0.7 0 0.75 + TI 18 M 4.17 3.22 SD 1.2 0 2.18 sk 1.05 0.37 ku 0.58 1.73 Overall 92 M 4.22 3.46 SD 0.9 0 2.02 sk 0.82 0.21 ku 0.4 0 1.65 A visual examination of Table 4.13 reveals that the group means for text comprehension appear to be similar for all five groups. The structured input with visual input enhancement group had the highest mean score for text comprehension (4.44) and the p rocessing instruction without visual input enhancement had the lowest mean score (4.05). For grammar comprehension, the processing instruction with visual input

PAGE 270

256 enhancement group had the highest mean score (4.50), while the structured input without visual input enhancement group had the lowest mean score (2.47). In order to determine if the group differences were significant, the data were subjected to a MANOVA with type of instruction as the independent variable and text and grammar comprehension scores as the dependent variables. However, before submitting the data to the MANOVA, multivariate normality, and homogeneity of covariance matrices were assessed, which are assumptions of the statistical test. In order to assess the assumption of normality, bo th univariate and mul tivariate normality were examined Univariate normality was assessed by examining skewness and kurtosis values for each dependent variable by group. For the distributions of text comprehension scores, skewness values ranged from 1.0 7 to .19 and kurtosis values ranged from 1.13 to .37. For the distributions of the grammar comprehension scores, skewness values ranged from .70 to 1.33 and kurtosis val ues ranged from 1.73 to 1.17. Thus, the distributions for these dependent measure s were not considered to be markedly skewed or kurtotic. Shapiro Wilk tests were also performed on each dependent variable by group. The Shapiro Wilk tests revealed that the scores for both text and grammar comprehension were not normally distributed. T he data set was checked for outliers by examining box plots ; however, there were no significant outliers in the data set. Data were also examined to determine if the assumptions of multivariate skewness and kurtosis were met. The results suggested depart ures from normality for both multivariate skewness and kurtosis. The data were screened for multivariate outliers

PAGE 271

257 distance. The maximum D value was 7.63, and two multivariate outliers were identified. The data were run without the mul tivariate outliers, and there was not a significant impact on multivariate skewness or kurtosis values. Therefore the multivariate outliers were retained in the data set. Consequently, the assumption of multivariate normality was not met. However, ther e is evidence to suggest that MANOVA is robust against lack of multivariate normality (Stevens, 2002). T o determine if the departure from normality power analysis was performed using statistica l analysis software (SAS). The analysis revealed that the power of the MANOVA te st was estimated to be .85. According to Stevens ( 2002), power of .80 is sufficient to detect gro up differences if they exist. Thus, it appears that the departure from normality did not adversely aff ect power. In addition, the MANOVA test is robust against violations of normality Thus, the decision was made to proceed with the analysis The final assumption that was checked was homogeneity of covariance matrices. highly sensitive to violations of normality. Examination of the chi square value from p value was not statistically significant, X (12 ) = 7.63, p > .05. Thus, it is r easonable to conclude that the assumption of homogeneity of covariance matrices was not violated.

PAGE 272

258 In addition, the degree of association was quantified by calculating . The obtained value was 0.17, which indicates that approximately 17% of the generalized variance in text and grammar comprehension scores was accounted for by type of instructional method. The proportion accounted for in the population was estimated to be somewhat less, c = .07. After assessing the assumptions of the test, t he data were subjected to a MANOVA with one independent between subjects variable (type of instruction) and two dependent variables (text and grammar comprehension scores). Th e results revealed a significant difference in group centroids, = 0.83 p < .05. The effect size for the MANOVA was calculated, = .17, which was a medium effect size. As the MANOVA was significant, follow up ANOVA tests were performed on each of t he dependent variables to determine on which of the variables the groups differed. The ANOVA with text comprehension scores as the dependent variable did not reveal a significant effect for type of instruction, F (4, 87) = 0.57, p > .05. However, the ANO VA with grammar comprehension scores as the dependent variable revealed a significant effect for type of instruction, F (4, 87) = 3.72, p < .01. The magnitude of the treatment effect was computed, R = 0.15, which w a s a small treatment effect The result s of the significant ANOVA are presented in Table 4.14.

PAGE 273

259 Table 4.14 ANOVA Summary Table for Grammar Comprehension Scores by Instructional Treatment Group Source df SS MS F Type of instruction (A) 4 54.45 13.61 3.72* (S/A) 87 318.45 ________________________________________________________________________ Total 91 372.90 Note N = 92. p < .01 Because the one way ANOVA test revealed significant group differences in mean grammar compr ehension scores, a po st hoc Tukey test was performed controlling alpha at the .05 level for the set of tests to determine which groups had statistically significant enh ancement and the processing instruction without visu al input enhancement groups had significantly higher mean grammar comprehension scores than the structured input without visual input enhancement group ( p < .05). However, the Tukey test did not reveal a ny significant differences between the two processing instruction groups and the structured input with visual input enhancement group or between the two processing instruction groups and the traditional instruction group.

PAGE 274

260 Correlational Analyse s of Comprehe nsion Test Scores In order to determine if there was a relationship between text comprehension and input processing (grammar comprehension), a Pearson product moment correlation coefficient was computed between the variables text comprehension and grammar comprehension for each of the four experimental groups (processing instruction without visual input enhancement, processing instruction with visual input enhancement, structured input without visual input enhancement, and structured input with visual inpu t enhancement) and for the comparison group (traditional instruction). For the processing instruction without visual input enhancement group, r = .19, which indicates a weak negative relationship between text comprehension and input processing. For the processing instruction with visual input enhancement group, r = .20, which indicates a weak positive relationship between text comprehension and input processing. Similarly, for the structured input without visual input enhancement group, r = .27, which i ndicates a weak positive correlation between text comprehension and input processing. However, the structured input with visual input enhancement group and the traditional instruction group demonstrated no correlation between text comprehension and input processing, r = .06 and r = .03 respectively. Summary of the Posttreatment Questionnaire At the end of the s tudy, participants completed a Posttreatment Q uestionnaire. The purpose of the questionnaire was t wo fold: (a) it was a retrospective measure of pa authentic input text, and (b) study related

PAGE 275

261 materials. The Posttreatment Q uestionnaire asked participants whether they believed t hat the study materials were more helpful and/or enjoyable than their regular course materials. Participants were also asked to rank the difficulty level of the study materials compared to their regular course materials, and they were asked if they felt t hat the study grammar activities were new and insightful. In addition, participants were asked to select the elements of the instructional treatments that they found to be the most and least helpful for learning the targeted grammatical form. Finally, pa rticipants were asked if they sought outside assistance while completing any of the study activities; and if so, they were asked to list which resource(s) they consulted. The results of the Posttreatment Q uestionnaire revealed that 82 students (90.22%) believed that the grammar activities presented in the study were new and insightful. Only 9 students (9.78%) stated that the study grammar activities were not new and insightful. Interestingly, 17 of the 18 students in the traditional instruction group s tated that they believed that the study grammar activities were new and insightful, even though the activities that they completed were almost identical to their regular course materials. It is important to note that all participants were told that they w ere receiving a novel instructional technique for teaching complex Spanish grammar online. By and large, participants in the traditional instruction group did not recognize that the method of instruction that they received was not new or unique. When a sked to rate the level of difficulty of the study materials compared to materials were easier than their regular course materials, 21 participants (22.83%) felt

PAGE 276

262 that the study materials were harder than their regular course materials, and 21 participants (22.83%) stated that the study materials were about the same as their regular course materials with respect to level of difficulty. When asked if the study materials wer e more helpful than their regular course materials, 72 participants (78.26%) expressed that the study materials were more helpful than their regular course materials, and 20 participants (21.74%) stated the study materials were not more helpful than their regular course materials. Participants were also asked if the study materials were more enjoyable than their regular course materials. A total of 52 students (56.52%) felt that the study materials were more enjoyable than their regular course materials, while 40 students (43.48%) felt that the study materials were not more enjoyable than their regular course materials. In an effort to uncover which elements of the instructional treatments the participants in the various groups believed were the most and least helpful for learning Spanish grammar, they were given two lists of the major components in the study materials, and they were asked to check the ones that they felt were the most and least helpful to them. Participants were able to check one, severa l, all, or none of th e components that were listed. The participants in all of the groups were exposed to the following three components: listening activities, written activities, and graphics. However, not all of the study participants selected these ac tivities as being the most or least helpful for learning. A total of 40 participants (43.48%) selected the listening activities as being the most helpful component of the study materials; while 41 participants (44.56%) stated that the listening activities were the least helpful component

PAGE 277

263 of the study materials. A total of 11 participants (11.96%) did not select the listening activities as being helpful or unhelpful for learning. For the written activities, 65 participants (70.65%) felt that they were hel pful, 16 (17.39%) felt that they were not helpful, and 11 (11.96%) did not rate the written activities as either helpful or unhelpful. Regarding the graphics, 16 participants (17.39%) found the graphics to be helpful and 30 participants (32.61%) found the graphics to be unhelpful. However, 46 participants (50%) did not rate the graphics as being particularly helpful or unhelpful for learning. The following components were specific to certain groups: speaking activities, word animations, grammar explanat ions, and information on processing strategies. Only the traditional instruction group had speaking activities. There were 18 participants in the traditional instruction group and 3 participants (16.67%) in this group stated that the speaking activities were helpful, while 3 participants (16.67%) stated that the speaking activities were not helpful. However, 12 participants (66.66%) in the traditional instruction group did not rate the speaking activities as being particularly helpful or unhelpful for le arning. Regarding the word animations, the participants in two groups received instructional treatments that included word animations: the processing instruction with visual input enhancement group and structured input with visual input enhancement group. There were a total of 37 participants in both of these groups, and 7 participants (18.92%) in these two groups stated that the word animations were helpful, while 8 participants stated that they were not helpful (21.62%). However, 22 participants

PAGE 278

264 (59.46 %) in the two groups that received visual input enhancement did not rate the word animations as being helpful or unhelpful for learning. Three groups received an explicit explanation of grammar: the processing instruction without visual input enhancemen t group, the processing instruction with visual input enhancement group, and the traditional instruction group. There were a total of 55 participants in these three groups, and 26 students (47.27%) in these three groups expressed that the grammar explanat ions were helpful, while only 2 students (3.64%) felt that the grammar explanations were not helpful. A total of 27 students (49.09%) did not rate the grammar explanations as being particularly helpful or unhelpful for learning. Finally, two groups recei ved information on processing strategies: the processing instruction without visual input enhancement group and the processing instruction with visual input enhancement group. There were a total of 37 participants in these two groups, and 15 of the 37 par ticipants (40.54%) found the information on processing strategies to be helpful, while 3 participants (8.12%) did not find the information on processing strategies to be helpful. Over half of the participants in these two groups (19 participants or 51.35% ) did not rate the information on processing strategies as being particularly helpful or unhelpful for learning. Finally, only 2 of the 92 participants (2.17%) in the study stated that they consulted outside resources when completing their web based ins tructional activities. A closer exa mination of their questionnaires revealed that one participant in the traditional instruction group consulted a dictionary for unknown vocabulary words and one participant in the processing instruction with visual input enhancement group consulted

PAGE 279

265 the Internet for unknown vocabulary words. None of the participants indicated that they consulted their textbooks, the Internet, their teacher, their peers, or any other resource for an explanation of the targeted grammatical f on the Posttreatment Questionnaire by group is provided in Table 4.15. Ended Responses from the Posttreatment Questionnaire Participants were asked to express their opinions on the Posttreatm ent expressed a clear preference for the PI materials over their regular course materials (traditional instruction), as 86.49% of participants in these two groups stated that the PI materials were more helpful than their regular course materials, 70.27% of participants expressed that the PI activities were easier than their regular course mater ials, and 91.89% of participants stated that the PI materials were a new and insightful way to learn Spanish grammar. Students had the opportunity to express their opinions on the Posttreatment Questionnaire, and the majority of students in the two PI gro ups expressed a clear preference for the study materials over their present course materials, mainly due to the explicit information that they received. Participant #178 from Class 3 stated, I think the grammar package presented the information in a way t hat was easy to understand. I picked up on it a lot easier than some other concepts activities broke things down into very understandable bits of information. A lot of times in the Spani sh course, things are just thrown at us without much of an explanation as to why things are the way they are. I actually understood what was going on in the grammar activities for the study.

PAGE 280

266 Table 4.15 Participant Responses f rom the Posttreatment Questio nnaire +PI VIE +PI+VIE +SI VIE +SI+VIE +TI ( n = 19) ( n = 18) ( n = 19) ( n = 18) ( n = 18) Study grammar activities new and insightful Agree (19) Disagree (0) Agree (15) Disagree (3) Agree (16) Disagree (3) Agree (16) Disagree (2) Agree (17) Disagree ( 1) Difficulty of study materials compared to regular course materials Easier (14) Harder (1) Same (4) Easier (12) Harder (2) Same (4) Easier (9) Harder (4) Same (6) Easier (7) Harder (8) Same (3) Easier (8) Harder (6) Same (4) Study materia ls more helpful than regular course materials Agree (18) Disagree (1) Agree (14) Disagree (4) Agree (14) Disagree (5) Agree (13) Disagree (5) Agree (13) Disagree (5) Study materials more enjoyable than regular course materials Agree (13) Disagree ( 6) Agree (10) Disagree (8) Agree (9) Disagree (10) Agree (7) Disagree (11) Agree (13) Disagree (5) Most/Least helpful aspects listening activities Helpful (10) Not helpful (7) Helpful (10) Not helpful (4) Helpful (4) Not helpful (15) Helpful (11) Not helpful (6) Helpful (5) Not helpful (9) written activities Helpful (13) Not helpful (2) Helpful (10) Not helpful (2) Helpful (15) Not helpful (4) Helpful (10) Not helpful (6) Helpful (17) Not helpful (2) gra phics Helpful (2) Not helpful (8) Helpful (2) Not helpful (6) Helpful (5) Not helpful (3) Helpful (2) Not helpful (9) Helpful (5) Not helpful (4) speaking activities N/A N/A N/A N/A Helpful (3) Not helpful (3) word an imations N/A Helpful (1) Not helpful (0) N/A Helpful (5) Not helpful (4) N/A grammar explanations Helpful (11) Not helpful (0) Helpful (8) Not helpful (2) N/A N/A Helpful (7) Not helpful (0) processing strategies Helpfu l (8) Not helpful (1) Helpful (7) Not helpful (2) N/A N/A N/A Note N = 92

PAGE 281

267 Another aspect of PI that participants found helpful was the information on processing strategies. Participant #122 from Class 2 stated the following: I think that the grammar p ackage expressed the subjunctive forms in a document that explained the common problems that Spanish language learners encounter when trying to understand the subjunctive form. This was very helpful for me and explained a different way of thinking when approaching a Spanish sentence in the subjunctive form. The materials in the activity package were significantly easier for me to understand, and in my opinion were much more informative than the book assigned to this course. Interestingly, several participants in the two PI groups stated that they felt the study them to teach the information to themselves. P articipant #165 from Class 2 stated the following: I feel like I have learned more from using this method where it is explained instead of the trial and error method of the current system There came a point in the activity package where all of a sudden I felt like I understood the concept of subjunctive vs. indicative. The best way I can phrase the difference is that with the explanations in the word documents I felt like I was being taught something whereas with the normal method, it seems like I am teac hing myself. As all three classes that comprised the sample in the presents study operated under the traditional instruction paradigm, students were expected to work independently and learn the course content through interaction with the materials. Inter estingly, many of the participants in the two PI groups expressed the belief that the study materials provided them with instruction, while their current course materials required them to teach the course content to themselves. Both the PI materials and p materials provided an explicit explanation of grammar; however, the PI materials were

PAGE 282

268 not paradigmatic as only one subjunctive form was presented (the third person) in order Further, the PI materials provided regular course materials. For the two structured input groups, the participants expressed more mixed feelings about the study ma terials. While 86.79 % of the participants in these two groups felt that the study grammar activities were a new and insightful way to learn Spanish grammar, only 43.24% of participants in the structured input groups expressed that the study materials wer e easier than their regular course materials, and 56.76% of participants expressed that the study activities were less enjoyable than their current revealed that the y sometimes felt lost and confused because they did not receive an explanation of the grammar rules for using the subjunctive. Participant #151 from Class 2 stated, understand how to do t enough information and the grammar pack that I worked with had even less information. Another participant, #164 from Class 2 stated the following: The grammar package presented the verbs in a way that was easy to me to conjugate them in sentences. It appears that the participants in the structured input groups felt that they were lacking key information that they needed to complete the activities correctly. Although these

PAGE 283

269 participants received structured input activities with an example of how to answer each activity type correctly, they did not receive an expl icit explanation of grammar or information on processing strategies. The responses from the participants in the traditional instruction group were very mixed as well. Even though these participants received instructional activities and grammar explanatio ns that were identical to their regular course materials, 91.67% of participants in the traditional instruction group stated that the study activities presented it appe ars that many of them were referring to the way that the information was displayed rather than the content of the instruction. For example, participant #54 from Class 1 stated, I found it to have some similar teaching styles that are found in the onli ne class except I like the style of this study more than that of the class. I found it to be more helpful and a little more informative, displaying the information in different ways. The traditional instruction group had the same format as the experiment al groups with respect to screen size navigation, tool bars, etc. However, it is possible that the learners found the interface and the way the instructional content was visually displayed to be more helpful than their regular course materials. In add ition, 41.67% of the participants in the traditional instruction group felt that the study materials were easier than their regular course materials, while 38.89% of participants felt that the study materials were harder than their regular course materials

PAGE 284

270 #7 from Class 1 stated, I found that it was a lot more helpful in helping me understand the subjunctive tense and be able to use it and write it in sentences. I think playing individual sentences and having to pick out verbs or determining if the sentence was in the subjunctive or indicative tense was new a proved to be a lot easier than the activities I normally do in my online Spanish class. Perhaps it was bec ause there were many little activities and individual questions and recordings instead of paragraphs to be read and lots of long activities. In the previous example, the participant expressed that the study materials helped her learn subjunctive verses in dicative contrasts. She also mentions that the format of the study grammar activities was helpful because there were many short activities rather than a few longer ones. This response and others like it reveal that the traditional instruction materials i n the study may have been slightly easier for participants than their regular course materials because they were only required to produce sentential level output, either orally or in writing. In contrast, their regular course materials required them to pr oduce discourse level output, or written and spoken output that spanned one or two paragraphs in length. Participant #142 from Class 2 felt that the study materials were not particularly helpful or unhelpful. She stated, understand how to change the verbs in each sentence. These activities seemed harder but I think that is only because I verbs at all. The way things were explained was t really understand most of the information in the book because it is in Spanish

PAGE 285

271 This participants points out another key difference between the study materials and re given in English for all four experimental groups and the comparison group in order to avoid confusion. With Spanish. Finally, 72.22% of the participants in t he traditional instruction group felt that the study materials were more helpful than their regular course materials. While this figure regular course materials, an examination of their responses revealed that there were three regular course activities: (a) the way the content was displayed visually, (b) the study activities only required participants to produce sentential level output, and (c) the instructions were given in English. It is important to note that these differences were equivalent across the treatment groups in the present study. The participants in all five g roups received the instructional activities in the same format (screen design, navigation, etc.), all groups received instructions in English, and while participants in the traditional instruction group were only required to produce sentential level output participants in the four experimental groups were only required to interpret sentential level input. Out of the 92 participants in the study, 72 (78.26%) expressed that the study activities were more helpful than their regular course materials. In addi tion, many participants in all five groups expressed a clear dislike for their regular course materials on the Posttreatment Questionnaire. The participants were very familiar with their

PAGE 286

272 regular course materials and had been working with them for almost a full academic year by the time that the study took place. Their regular course materials fit into what Paultson (1972) describes as traditional instruction, and they were delivered within the traditional instruction distance learning paradigm, which reli es on the materials rather than on the teacher to provide instruction. Overall Summary of the Results of the Major Statistical Analyses Repeated Measures Analyses interpretation of the s ubjunctive at three poin ts in time (Pretest, Posttest 1, and P osttest 2), the results indicate d that there was a significant Group x Time interaction effect F (8, 174) = 2.10, p < .05. A Post hoc Tukey test revealed that the processing instruction with v isual input enhancement group had a significantly higher mean interpretation test score than the structured input w ithout visual input enhancement group at Posttest 1 compared to Pretest ( p < .05). There was also a significant main effect for time, F (2, 174) = 100.26, p < .0001 ; however, there was not a significant between subjects effect for type of instruction. As a follow up to the significant main effect for time, post hoc contrast tests were performed. The contrast tests revealed that the mean inte rpretation test score was significantly higher at Posttest 1 compared to Pretest, F (1, 87) = 182.86, p < .0001. The mean interpretation test score was also significantly lower at Posttest 2 compared to Posttest 1, F (1, 87) = 17.25, p < .0001. However, the mean interpretation test score was still significantly higher at Posttest 2 compared to Pretest, F (1, 87) = 87.30, p < .0001.

PAGE 287

273 the sub junctive over time the res ults indic ate d that there was no t a significant Group x Time interaction effect ; however, there was a highly significant main effect for time, F (2, 174) = 137.31, p < .0001. There was not a significant between subjects effect for type of instruction Post hoc con trast tests were performed as a follow up to the significant main effect for time. The contrast tests revealed that the mean production test score was significan tly higher at Posttest 1 compared to Pretest F (1, 87) = 212.11, p < .0001. In addition, the mean production test score was significantly lower at Posttest 2 compared to Posttest 1, F (1, 87) = 10.63, p < .01 However the mean production test score was still significan tly higher at Posttest 2 compared to Pretest, F (1, 87) = 136.05, p < .0001. The repeated measures analysis tion of the indicative did not reveal a significant G roup x Time interaction effect. The test did reveal a significant main effect for time, F (2, 174) = 7.77, p < .001 ; however, the re was not a significant between subjects effect for type of instruction As a follow up to the main effect for time, post hoc contrast tests were performed, which revealed that the mean interpretation test score was significan tly higher at Posttest 1 com pared to Pretest F (1, 87) = 14.77, p < .001. Similarly, the mean interpretation test score was significant ly higher at Posttest 2 compared to Pretest F (1, 87) = 7.78, p < .001. Conversely, there was no t a significant difference in the mean interpretat ion test score from Posttest 1 to Posttest 2 F (1, 87) = 1.50, p > .05. dicative over time the repeated me asures analysis did not reveal a significant Group x Time interaction

PAGE 288

274 effect. Howe ver, there was a significant main effect for time, F (2, 174) = 04.92, p < .05, and there was no t a significant between subjects effect for type of instruction. Post hoc contrast tests were performed as a follow up to the significant main effect for time. The contrast tests revealed that the mean production test score was significan tly higher at Posttest 1 compared to Pretest F (1, 87) = 8.79, p < .01. H owever, there was no significant difference in the mean production test score from Posttest 1 to Post test 2 F (1 87) = 4.99, p = 03. Nor was there a significant difference in the mean production test score at Posttest 2 compared to Pretest F (1 87) = 1.03, p > .05. In order to control the Type I error rate, the B onferroni adjustment was applied, with alpha s et at .0167 for the set of post hoc contrast tests. Multivariate Analyses and awareness scores revealed a statistically significant difference in group centroids, = 0.83, p < .05. Follow up ANOVAs on both dependent measures revealed that there were significant differences in mean awareness scores by instructional treatment group, F (4, 87) = 3.98, p < .01, and a post hoc Tukey test revealed that the processing in struction with visual input enhancement group performed significantly better than three other groups: the traditional instruction group, the structured input without visual input enhancement group, and the structured input with visual input enhancement gro up ( p < .05). However, there were no significant differences in mean awareness scores between the processing instruction with visual input enhancement group and the processing instruction without visual input enhancement group.

PAGE 289

275 The MANOVA that examined p scores also revealed a significant difference in group centroids, = 0.83, p < .05. Follow up ANOVAs on both dependent measures revealed that there were significant differences in mean grammar comprehension tes t scores by instructional treatment group, F (4, 87) = 3.72, p < .01. A post hoc Tukey test revealed that the processing instruction with visual input enhancement group and the processing instruction without visual input enhancement performed significantl y better than the structured input without visual input enhancement group ( p < .05). However, the two processing instruction groups (with and without visual input enhancement) did not perform significantly better than the structured input with visual inpu t enhancement group and the traditional instruction group. Correlational Analyses There was no correlation found between text and grammar comprehension scores for the traditional instruction group ( r = .03) or the structured input with visual input enhanc ement group ( r = .06). There was a weak negative correlation between text and grammar comprehension scores for the processing instruction without visual input enhancement group ( r = .19). There was a weak positive correlation between text and grammar co mprehension scores for the structured input without visual input enhancement group ( r = .27) and the processing instruction with visual input enhancement group ( r = .20).

PAGE 290

276 Chapter 5 Discussion Introduction This chapter provides a discussion of the resu lts of the experiment that compared four novel instructional techniques (processing instruction without visual input enhancement, processing instruction with visual input enhancement, structured input without visual input enhancement, and structured input with visual input enhancement) with traditional instruction for the acquisition of the subjunctive in adjectival clauses by intermediate level distance learners of Spanish. The chapter begins with a discussion of the results of the experiment in terms of the research questions and hypotheses. After discussing the findings of the experiment, the chapter also presents a discussion of the theoretical and pedagogical implications of the research findings. Finally, the limitations of the study are presented a nd discussed, and some suggestions are made for future research. Discussion of Results Discussion of Findings in Relation to the Research Questions The present study investigated the components of processing instruction, a novel instructional technique th at is informed by second language acquisition (SLA) research. Processing instruction (PI) consists of an explicit explanation of grammar that is not

PAGE 291

277 paradigmatic, information on processing strategies, and structured input activities. VanPatten ( 1993, 199 6, 2002, 2004) claims that only structured input activities are needed to bring about improved performance on interpretation and production tasks. Interpretation tasks require learners to comprehend the referential meaning of targeted grammatical forms, a nd production tasks require learners to produce targeted forms either orally or in writing. However, Farley (2004) and Fernndez (2008) found that the explicit explanation component of PI is necessary when the targeted grammatical form is complex. The pr esent study combi ned PI and structured input activities with visual input enhancement (VIE) in an attempt to increase the salience of subjunctive verb forms for that are in the sentence medial position are processed last by second language (L2) learners. The targeted form of the present study was the subjunctive in adjectival clauses when the referent is uncertain, hypothetical, or unknown to the speaker. In Spanish, the subjunctive in adjectival clauses always occurs in the sentence medial position i n natural speech. Thus, VIE, which was operationalized as word animation in the present study, was utilize d to facilitate noticing of targeted verb forms as participants rea d input senten ces online. In addition, VIE was combined with structured input activities that were desig ned to help learners process targeted verb fo rms correctly. In order to determine if the explicit explanation component of PI is necessary, learners i n the experimental groups either received PI, which contained an explicit explanation of grammar, or structured input activities which did not The present study also investigated whether the addition of VIE to PI and structured input activities was able to

PAGE 292

278 increase the beneficial effect of these instructional techniques for learning complex Spanish grammar online. A meta analysis on VIE only revealed a slight positive effect for VIE ( S. Lee & Huang, 200 8 ) on grammar learning; however, in past language a cquisition studies VIE was only operationalized as simple typographical enhancements such as underling or bolding targeted forms. The present study updated VIE for web based delivery by using flash programming la nguage to animate subjunctive forms by maki ng the m grow larger and smaller for a period of seven seconds after participants ope ned the link for an input sentence that contained VIE. In addition, the word animations were delivered sequentially rather than simultaneously to avoid overloading learner items on the screen. The novel instructional techniques investigated in the present study were compared to traditional instruction, which is currently the dominant paradigm for foreign language instruction at the secondary and postsecondary levels in the United States (VanPatten, 2004). Traditional instruction requires learners to produce target language (TL) output immediately after they receive an explicit explanation of g rammar that is paradigmatic. Further, traditional instruction places a heavy emphasis on mechanical drill activities, which are vestiges from the audiolingual method, a foreign language teaching method that was founded on the principles of behaviorism. M echanical drill activities do not require learners to comprehend the stimulus in order to formulate a correct response in the TL. Thus, mechanical drills require learners to focus on form rather than on TL meaning. Conversely, structured input activities require a focus on

PAGE 293

279 both form and meaning in order for learners to an swer questions correctly. P ast research in the PI strand has found that PI is superior to traditional instruction for interpretation tasks and that PI is equivalent to traditional instru ction for production tasks (Benati, 2001, 2005; Cadierno, 1995; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004). Below are the first two research questions that were addressed within the context of the present study: 1. Is there a differential pe rformance between treatment groups for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks over time? 2. Is there a differential performance between treatment groups for the acquisition of the pres ent subjunctive in adjectival clauses in Spanish as measured by production tasks over time? Regarding the first research question, the results of the present study differ from the finding s of past studies in the PI strand that compared processing instruc ti on with traditional instruction as past studies found that processing instruction was superior to traditional instruction for interpretation tasks (Benati, 2001, 2005; Cadierno, 1995; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004) A repea ted measures ANOVA performed on the interpretation test scores found a significant Group x Time interaction effect, and a post hoc Tukey test revealed that participants who were exposed to processing instruction with visual input enhancement (+PI +VIE) per formed significantly better than participants who were exposed to structured input without visual

PAGE 294

280 input enhancement (+SI VIE) across time (from Pretest to Posttest 1) as measured by interpretation tasks The findings of the present study did not indicate that exposure to processing instruction with or without visual input enhancement wa s superior to traditional instruction for interpreting the subjunctive as there were no significant differences between the processing instruction groups and the traditi onal instruction group across time Although the processing instruction with visual input enhancement group (+PI +VIE) outperformed the struct ured input without visual input enhancement group (+SI VIE) for interpreting the subjunctive in the short term significant group diff erences were not retained from P osttest 1 to P osttest 2 However the repeated measures analysis also revealed a highly significant main effect for time. When the overall mean interpretation test score was examined over time, it app ears that exposure to terpretation of the subjunctive, as the mean in terpretation test score was significantly higher at both Posttest 1 and Posttest 2 compared to Pretest. For the second research question, the results of the pre sent study support the findings of past studies in the PI strand Past studies (Benati, 2001, 2005; Cadierno, 1995; VanPatten & Cadierno, 1993a, 1993b; VanPatten & Wong, 2004 ) found PI to be equivalent to traditio nal instruction for production tasks A repeated measures ANOVA performed on production test scores revealed that there was not a significant Group x Time interaction effect In addition, there was not a significant between subjects effect for type of in struction. However, there was a highly significant main effect for time. Thus, when the overall mean production test score was examined over time it appears

PAGE 295

281 that exposure to the instructional treatments had a beneficial effect on lity to pr oduce the subjunctive, as the mean production test score was significantly higher at both Posttest 1 and Posttest 2 compared to Pretest. A lthough participants in the four experimental gr oups never produced the subjunctive verb forms during their instructi onal treatments, they performed equally as well on production tasks as participants who did as there was no significant between subjects effect for type of instruction. Research Question 3 which is listed below, investigates whether the instructional tr eatments were able to help participants notice t argeted verb form s when they encountered them in authentic input following the instructional treatments. 3. Is there a ability to notice target e d forms in subsequent authentic input as measured by note scores and awareness scores? To answer this question, noticing was measured two ways: (a) the amount of noticing that examining their level of awareness. In order to assess the amount of noticing that took place, participants took notes while they read an authentic input text online one to three days after completing their instructional treatments. They were asked to no te all of the vocabulary words and verb forms that were necessary to comprehend the text. The number of targeted verb forms that participants noted was tallied, and as there were 15 subjunctive forms embedded in the text, the highest note score possible w as 15. of awareness was measured by a Posttreatment Q uestionnaire that they completed after reading an authentic input passage and tak ing a comprehension

PAGE 296

282 test. The Posttreatment Q uestionnaire asked participants if they could state t he targeted grammatical form that was present in the authentic input text and give an example of it, which constituted awareness at the level of noticing, or a medium level of awareness (Leow, 2000) Participants were also asked if they could state the mo rphological rule for using the grammatical form that they listed, and if they were able to do so, they demonstrated awareness at the level of understanding, or a high level of awareness (Leow, 2000) Participants demonstrated a low level of awareness if t hey were only able to name the grammatical form or provide an example of a subjunctive form from the authentic input text. By examining note scores and awareness scores, the present study found that there were no significant differences between the group s for the amount of noticing that took place. In other words, the mean number of targeted verb forms that were noted while participants read an authentic input passage online did not differ significantly by instructional treatment group. However, when pa measured by examining their level of awareness, the results revealed that the participants in the processing instruction with visual input enhancement group (+PI +VIE) outperformed the following groups: traditional instruc tion (+TI) structured input without visual input enhancement (+SI VIE) and structured input with visual input enhancement (+SI +VIE) Interestingly, the processing instruction with visual input enhancement group (+PI +VIE) did not outperform the process ing instruction without visual input enhancement group (+PI VIE)

PAGE 297

283 It appears that exposure to processing instruction with visual input enhancement helped participants notice the targeted forms in subsequent authentic input with a deeper level of awaren ess In other words, participants in the processing instruction with visual input enhancement group not only noticed the targeted verb forms in subsequent authentic input, they were also able to explicitly state the morphological rule for the use of the t argeted grammatical form as it appeared in authentic input, which Leow (1997, 2000) defines as noticing with metalinguistic awareness, or noticing at the level of understanding. Research Question 4 d grammar comprehension scores. The Comprehension Test measured two constructs: (a) comprehension of the propositional content of the input passage, and (b) comprehension of the referential meaning of the targeted verb forms. The second construct refers to grammar comprehension, which is also known as input processing. 4. Following the instructional treatments, is there a differential performance meaning of the targeted grammatic al form (input processing) and the message of the authentic input text in which it is embedded as measured by grammar comprehension and text comprehension scores? abilit y to comprehend the propositional content of the input text. However, there were meaning of the subjunctive forms that were embedded in the authentic input text that was

PAGE 298

284 received one to three days after completing the instructional treatments. The two processing instruction groups, processing instruction with visual input enhancement (+PI +VIE) and processing instruction w ithout visual input enhancement (+PI VIE), outper formed the structured input without visual input enhancement group (+SI VIE) as measured by grammar c omprehension uent processing of L2 input In addition to the beneficial effect on awareness, it appears that exposure to were embedded in an authentic input text that was received post experimental expo sure. Research Question 5 which is listed below, investigated the relationship between text comprehension and input processing, which are two related but separate constructs. 5. What is the relationship between text comprehension and input processing w hen learners encounter the targeted grammatical form in subsequent authentic input? Text comprehension and grammar comprehension scores were examined by group to determine if there was a relationship between the t wo constructs. The r esults of Research Que stion 5 do not indicate that there is a strong relationship between text comprehension and grammar comprehension for any of the groups that were investigated in the present study. There was no correlation found between text and grammar comprehension score s for the traditional instruction (+TI) group or the structured input with visual input enhancement group (+SI +VIE) There was a weak negative correlation between text and grammar comprehension scores for the processing instruction without visual input

PAGE 299

285 e nhancement group (+PI VIE) Several participants in this group appeared to have low text comprehension scores, but high grammar comprehension scores, which may account for the inverse relationship that was found. There was a weak positive relationship b etween text and grammar comprehension scores for the structured input without visual input enhancement group (+SI VIE) and the processing instruction with visual input enhancement group (+PI +VIE) These results mirror the findings of J.F. Lee (1998) and J.F. Lee and Rodrguez (1997). J.F. Lee (1998) found no correlation between words, and J.F. Lee and Rodrguez (1997) found a weak positive relationship between the two constr theoretically there should be a strong positive relationship between comprehension and input processing; thus far, no study has been able to support this claim. It is not possible to clai m that good comprehenders are also good input processors, or conversely that poor comprehenders are poor input processors. Discussion of Findings in Relation to the Research Hypotheses There were seven hypotheses related to the first two research question s in the present study, which are presented below. The hypotheses were formulated based upon previous empirical research and theory in the areas of processing instruction and visual input enhancement. The research studies and theory associated with these areas were presented and discussed in the review of literature. Each hypoth esis is listed below with a brief synopsis of whether the results of the present study either support or refute each hypothesis.

PAGE 300

286 Hypothesis 1 : L earners who are exposed to process ing instruction and structured input activities will outperform learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks over time. ( +PI VIE an d +SI VIE > TI) The results of the present study do not support Hypothesis 1, as there were no statistically significant differences between learners who received processing instruction and structured input activities and learners who received tradition al instruction as measured by interpretation test scores across time Hypothesis 2 : L earners who are exposed to processing instruction and structured input activities will perform as well as learners who are exposed to traditional instruction for the acqu isition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks over time. ( +PI VIE and +SI VIE = TI) The results of the present study support Hypothesis 2. T here were no statistically significant differences betwe en learners who received processing instruction and structured input activities and learners who received traditional instructi on as measured by production test scores across time Hypothesis 3 : L earners who are exposed to processing instruction and struc tured input activities with visual input enhancement will outperform learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by i nterpretation tasks over time. ( +PI +V IE and +SI +VIE > TI)

PAGE 301

287 Hypothesis 3 is not supported by the findings of the present research study, as there were no significant differences between learners who received processing instruction and structured input activities with visual input enhancement a nd learners who received traditional instruction as measured by interpretation te st scores across time Hypothesis 4 : L earners who are exposed to processing instruction and structured input activities with visual input enhancement will perform as well as learners who are exposed to traditional instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by production tasks over time. ( +SI +VIE and +SI +VIE = TI) The results of the present study support Hypo thesis 4. The present study did not find any significant differences between the groups as measured by production test scores ; thus, p articipants in the experimental groups performed equally as well as participants who received traditional instruction for produ ction tasks, even though participants in the experimental groups never produced the targeted forms during their instructional treatments Hypothesis 5 : L earners who are exposed to processing instruction and structured input activities with visual input enhancement will outperform learners who are exposed to processing instruction and structured input activities without visual input enhancement for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by i nterpretation ta sks over time. ( +PI +VIE and + SI +VIE > +PI VIE and +SI VIE) Hypothesis 5 is partially supported by the findings of the present research study. The processing instruction with visual input enhancement group had significantly higher

PAGE 302

288 mean interpretation test scores than the structured input without visual input enhancement group from Pretest to Posttest 1. However, there were no significant group differences between the processing instruction with visual input enhancement group and the following groups: the processing instruction without vi sual input enhancement group, the structured input with visual input enhancement group and the traditional instruction group It is also important to note that significant group diff erences were not retained from Pos ttest 1 to P osttest 2. Hypothesis 6 : L earners who are exposed to structured input activities alone will not perform as well as learners who are exposed to processing instruction for the acquisition of the present subjunctive in adjectival clauses in Spani sh as measured by interpretation tasks over time. ( +SI VIE < +PI VIE ) The results of the research study do not support Hypothesis 6. There were no significant differences in mean interpretation test scores between the structured input without visual i nput enhancement group and the processing instruction without visual inp ut group across time. Hypothesis 7 : L earners who are exposed to structured input activities combined with visual input enhancement will perform as well as learners who are exposed to processing instruction for the acquisition of the present subjunctive in adjectival clauses in Spanish as measured by interpretation tasks over time. ( +SI +VIE = +PI VIE ) Hypothesis 7 was supported by the results of the present research study. The r e were no significant differences in mean interpretation test scores between the structured

PAGE 303

289 input with visual input enhancement group and the processing instruction without visual input enhancem ent across time Hypothesis 8a : L earners who are exposed to processing instruction and structured input activities with visual input enhancement will notice more targeted verb forms that are embedded in a subsequent authentic input passage than learners who are exposed to traditional instruction and learners who ar e exposed to processing instruction and structured input activities without visual input enhancement as measured by note scores. (+ PI +VIE, +SI +VIE > +PI VIE, +SI VIE, and +TI) Hypothesis 8 a was not supported by the results of the present research stu dy. There were no significant differences in mean note scores by instructional treatment group. Hypothesis 8b : L earners who are exposed to processing instruction and structured input activities with visual input enhancement will have a higher level of aw areness (or a deeper level of noticing) of the targeted verb forms that are embedded in a subsequent authentic input passage than learners who are exposed to traditional instruction and learners who are exposed to processing instruction and structured inpu t activities without visual input enhancement as measured by awareness scores. ( +PI +VIE, +SI + VIE > TI, +PI VIE and +SI VIE) The results of the present study partially support Hypothesis 8b. The processing instruction with visual input enhancement gr oup had higher mean grammar comprehension scores than the following groups: traditional instruction, structured input without visual input enhancement, and structured input with visual input enhancement.

PAGE 304

290 However, there were no significant differences in m ean grammar comprehension scores between the processing instruction with visual input enhancement group and the processing instruction without visual input enhancement group. Hypothesis 9a : Learners who are exposed to processing instruction and structured input activities with and without visual input enhancement will perform as well as learners who are exposed to traditional instruction for comprehending the message of a subsequent authentic input text in which the targeted grammatical form is embedded as measured by text comprehension scores. ( +PI +VIE, +PI VI, +SI +VIE, +SI VIE = TI) The results of the present research study support Hypothesis 9a. There were no significant differences in mean text comprehension scores by group. Hypothesis 9b : Le arners who are exposed to processing instruction and structured input activities with and without visual input enhancement will outperform learners who are exposed to traditional instruction for processing targeted forms that are embedded in a subsequent a uthentic input text as measured by grammar comprehension scores ( +PI +VIE, +PI VI, +SI +VIE, +SI VIE > TI) Hypothesis 9b is not supported by the results of the present research study. There were no significant differences in mean grammar comprehension scores between the four experimental groups and the comparison group (traditional instruction). However, the two processing instruction groups (with and without visual input enhancement) outperformed the structured input without visual input enhancement g roup as measured by grammar comprehension scores.

PAGE 305

291 Hypothesis 10 : T here will be a significant positive correlation between input processing and text comprehension. The results of the present study do not support Hypothesis 10. The traditional instructio n group and the structured input with visual input enhancement groups had no correlation between text and grammar comprehension scores. The processing instruction without visual input enhancement group demonstrated a weak negative relationship between tex t and grammar comprehension scores, and the processing instruction with visual input enhancement group as well as the structured input without visual input enhancement group demonstrated a weak positive correlation between text and grammar comprehension sc ores. Discussion of Results for Interpretation and Production of the Indicative Before the instructional treatments took place, participan ts already understood how to form and use the indicative mood in Spanish. During the instructional treatments, they were required to make numerous subjunctive versus indicative contrasts. In order to deter mine if participants overgeneralized the subjunctive by using subjunctive forms in adjectival clauses when indicative forms were required scores from the interpretat ion of the indicative component of the Interpretation Subtest and scores from the production of for interpretation or production of the indicative decreased over tim e, it could indicate that participants overgeneralized the subjunctive The findings of the repeated measures ANOVA that was performed on scores for interpreting the indicative revealed that although there were no significant differences

PAGE 306

292 between the group s over time, there was a significant main effect for time. Post hoc contrast tests revealed that the overall mean interpretation test score was significantly higher at both Posttest 1 and Posttest 2 compared to Pretest However, the overall mean interpre tation test sco re was no t significantly different from Posttest 1 to P osttest 2. These results indicate that the instructional treatments helped participants improve their use of the indicative in adjectival clauses in Spanish over time and subjunctive f orms do not appear to have been overgeneralized. Regarding production of the indi cative, the results of the repeated measures ANOVA did not reveal any significant dif ferences in mean production test scores over time by instructional treatment group. Howe ver, the re was a significant main effect for time. Post hoc contrast tests revealed that the overall mean production test score was significan tly higher at Posttest 1 compared to Pretest. Conversely, the overall mean production test score was not signif icantly different at Posttest 2 compared to Posttest 1 or at Posttest 2 compared to Pretest. It appears that exposure to the instructional treatments helped participants improve in their production of the indicative in adjectival clau ses in the short term but the improvements were not retained ove r time. As the overall mean production test score did not decrease over time, it does not appear that over generalization of the subjunctive took place as a result of exposure to the instructional treatments.

PAGE 307

2 93 The oretical Implications Processing Instruction and the Spanish Subjunctive The findings of the present study have a number of theoretical implications. First and foremost, the present study did not find that learners who received processing instruction and /or structured input activities, with or without visual input enhancement, outperformed learners who received traditional instruction for interpretation t asks; rather, there were no significant differences between the experimental groups and the traditiona l instruction group as measured by interpretation tasks. For production tasks, however, the present study found that learners who received processing instruction and structured input activities, with or without visual input enhancement, performed equally as well as learners who received traditional instruction. Thus, the results of the present study only partially support the findings of past studies that compared PI with traditional instruction (Benati, 2001, 2005; Cadierno, 1995; VanPatten & Cadierno, 1 993a, 1993b; VanPatten & Wong, 200 4), as these studies found that PI is superior to traditional instruction for the interpretation o f targeted forms and that PI is equal to traditional instruction for the production of targeted forms. The results of the b ) findings; he found that both PI and traditional instruction were equally effective for the acquisition of complex Spanish grammar for both interpretation and production tasks. Collentine also investigated the subjunctive in adjectival clauses; however, his research study was heavily criticized by VanPatten and his colleagues because he failed to maintain treatment fidelity to PI. In particular, he did not follow the appropriate guidelines for the

PAGE 308

294 developm ent of structured input activities, which are a key component of PI. Collentine (1998) asserts that the pushed output tasks that were included in his traditional instruction atical form. In the present study, 80% of the activities that comprised traditional instruction did not represent authentic communication in the TL (40% of the activities were mechanical or transformational drills and 40% were meaningful drills). However 20% of the activities were open ended communicative, where learners were required to produce sub junctive forms either verbally or in writing in order to communicate a message in the TL. The two open ended communicative activities represented authentic c ommunication in the TL, and during these activities learners may have become aware of gaps in their interlanguage knowledge regarding the appropriate use of the subjunctive in adjectival clauses. Swain and Lapkin (1995) assert that the production of outp ut compels learners to or do not know in the TL. Therefore, when learners attempt to produce output, they notice what they do not know, which prompts them to pay cl oser attention to relevant forms and structures in subsequent L2 input. Swain (1985) claims that pushed output forces learners to shift to a deeper level of language processing and that the act of producing either spoken or written language contributes to the acquisition process. It is possible that the two open ended communicative activities in the traditional instruction

PAGE 309

295 facilitated the acquisition process which could have blurred the differences between the effects of processing and traditional instruction for interpretation tasks. O ther studies in the PI strand that examined the subjunctive mood did not compare PI with traditional instruction; rather, Farley (2001a, 20 01b) compared PI with meaning based output instruction (MOBI) for the acquisition of the subjunctive in nominal clauses following expressions of doubt. Rather than examine traditional instruction where most of the activities focus on form rather than on m eaning, MOBI required learners to focus on both meaning and form simultaneously, which is similar to the structured input activities that comprise PI. While Farley (2001a) found that PI was superior to MOBI for interpretation tasks and equal to MOBI for p roduction tasks, Farley (2001b) did not find any significant differences between the PI group and the MOBI . there is something about the nature of the subjunctive that causes the results in o ur two studies [ his and ] to be different from other PI ). Thus, PI may be more effective than traditional instruction for grammatical forms that are simple, but for complex forms such as the Spanish sub junctive, it appears that both PI and traditional instruction are equally effective. Further, there was no significant difference between the structured input groups with or without visual input enhancement (+SI +VIE, +SI VIE) and the processing instru ction without visual input enhancement group (+PI VIE) as measured by interpretation and production tasks over time. The key difference between full PI and structured input activities is that PI provides learners with an explicit explanation of

PAGE 310

296 grammar a nd information on processing strategies before learners begin structured input activities. The results of the present study indicate that when PI is delivered without VIE, the presence of an explicit grammar explanation and the provision of information on processing strategies do not result in significant gains when compared to groups that only received structured input a ctivities, with or without VIE. Processing Instruction and Input Enhancement The present study combined PI with computerized visual i nput enhancement in an effort to increase the salience of subjunctive forms for web based delivery. Past studies that utilized input enhancement employed simple typographical enhancements such as bolding and underlining, which may not be effective for cap multimed ia and web based environments where learners are often exposed to multiple layers of information such as text, video, and audio simultaneo usly. The present study attempted to optimize the capabilities of the web based learning environment by using flash programming language to create word animation, where the movement of sentences online. A meta analysis on the efficacy of VIE conduc ted by S. Lee and Huang (200 8 ) only found a very small positive effect for VIE on grammar learning, ( d = .22). In addition, scholars such as J. White (1998), Izumi (2002), and Hwu (2004) assert that VIE should be combined with other pedagogical techniques because VIE is more effective for facilitating noticing rather than learning of targeted forms. Given the results of the meta analysis and the assertions of SLA scholars regarding the efficacy of VIE, the present study did not examine VIE in isolation. Rather, VIE was combined with

PAGE 311

297 structured input, which is another type of input enhancement technique. VIE was structured input activities while the structured input activities themselves were designed to help learners process subjunctive forms correctly once they were noticed. Hence, the two types of input enhancement techniques were designed to work synergistically in the present study. The results of the present study indicate that learners who were exposed to processing instruction with visual input enhancement (+PI +VIE) outperformed learners who were exposed to structured input wit hout visual input enhancement (+SI VIE) on Posttest 1 compared to Pretest as mea sured by scores for interpreting the subjunctive. With PI learners were provided with an explicit explanation of grammar and information on processing strategies, which participants in the structured input groups did not receive. Sharwood Smith (1991) a sserts that grammar explanations that provide metalinguistic rule explanations are an explicit and elaborate form of input enhancement. Thus, learners who received processing instruction with visual input enhancement (+PI +VIE) actually received four type s of enhanced input: (a) metalinguistic information in the form of an explicit explanation of L2 grammar rules, (b) metalinguistic information about processing strategies, (c) structured input, and (d) computerized visual input enhancement operationalized as word animation of targeted verb forms. Conversely, participants who received structured input without visual input enhancement (+SI VIE) only received one type of input enhancement: structured input. Learners who were exposed to structured input with out visual input enhancement did not receive an explicit explanation of

PAGE 312

298 grammar rules, information on processing strategies, or word animation of targeted forms. Consequently, processing instruction with visual input enhancement (+PI +VIE) is a highly exp licit type of focus on form instruction with multiple layers of input enhancement while structured input without visual input enhancement (+SI VIE) only contains one layer of input enhancement and presumes that gr ammar will be learned inductively through exposure to structured input activities alone The findings of the present study indicate that instructional techniques that are highly explicit and that contain multiple layers of input enhan cement are superior to inductive instructi onal techniques that only contain one lay er of enhanced input for learning complex L2 grammar online. Interestingly, the present study found that participants who received processing instruction with visual input enhancement (+PI +VIE) did not outperform participants who rece ived processing instruction without visual input enhancement (+PI VIE) structured input with visual input enhancement (+SI +VIE) or traditional instruction (+TI) for inte rpreting the subjunctive in the short term T raditional instruction provided learn ers with an explicit explanation of L2 grammar (an elaborate form of input enhancement), str uctured input with visual input enhancement provided learners with two layers of input enhancement: structured input and word animation. Processing instruction wit hout visual input enhancement provided learners with three layers of input enhancement: an explicit explanation of L2 grammar, information on processing strategies, and structured input. These findings support the claim that language learners

PAGE 313

299 benefit f ro m exposure to multiple layers of input enhancement for learning complex L2 grammar online. Processing Instruction and Authentic Input Past studies in the PI strand have not investigated whether PI is able to facilitate how learners notice and process targ eted grammatical forms when they are embedded in authentic input that is received after exposure to PI treatments. Collentine (2004) states, . we do not know if learners respond to forms constituting the targeted grammatical phenomenon in normal inp ut conditions (i.e., authentic input) once they have left the Collentine asserts that it is impossible to determine if PI has a beneficial effect on the acquisition process until there is evidence of h exposure to PI. Thus far, all experiments in the PI strand, including the present study, provided participants with input that was structured, or manipulated, during the experimental tre atments in order to facilitate input processing. For example, many of the subjunctive forms in the instructional activities of the present study appeared in the sentence initial position in order to facilitate noticing for the experimental groups. Howeve r, in authentic input the subjunctive in adjectival clauses always appears in the sentence medial position, which is the most difficult place for learners to notice it. Further, structured input activities tend to embed targeted forms within short input s entences that contain basic authentic input passage that participants received post experimental exposure in the

PAGE 314

300 present study contained longer input sentences as wel l as more advanced vocab ulary items and colloquial expressions to which the participants were not accustomed. One of the main goals of the present study was to determine if PI facilitates noticing and pro cessing of targeted forms that we re p resent in a uthentic input that wa s received following the instructional treatments. By exposing participants to an authentic input text that contained 15 subjunc tive forms in the adjectival clause where the referent was uncertain, hypothetical, or unknown one to thr ee days after participants completed their experimental treatment s the present study was able to provide evidence that PI has a beneficial effect when learners encounter targeted forms that are present in subsequent authentic input. Although the authenti c input passage was more difficult for participants to comprehend than the input that they received during the experimental treatments, the results of the present study indicate that exposure to processing instruction with visual input enhancement facilita ted a dee per level of noticing (noticing with metalinguistic awareness) than exposure to traditional instruction or exposure to structured input with or without VIE In addition, exposure to processing instruc tion, with or without VIE resulted in correct input processing of targeted forms when they appeared in subsequent authentic input The present s tudy found that both processing instruction group s outperformed the structured input without v isual input enhancement group as measured by grammar comprehen sion scores (input processing). Attention and Awareness in SLA The results of the present study have theoretical implications for the areas of attention and awareness in SLA. The concept of attention has been a matter of

PAGE 315

301 controversy in the field of SLA o ver the past two decades. Some scholars assert that SLA is a conscious process (Robinson, 1995; Schmidt, 2001; VanPatten; 1994) and that focal attention is a necessary prerequisite in the noticing, storing, and learning of TL forms. Others (Krashen, 1980 1981, 1985; Schachter, 1998; Truscott, 1998) contradict this view and hypothesize that SLA is a largely unconscious process with learners acquiring TL forms subliminally, or without focal attention. The Noticing Hypothesis (Schmidt, 1990, 1993, 1995; Schmidt & Frota, 1986), (1981, 1991) input (1993, 1996, 2002, 2004) model of i nput p rocessing, states that learners must notice features of their TL inpu t with conscious attention before form learning can take place. A major goal of both input enhancement otherwise notice. However, Truscott (1998), one of the m Noticing Hypothesis, asserts that second language learning oc curs sub consciously, with conscious noticing merely leading to the acquisition of metalinguistic knowledge, or knowledge about the language, and not to the abi lity to use the language. Further, he suggests that metalinguistic knowledge may actually impede L2 performance, although he acknowledges that more research is needed in this area before any definitive claims can be made about the role of metalinguistic k nowledge in SLA. Schachter (1998) also refutes the belief that learners must attend to all linguistic input with focal attention. She claims that while certain aspects of L2 learning require conscious attention, namely the

PAGE 316

302 learning of individual sounds, individual words, and writing systems, the bulk of learning with respect to phonological, morphological, and syntactic rules does not require focal attention on the part of the learner. Further, Schachter (1998) suggests that it is practically imposs ible to ascertain whether the language acquisition process for children or adults is In contrast to the aforementioned views, Schmidt (1995, 2001) set forth the strong version of the Noticing Hypothesis, which states that there is no subliminal learning in SLA, although there may be subliminal perception. In other words, nearly all L2 learning is a conscious process that requires at least some attention to form. There has been 1994; N. Ellis, 1994, 1996; R. Ellis, 1997; Gass, 1988, 1997; Hatch 1983; Pie nemann, 1989; Pienemann & Johsnson, 1987; Robinson, 1995; Skehan, 1998, Swain, 1993, 1995; VanPatten, 1990, 1994, 1996; Wolfe Quintero, 1992), in particular research that stems from cognitive accounts of L2 development. Research from this theoretical fram ework presumes that conscious attention is what mediates input and intake, especially given that not all input becomes intake for learning (Gass, 1997; VanPatten, 1994). Further, attention to input is essential for storage into short term memory, and it i s a critical precursor to hypothesis formation and testing (Schmidt, 2001). It is important to note Noticing Hypothesis; Carr and Curran (1994) and Gass (1997) deny that all L2 learning

PAGE 317

303 requires focal attention. Conversely, VanPatten (1994) asserts that attention is the only necessary and sufficient condition for learning L2 structure. In addition to the disagreement over the role of attention in SLA, there is also c ontroversy over the amount and type of attention that is needed for L2 learning. Schmidt (1990, 1993, 1995) contends that L2 learners must allocate attention with awareness for learning to take place. Further, he argues that learning can occur without in tention and without knowledge of metalinguistic rules, but it cannot occur without awareness. include attention plus rehearsal in short term memory According to Robinson, focal at tention with rehearsal in short term memory is necessary but insufficient for SLA to take place. In contrast, Tomlin and Villa (1994), who drew upon the work of Posner (1994) and Posner and Peterson (1990), hypothesize that there are three subsystems of a motivation, interest, and overall readiness to learn, while orientation is associated with an allocation of attentional resources to form, which may increase the likelihood of detection. Tomlin and Villa (1994) assert that alertness and orientation may assist with detection, which is the cognitive registration of stimuli, but only detection is necessary for further L2 processing and learning. Additionally, detection may o c cur with or without awareness. D etection without awareness signifies the mere registration of word. Interestingly, Tomlin and Villa (1994) theorize that only detec tion without

PAGE 318

304 awareness is necessary for L2 processing and learning, a point of view that directly opposes the Noticing Hypothesis. the participants in his study were able to make generalizations in an artificial grammar even though they expressed no awareness of the targeted forms during retrospective self ( p. 2 72 ) However, Leow and Hama (2008) replicated to use think aloud protocols to measure awareness rather than retrospective self reports. Leow (2000) found that particip ants who demonstrated noticing at the level of understanding were able to make generalizations, but those who demonstrated awareness at the level of noticing were not. Izumi (2002) also found that learners who noticed targeted forms did not necessarily l earn them. Izumi compared the production of output with exposure to textual input enhancement for the acquisition of relative clause formation in English. He measured noticing by examining the notes that participants took while they read an input passage in the TL, which he converted to note scores. Izumi found that participants who were exposed to textual input enhancement operationalized as underling, bolding, and/or a change of font style or size demonstrated significantly higher note scores than part icipants who produced output; however, he also found that the increased noticing did not lead to learning of targeted forms. Textual enhancement appeared to facilitate

PAGE 319

305 been unsure about their meaning or relevance, which could have prohibited them from making the necessary form meaning connections for intake to occur. Izumi did not t hat took place. Thus, it is possible that the parti cipants in his study only had a low level of awareness (simple detection) rather than awareness at the level of understanding. According to Schmidt (2001), noticing, by itself, is not the only necessar y and notice in target language input and what they understand the significance of noticed Similarly, VanPatten (1993, 1996, 2002, 2004) posits t hat in addition to noticing linguistic features of input, learners must also be able to make form meaning connections, or understand the relationship between a linguistic form and the referential meaning that it encodes. Further, N. Ellis (1994) asserts t hat form meaning connections recall might play a large part in L2 learning Empirical research supports the facilitative effects of awareness on foreign language learn 1999) Further, Leow (2000) asserts that awareness at the level of noticing is a necessary precursor for learning, but awareness at the level of understanding plays a critical role in le grammatical forms. The findings of the present study lend weight to these assertions and reveal that learners who demonstrated awareness at the level of understanding were also better processors of L2 i nput. Thus, the results of the present study indicate that when learners consciously

PAGE 320

306 attend to targeted forms and notice them with understanding, there is a positive facilitative effect on input processing. Pedagogical Implications The results of the pre sent study offer a number of pedagogical implications. The first major pedagogical implication is that distance language learners benefit from having a combination of the following features for short term form learning of complex L2 grammar: an explicit explanation of grammar that is not paradigmatic, information on processing strategies, structured input activities, and computerized visual input enhancement. The results of the study indicate that processing instruction with visual input enhancement (+PI +VIE) was superior to structured input without visual input enhancement (+SI VIE) for short term learning gains on interpretation tasks. Interestingly, the most explicit and elaborate instructional method was found to be significantly better than the le ast explicit and least enhanced method when the immediate effects of the exper imental treatments were examined for interpreting the subjunctive. It appears that providing distance language learners with multiple layers of input enhancement, including a me talinguistic explanation of grammar rules, is beneficial. Sharwood Smith (1991) posits that providing an explanation of grammar rules is as an elaborate way to enhance input using the technical terminology that describes language. C White (2003) asserts that under the traditional distance learning paradigm, where the emphasis is on independent learning and self instruction, learners rely on the course materials rather than on their teacher for instruction. Thus, web based mate rials that are highly expli cit and elaborate, such as processing instruction with visual input

PAGE 321

307 enhancement, appear to facilitate the self instruction of complex aspects of L2 grammar in the short term. T he results of the present study suggest that combining computerized visual inp ut enhancement with processing instruction is an effective way to teach complex L2 grammar online. In the present study, computerized visual input enhancement was operationalized as word animation of targeted verb forms. This type of input enhancement te chnique optimized the capabilities of the web based learning environment, and it would not have been possible with the traditional print medium. The present study also found that exposure to processing instruction with visual input enhancement (+PI +VIE) resulted in significantly higher awareness of targeted forms in subsequent authentic input than exposure to the following techniques: traditional instruction (+TI) structured input without visual input enhancement (+SI VIE) and structured input with visual input enhancement (+SI +VIE) These results indicate that ex posure to PI helps learners notice and process targeted forms when they encounter them in subsequent authentic input. This finding has clear implications for foreign language pedagog y, especially for instruction that stems from the Communicative Language Teaching Approach (CLT), which is a teaching philosophy that is advocated by the American Council on the Teaching of Foreign Languages (2002) a professional organization that sets th e program standards for the preparation of foreign language teachers. CLT emphasizes authentic communication in the TL. Richards and Rodgers (1986) assert that CLT is founded upon three foundational principles as follows:

PAGE 322

308 1. The communication principle: Act ivities that involve communication promote language learning. 2. The task principle: Activities that involve the completion of real world tasks promote learning. 3. The meaningfulness principle: Learners must be engaged in meaningful and authentic language us e for learning to take place. (p. 72) Thus, the findings of this study indicate that exposure to processing instruction has the pot ential to facilitate the meaning fulness principle listed above, as PI appears to prime learners to notice and process targe ted forms that appear in subsequent authentic language input. Finally, learner preference should be taken into account when developing web based instr uctional materials. Responses from the Posttreatment Questionnaire revealed that learners who received PI expressed an overwhelming preference for the PI materials over their regular course materials. Participants in the two PI groups fe lt that the explicit grammar explanation and the information on processing strategies were extremely helpful for learning complex grammar online. According to C. White ( 2003) distance language the two most important factors for success in a distance language course. Thus, materials that learners perceive as being highly beneficial, such as information on processing efficacy and motivation to continue with the distance course. The results from the Posttreatment Questionna ire for the two structured input groups and for th e traditional instruction group were more mixed, with some participants

PAGE 323

309 expressing that the materials were highly effective for learning the subjunctive and others stating that they were not very effective for learning Spanish grammar online. In addition, learners across all five groups expressed dissatisfaction and frustration with their regular course materials, which fall under the traditional instruction paradigm. It is preferences when selecting or creating course materials because materials that learners perceive as being too difficult or lacking in clear explanations could dampen their feelings of self efficacy as well as their motivation to continue with the course. Another important consideration is that the present study was conducted entirely online. By and large, past studies in the PI strand were classroom based and used the individual rather than the class as the unit of analysis. With classroom based studies, a number of factors other than the instructional treat ment, such as interaction between participants, could potentially influence the outcome of the study. According to Stevens (2002), even a small amount of dependence am ong observations can cause the Ty pe I error rate to increase several times greater than the level of significance. Conversely, with instructional method studies that are conducted online, the instructional treatment is individually administered and there is no interaction between partici pants. It is possible that the present study provides a clearer picture of the effects of processing instruction than prior classroom based studies Limitations of the subju nctive was not taken into account, visual input enhancement was not

PAGE 324

310 examined in isolation as an independent variable, the traditional instruction group received differential feedback for one activity, the instructional treatment was relatively short in dur ation, and awareness was measured retrospectively. Rather than take into account learners past knowledge of the subjunctive, the present study used a 60% cutoff level for scores on the Interpretation and Production Subtests, which was in keeping with past studies in the PI strand. However, a better the subjunctive by administering a pretreatment subjunctive knowledge test before the pretest. Thus, the pretreatment subju nctive knowledge test means and the pretest means could have served as covariates in the analysis of posttest means, which is the design that Collentine (1998) utilized. The present study did not take prior subjunctive knowledge into account because the s tudents enrolled in the Spanish classes at the two institutions that comprised the sample were required to take a placement exam if they took two or more years of Spanish in High School. If students had already mastered the subjunctive mood as evidenced b y their placement test score they would have been placed higher pretest, it was apparent that many students had intuitions about how the subjunctive mood is used in adjectiv al clause s. A total of 43 of the 44 participants that were excluded from the study due to their performance on the pretest scored 60% or higher on the Interpretat ion Subtest, which required learners to choose between subjunctive and indicative forms in or der to interp ret sentences correctly. Only one student from the initial pool scored 60% or higher on the Production Subtest Thus, if previous exposure

PAGE 325

311 to the subjunctive had been taken into account by administering a pretreatment subjunctive knowledge t est, then participants who sco red 60% or higher on the Interpretation Subtest could have had their prior knowledge accounted for statistically rather than be elim in ated from the study. Another limitation of the present study is that VIE was not examined i n isolation as an independent variable. Since the instructional techniques were grouped in the present study, it was not possible to determine if VIE had an effect on its own. The decision was made to pair VIE with structured input activities based on th e results of a meta analysis of past empirical research ( S. Lee & Huang, 200 8 ), which only found a very small effect for VIE on grammar learning. Theory also supported the decision to pair VIE with structured input activities as several prominent scholars in the field of SLA (Hwu, 2004; Izumi, 2002; J. White, 1998) assert that VIE should be used to promote noticing, or detection of targeted forms, while another pedagogical technique should be used to facilitate form learning. The present study found that for interpreting the subjunctive in the short term, the processing instruction with visual input enhancement group outperformed the structured input without visual input enhancement group. However, it was not possible to determine whether the explicit gra mmar explanation, the informatio n on processing strategies, the computerized visual input enhancement, or a combination of two or more of these factors was r esponsible for the significant differences that were found. In addition, although the processing i nstruction with visual input enhancement group outperformed the structured input without visual input enhancement group for interpreting the subjunctive in the short term, the processing

PAGE 326

312 instruction without visual input enhancement group did not. However, the processing instruction with visual input enhancement group did not outperform the processing instruction without visual input enhancement group on any of the measures used in the present study (interpretation scores, production scores, note scores, aw areness scores, and grammar comprehension scores). Thus, VIE could be responsible for the significant Group x Time interaction effect th at was found for interpreting the subjunctive but due to the design of the present study, it is not possible to make a ny definitive claims regarding the efficacy of VIE in isolation. The differential feedback that the traditional instruction group received for the open ended communicative activity that required an oral response is another limitation of the present study. With this activity, participants were required to record an oral response to five prompts, and correct answers necessitated a subjunctive verb form. Answers were recorded using an audio drop box that was embedded on a web page within the instructional m aterials. The traditional instruction participants received delayed feedback for this activity as voice recognition software is still not widely available, and its use in the present study would have been too costly an d time consuming to implement. As v oice recognition technology improves and becomes less cost ly and more readily available, future studies will be able to avoid this design flaw. In an attempt to mitigate the effects of receiving delayed feedback, participants in the traditional instructio n group were asked to self reflect on their answers to the oral activity. For all other study activities, participants in the comparison and experimental groups received implicit feedback that was immediate In other words, participants were only told if their answers

PAGE 327

313 were correct or incorrect, and they were not given the correct forms if the answers that they provided were incorrect Correct forms were not given to participants in order to avoid providing them with incidental input of the targeted gramm atical form. Another limitation of the present study is the relatively short treatment period that participants had to learn the targeted grammatical form. The mean time for all 92 participants to complete their instructional treatment s was 75.15 minutes However, the second semester face to face (FTF) Spanish classes that are offered at the participating institutions meet for 50 minutes four times per week. Teachers in the FTF courses typically spend no more than one 50 minute class period on the instr uction of the subjunctive in adjectival clauses (the targeted grammatical form in the present study.) Thus, while the treatment period in the present study was relatively short, it still provided participants with th e equivalent of 1.5 FTF classes of inst ruction on the targeted grammatical form. The final limitation that will be discussed is the data collection procedures that embedded in an authentic input text that wa s received after the experimental exposure A Posttreatment Q retrospectively in the present study. Leow (2000) criticizes this technique because it is an off line measure, and it may not ca pture what learners actually paid attention to or became aware of during the experimental exposure. The issue of how to operationalize and measure awareness is a thorny issue in SLA research. There are three prominent data collection procedures that are ess : (a)

PAGE 328

314 online elicitation measures such as think aloud protocols, (b) off line elicitation measures such as postexposure questionnaires, and (c) a combination of online and off line elicitation measur es (Leow, 2000, p. 559). The present study used an online elicitation present in an authentic input text. Participants were asked to write down all of the vocabulary it ems and verb forms that were necessary to comprehend the text, and the subjunctive forms that were noted were tallied to compute a note score. All 92 participants noted at least one subjunctive form during this activity. Thus, it was possible to establis h that all participants in the present study demonstrated at least a low level of awareness of the targeted grammatical form that was present in an authentic input passage. The post exposure questionnaire elicited metalinguistic information from participan ts to determine if they demonstrated awareness at the level of noticing and/or awareness at the level of understanding. In order to demonstrate awareness at the level of noticing, participants had to specifically state that the subjunctive mood was presen t in the input text and provide an example of it, and in order to demonstrate awareness at the level of understanding, participants had to explicitly state the grammatical rule for using the subjunctive in adjectival clauses. While two types of elicitatio n were used to assess noticing and awareness, it was not feasible to use think aloud protocols because the experiment was conducted entirely online and the present study employed a purely quantitative research design with a large number of participants. L eow (1997, 2000) asserts that think aloud protocols are the optimal way to collect information on

PAGE 329

315 aloud while completing a task has the potential to interfere with le processes. More recently (Matsunaga & Crosby, 1997; Godfroid, Boers, & Housen, 2008) eye tracking technology has been used to measure noticing of targeted forms, but this method is also flawed because it only reveals what participants dete ct in their L2 input and not whether they understand the significance of detected forms and structures. studies that use multiple data elicitation techniques and that an alyze data both quantitatively and qualitatively are preferable for assessing the construct of awareness in SLA. Suggestions for Future R esearch Future research studies are needed that examine the effects processing instruction and structured input activi ties qualitatively. By examining the effects of processing instruction and structured input with smaller, more focused samples, it may be possible to aquire it affect instr Posttreatment Questionnaire suggest that learners who identified themselves as poor language learners found processing instruction to be extremely helpful for learning complex Spanish grammar online. Future research could take into account individual differences such as age, gender, language apti tude, and developmental readiness when investigating the efficacy of processing instruction and structured input activities. Research s tudies are nee ded that investigate whether processing instruction is more

PAGE 330

316 effective than traditional instruction for learners with a low aptitude for language learning or for learners who are older. Further, more research studies are needed that examine computerized v isual input enhancement in isolation, as it was not possible in the present study to determine if this type of input enhancement had an effect on its own. The type of enhancement used in this study could be compared with other types of computerized input enhancement to formal features of their L2 input as they work online. Finally, more research studies are needed that examine the cumulative effects of processing instruction. The present study only investigated the effects of processing instruction and structured input for the acquisition of the subjunctive in adjectival clauses. However, future studies could examine the effects of repeated exposure to processin g instruction and its components. For example, processing instruction could be investigated over the course of a semester for the acquisition of the subjunctive in nominal clauses, followed by the subjunctive in adverbial clauses and the subjunctive adjec tival clauses. This type of study would help uncover whether the effects of processing instruction for the acquisition of the Spanish subjunctive are more durative with repeated exposure. Conclusion The results of the present study are encouraging for t he use of processing instruction combined with visual input enhancement for the instruction of complex Spanish grammar online. Although the analyses of the pre and posttests did not reveal

PAGE 331

317 any significant difference s between the experimental groups and traditional instruction for interpretation and production tasks over time, the findings of the present study suggest that instructional techniques that are highly explicit and that contain multiple layers of input enhancement are superior to instructional techniques that are inductive with only a single layer of input enhancement for short term learning of complex grammar online As distance language learners typically rely more heavily on the materials than on their teacher for instruction, techniques suc h as processing instruction with visual input enhancement appear to facilitate the self instruction process. In addition, the present study also examined whether the instructional treatments targete d forms that were embedded in an authentic input passage that was received following the experime ntal exposure. Thus far, past studies in the PI strand have only examined how learners interact with structured, or manipulated, input. The results of the pr esent study indicate that exposure to processing instruction increases the likelihood that learners will notice targeted forms in subsequent authentic input with metalinguistic awareness, which Leow (2000) claims nd subsequent processing of targeted forms. In addition to the beneficial effect on noticing, processing instruction also appears to facilitate correct input processing of targeted forms when learners encounter them in subsequent authentic input, which ha s the potential to facilitate authentic communication in the target language

PAGE 332

318 List of References Alanen, R (1995) Input enhancement and rule presentation in second language acquisition In R. Schmidt (Ed.), Attention and Awareness in Foreign Language L earning (Tech. Rep. No. 9, 259 302) Honolulu: University Allen, L. (2000) Form meaning connection s and the French causative: An experiment in processing instruction Studies in Second Language Acquisition 22, 69 84. Allen, M (2003) Learning Hoboken, NJ: John Wiley & Sons, Inc. Allen, E., & Seaman, J (2008) Staying the Course: Online education in the United States, 2008 (Babson Survey Research Group Report, (pp. 1 31) The Sloan Consortium. American C ouncil on the Teaching of Foreign Languages (2002). ACTFL/NCATE program standards for the preparation of foreign language teachers Anderson, J.R. (1976). Language, memory, and thought Hillsdale, NJ: Erlbaum. Andrews, D. Carroll, L., & Bell, H (1995 ) The future of selective fidelity in training devices Educational Technology 35, 32 36. Auble, P. M. & Franks, J. J. (1978) The effects of effort toward comprehension on recall Memory and Cognition 6, 20 25. Baek Y. K., & Layne, B. H. (1988). Col or, graphics, and animation in a computer assisted learning tutorial lesson. Journal of Computer Based Instruction, 15, 4, 131 135. Ba r croft, J., & VanPatten, B (19 9 7) Acoustic sali ence of grammatical forms: The effect of location, stress, and boundedn ess on Spanish L2 input processing In T. Prez Leroux & W. R. Glass (Eds.), Co ntemporary perspectives on the acquisition of Spanish: Vol 2 Production, processing, and comprehension (pp. 109 121) Somerville, MA: Cascadilla Press.

PAGE 333

319 Barrow, L. C. & Westley B. H. (1959) Comparative teaching effectiveness of radio and television Audio Visual Communication Review 7, 14 23. Benati, A (2001) A comparative study on the effect s of processing instruction and output based instruction on the acquisition of the Italian future tense Languag e Teaching Research 5, 95 127. Benati, A (2004a) The effects of structured input activiti es and explicit information on the acquisition of the Italian future tense In B. VanPatten (Ed.), Processing instruction: Theory, re search, and commentary (pp. 207 226) Mahwah, NJ: Erlbaum. Benati, A (2004b) The effects of processing instruc tion and its components on the acquisition of gender agreement in Italian Language Awareness 13, 2, 67 80. Benati, A (2005) The effects o f processing instruction traditional instruction, and meaningful based output instruction on the acquisit ion of the English past simple tense Language Teaching Research 9, 67 93. Bialystok, E (1994) Analysis and control in the development of second l anguage proficiency Studies in Second Language Acquisition, 16, 157 168. Blake, R (1985) Mood selection among Spanish speaking children ages 4 12 The bilingual review 10, 21 32. Blau, E. K (1990) The effect of syntax, speed and pauses on listening comprehension TESOL Quarterly, 24, 746 753. Bock, M (1978) Levels of processing of normal and am biguous sentences in different contexts Psychological Research 40, 37 51. Bosco, J (1986) An analysis of evaluations of interactive video Edu cationa l Technology 25, 7 16. Cadierno, T (1995) Formal Instruction from a processing perspectiv e: An investigation into the Spanish past tense Modern Language Journal 79, 179 193. Cam p bell, D. T., & Stanley, J. C (1963) Experimental and quasi experimen tal designs for research Chicago: Rand McNally. Caron, J. (1992). An introduction to psycholinguistics. Toronto: University of Toronto Press.

PAGE 334

320 Carr, T. & Curran, T (1994) Cognitive factor s in learning about structured sequences: Applications to syn tax Studies in Second Language Acquisition, 16, 205 230. Carterette, E. C. & Jones, M. H (1967) Visual and audi tory information processing in children and adults Science 156, 986 988. Chan, A., Travers, R. M. W., & Van Mondfrans, A. P (1965). The e ffects of colored embellishments of visual array on a simultaneously presented audio array Audio Visual Communication Review 13, 159 164. Chastain, K (1976) Developing second language skills: Theory to practice Chicago: Rand McNally Cheng, A (19 95) Grammar instruction and input processing: The acquisition of Spanish ser and estar Unpublished doctoral dissertation, University of Illinois at Urbana Champaign. Cheng, A (2002) The effects of processing instruction on the acquisition of ser and e star Hispania 85, 308 323. Clark, J. M. & Paivio, A (1991) Dual coding theory and education Educ ational Psychology Review 3, 149 210. Clifton, C., Frazier, L., & Ryner, K. (Eds.) (1994) Perspectives on sentence processing Hillsdale, NJ: Lawrence Elrbaum Associates. Cohen, J (1992) A power primer Psychological Bulletin, 112, 155 159. Collentine, J.G. (1995). The development of com plex syntax and mood selection abilities by intermediate level learners of Spanish Hispania 78, 122 136. Collen tine, J.G (1998a) Cognitive principles and CA LL grammar instruction: A mind centered, input approach CALICO journal, 15, 1 18. Collentine, J.G (1998b) Processing instruction and the subjunctive Hispania 81, 57 6 587. Collentine, J.G (2000) The re lationship between syntact ic and morphological abilities in FL learners of Spanish In R. Leow and C. Sanz (Eds.) Spanish Applied Linguistics at the Turn of the Millennium (pp. 20 35). Somerville, MA: Cascadilla Press.

PAGE 335

321 Collentine, J.G (2003) The develop ment of subjunctive a nd complex syntactic abilities among FL Spanish learners In B. Lafford and R. Salaberry (Eds.), Studies in Spanish Second Language Acquisition: the State of the Science (pp. 74 97). Washington, DC: Georgetown University Press. Collen tine, J G. (2004) Commentary: Where PI research h as been and where it should be going In B. VanPatten (Ed.), Processing inst ruction: Theory, research, and commentary (pp. 169 181) Mahwah, NJ: Erlbaum Corder, S (1981). Error Analysis and Interlanguage Oxford: Oxford University Press. Cowan, N (1995) Attention and memory: An integrated framework New York: Oxford University Press. Craik, F. I. M. & Lockhart, R. S. (1972) Levels of processing: A framework for memory research Journal of Verbal Lea rning and Verbal Behavior 11, 671 684. Craik, F. I. M. & Tulving, E (1975) Depth of processing and the retention of words in episodic memory Journal of Experimental Psychology 104, 268 294. de Gr a aff, R (1997). The eXperanto experiment: Effects of explicit instruction on second language acquisition Studies in Second Language Acquisition 19, 249 297. DeKeyser, R. M., Salaberry, M. R., Robinson, P., & Harrington, M (2002) What gets processed in processing instruction? A commentary on Bill VanPat Language Learning 52, 805 823. DeKeyser, R. M., & Sokalski, K. J (1996) The differential role of comprehension an d production practices Language Learning 46, 613 642. Doughty, C (1988) The effect of inst ructon on the ac quisition of relativization in English as a second language Unpublished doctor al dissertation, University of Pennsylvania, Philadelphia. Doughty, C (1991) Second language instruction do es make a difference: Evidence from an empirical st udy of SL relativization Studies in Second Language Acquisition 13, 431 469. Doughty, C (2004) Commentary: When PI is focus on form it is very very good, but when it is focus on forms . In B. VanPatten (Ed.), Processing instruction: Theory, resea rch, and commentary (pp. 257 270) Mahwah, NJ: Erlbaum.

PAGE 336

322 Ellis, N (1993) Rules and instances in foreign language learning: Interaction of explicit and implicit knowledge European Journal of Cognitive Psychology 5, 289 318. Ellis, N (1994) Consciou sness in second language learning: Psychological perspectives o n the role of conscious processes in vocabulary acquisition AILA Review, 11, 37 56. Ellis, N (1996) Sequencing in SLA: Phonological memory, chunking, and points of order Studies in Second Language Acquisition, 18, 91 126. Ellis, R (1994) The Study of Second Language Acquisition New York: Oxford University press. Ellis, R (1995) Interpretation Tasks for Grammar Teaching TESOL Quarterly 29, 1, 87 105. Ellis, R (1997) SLA research and second language teaching Oxford: Oxford University Press. Erlam, R (2003) Evaluating the relative effectiveness of structured input and output based instruction in foreign language learning Studies in Second Language Acquisition 25, 559 582. Erv in Tripp, S (1974) Is second language learning like the first? TESOL Quarterly 8, 111 127. Frch C., & Kasper, G (1986) The role of co mprehension in second language learning Applied Linguistics 7, 257 284. Farley, A (2001a) Authentic processi ng instruction and the Spanish subjunctive Hispania 84, 289 299. Farley, A (2001b) Processing instruction and meani ng based output instruction: A comparative study Studies in Applied Linguistics 5, 57 93. Farley, A (2002) Processing instruction, communicative v alue, and ecological validity: Hispania 85, 889 895. Farley, A (2004a) The relative effects of processing instruction and meaning based output instruction In B. VanPatten (Ed.), P rocessing Instruction : Theory, Research, and Commentary (pp. 227 239) Mahwah, NJ: Lawrence Erlbaum Associates.

PAGE 337

323 Farley, A (2004b) Processing instruction and the Sp anish subjunctive: Is explicit information needed? In B. VanPatten (Ed.), P rocessing Instruction: Theory, Rese arch, and Commentary (pp. 227 239) Mahwah, NJ: Lawrence Erlbaum Associates. Fern ndez, C (2008) Reexamining the role of expl icit information in processing instruction Studies in Second Language Acquisition 30, 3, 277 305. Fletcher, D (1989) The ef fectiveness and cost of interactive videodisc instruction Machine Mediated Learning 3, 361 385. Fletcher, D (1990) The effectiveness and cost of intera ctive videodisc instruction in defense training and education (IDA Paper P 2372). Alexandria, VA: In stitute for Defense Analysis. Fotos, S (2002) From theory to practice: A teacher's view. In E. Hinkel & S. Fotos (Eds). New Perspectives on Grammar Teaching in Second Language Classrooms (pp. 1 12) Mahwah: Lawrence Erlbaum Associates. Fraser, A (199 9) Colleges should tap the pedagogical potential of the world wide web Chronicle of Higher Education 48, B8 Gall, M. D., Gall, J. P., & Borg, W (2007) Educational research ( eighth edition ) New York: Pearson. Gass, S. (1988) Integrating research ar eas: A framework for second language studies Applied Linguistics 9, 198 217. Gass, S. (1997) Input, interaction, and the second language learner Hillsdale, NJ: Erlbaum. Gibson, C (1998) The distance learner in context In C. Gibson (Ed.), Distance learners in higher education: Institutional responses for quality outcomes (pp. 113 125) Madison, WI: Artwood Publishing. Godfroid, A., Boers, F., & Housen, A (2008) Noticin g new words while reading in a foreign language: An eye tracking study into qua ntitative differences Paper presented at the American Association for Applied Linguistics Annual Conference, Washington DC March 30, 2008.

PAGE 338

324 Ha, J. (2005). Developing English determiners through Internet chat: An experiment with Korean EFL students. Unpu blished doct oral dissertation, University of Florida, Gainesville. Hannifin M. J. & Peck, K. L. (1988) The design, development and evaluation of instructional software NY: McMillan Publishing Company. Harrell, W (1998) Language learning at a distance via computer Indian Journa l of Open Learning 7, 2, 179 190. Harrington, M (2001) Sentence processing In P. Robinson (Ed.), Cognition and Second Language Instruction (pp. 3 32) New York: Cambridge University Press. Harrington, M (2004) Commentary : Input processing as a theory of processing input. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 79 92) Mahwah, NJ: Erlbaum. Harris, C (1995) What do the learners think: A study of how learn ers define successful learning at a distance In S. Gollin (Ed.), Language in distance education: How far can we go? Proceed ings of the NCELTR Conference, Sydney: NCELTR. Hatch, E (1983) Simplified input and second language acquisition In R. W. Anderso n (Ed.), Pidginization and creolization as language acquisition (pp. 64 88) Rowley, MA: Newbury House. Hawkins, R., & Trowell, R (1992) Second languag e acquisition research and the second language acquisition of french French Language Studies 2 97 1 21. Hong, W., Thong, J., & Tam, K (2004) Does animation attract onlin The effects of flash on information search performance perceptions Information Systems Research 15,1, 60 86. Hulstijn J. & Hulstijn W (1984) Grammatical er ro rs as a function of processing constraints and explicit knowledge Language Learning 34 24 43. Hunt, R. R. & Elliot, J. M. (1980) The role of nons emantic information in memory: Orthographic distinctiveness effects on retention Journal of Experiment al Psychology 109, 49 74. Hwu, F (2004) On the applicability of the input en hancement hypothesis and input processing theory in multimedia CALL: The case of Spanish preterite and imperfect instruction in an input application CALICO Journal 21, 2, 317 338.

PAGE 339

325 Izumi, S. (2002) Output, Input Enhancement, and the Noticing Hypothesis: An Experimental Study on ESL Relativization Studies in Second Language Acquisition 24, 541 577. Izumi, S., & Bigelow, M (2000) Does output promot e noticing and second la nguage acquisition? TESOL Quarterly 34, 2, 239 278. Izumi, S, Bigelow, M., Fujiwara, M., & Fearnow, S (1999) Testing the output hypothesis: Effects of output on noticing and second language acquisition Studies in Second Language Acquisition 21 421 452. Jagaroo, V., & Wilkinson, K (2008) Further cons iderations of visual cognitive neuroscience in aided AAC: The potential role o f motion perception systems in maximizing design display Augmentative and Alternative Communication 24, 1, 29 42. Johnso n, K. (2001). An introduction to foreign language learning and teaching. Reading, MA: Pearson Education Limited. Jourdenais, R (1998) The effects of textual enhance ment on the acquisition of the Spanish preterit and imperfect Unpublished doc toral dis sertation, Georgetown University, Washington, D.C. Jourdenais, R., Ota, M., Stauffer, S, Boyson, B., & Doughty, C (1995) Does textual enhancement promote noticing? A think aloud protocol analysis In R. Schmidt (Ed.), Attention and Awareness in Foreign Language Learning (Tech. Rep. No. 9, pp. 259 302) Just, M. A., & Carpenter, P. A (1992) A capacity theory of comprehension: Individual differences in working memory Psychological Review 99, 122 149. Klein, W (1986) Second Language Acquisition C ambridge: Cambridge University Press. Kolers, P. A. (1979) A pattern analyzing basis of recognition In L S. Cermak & F. I. M. Craik (Eds.), Levels of processing in human memory (pp. 363 384) Mahwah, NJ: Erlbaum K rashen, S (1980) The input hypothesis In. J. Alatis (Ed.), Current issues in bilingual education (pp. 175 183) Washington D.C.: Georgetown University Press.

PAGE 340

326 Krashen, S. (1981) Second Language Acquisition and Second Language Learning Oxford: Pergamon Krashen, S. (1982) Principles and practice in second language acquisition Pergamon. Krashen, S. (1985) The Input Hypothesis: Issues and Implications New York: Longman. Kubota, S. (2000). Input enhancement in classroom second language acquisition o f Japanese. Unpublished doctoral dissertation. Purdue University, West Lafayette, IN. Larsen Freeman, D. & Long, M. (1991) An introduction to second language acquisition research London: Longman. Lee, J. F (1987) The Spanish subjunctive: An informati on processing perspective The Modern Language Journal 71, 50 57. Lee, J. F. (1998) The relationship of verb morphology to second language reading comprehension and input processing. The Modern Language Journal 82, 33 48. Lee, J. F (2002) The In cidental Acquisition of Spanish: Future Tense Morpho logy Through Reading in a Second Language Studies in Second Language Acquisition, 24 55 80. Lee, S (2007) The effects of textual enhancement and t opic familiarity on Korean EFL rehension and learning of passive form Language Learning 57, 87 118. Lee, J. F. & Rodrguez R. (1997) The effect of leximic and morphosyntactic modifications on L2 reading comprehension and input processing. In T. Prez Leroux & W. R. Glass (Eds.), Co ntemporary perspectives on the acquisition of Spanish: Vol 2 Production, processing, and comprehension (pp. 135 137 ) Somerville, MA: Cascadilla Press. Lee, J. F., & VanPatten, B (1995) Making communicative language teaching happen New York: McGraw Hill Lee, J. F., & VanPatten, B (2003) Making communicative language teaching happen (2nd Edition) New York: McGraw Hill Lee, S., & Huang, H (2008) Visual input enhan cement and grammar learning: A meta analytic review Studies in second language a cquisition, 30, 307 331.

PAGE 341

327 Lehrer, R (1993) Authors of knowledge: Patterns of hypermedia design In S. LaJoie & S. Derry (Eds.), Computers as cognitive tools (pp 197 227) Hillsdale, NJ: Lawrence Earlbaum. Leow, R (1995) Modality and intake in second language acquisition Studies in second language acquisition 17, 79 90. Leow, R (1997) Attention, Awareness, and Foreign Language Behavior Language Learning 47, 3, 467 505. Leow, R (2000) A study of the role of awareness in foreign language behavi or Studies in Second Language Acquisition 22, 557 584. Leow, R (2001) Do learners notice enhanced forms while interacting with the L2 input? An online and offline study of the role of w ritten input enhancement in L2 reading Hispania 84, 496 509. L eow, R., & Hama, M (2008) The role of awareness in learning: Revisiting Williams ( 2005) Paper presented at the American Assoc iation for Applied Linguistics Annual Conference, Washington DC March 31, 2008. Leow, R., Nuevo, M. & Tsai, Y. (2003). The ro les of textual enhancement and type of Applied Language Learning, 13, 2, 1 16. Levin, J. R., Bender, B., G., & Lesgold, A. M (1976) Pi ctures, repetition, and young c rning Audio Visual Communication Review 24, 367 380. Lieberman, L. R. & Culpepper, J. T (1965) Words vers us objects: Comparison of free verbal recall. Psychological Reports 17, 983 988. LoCoco, V (1987) Learner comprehension of oral a nd written se ntences in German and Spanish: The importance of word order In B. VanPatten, T. R. Dvorak, & J. F. Lee (Eds.), Foreign Language Learning: A research perspective (pp. 119 129) Cambridge, MA: Newbury House. Long, M. (1985) Input and second language acqui sition theory In Gass, S. & Madden, C. (Eds.), Input in second language acquisition (pp. 377 393) Rowley, MA: Newbury House. Marsden, E (2006) Exploring input processing in the classroom: An experimental comparison of processing instruction and enrich ed input Language Learning, 56, 3, 507 566

PAGE 342

328 Matsunaga, S., & Crosby, M (1997) The relationship bet ween spatial ability of native speakers of Japanese and their coding strategy when reading Kanji, Computer Assisted Language Learning 10, 4, 321 337. Mayer, R. E (2005) Multimedia Learning: Guid ing visuospacial thinking with instructional animation In P. Shah, & A. Miyake (Eds.), The Cambridg e handbook of visuospatial thinking (pp 477 508.) New York, NY: Cambridge University Press. Mayer, R. E. & Anderson, R. B (1991) Animations ne ed narrations: An experimental test of dual coding hypothesis Journal of Educational Psychology 83, 484 490. McConnell, J. V. (1978) Confessions of a textbook writer American Psychologist 33, 159 169. McL aughlin, B. (1987). Theories of second language learning London: Edward Arnold. McLaughlin, B (1990) . TESOL Quarterly 24, 617 634. McLaughlin, B., Rossman, T., & McLeod, B. (1983) Second language learning : An information processing perspective. Language Learning, 33 2, 135 158. Mejias account of mood in Spanish. Hispania 77, 4, 892 902. Menne, J. M. & Menne, J. W (1972) T he relative efficiency of bimodal presentation as an aid to learning Audio Visual Communication Review, 20, 170 180. Montrul, S. A (2004) The acquisition of Spanish: Morphosyntactic development in monolingual and bilingual L1 acquisition and adult L2 a cquisition Philadel phia, PA: John Benjamins Publishing Company. Morgan Short, K., & Wood Bowden, H (2006) Process ing instruction and meaningful output based instruction: Effects on second language development Studies in Second Language Acquisition 2 8, 31 65. Nagata, N (1998) Input vs. output practice in educationa l software for second language acquisition Language Learning and Technology 1, 23 40.

PAGE 343

329 Najjar, L. J (1998) Principles of educational multimedia user interface design Human Factors 4 0, 311 323 Nealis, P. M., Harlow, H. F., & Suomi, S. J. (1977) Th e effects of stimulus movement on discrimination learning by rhesus monkeys Bulletin of Pscyhonomic Society 10, 161 164. Nelson, D. L., Reed, V. S., & Walling, J. R (1976) Pictorial su periority effect Journal of Experimental Psychology: Human Learning and Memory 2, 523 528. Nielsen J (1993) What is usability? In J. Nielsen (Ed.) Usability Engineering (pp. 23 48) New York: Morgan Kaufmann. Nielsen, J., & Landauer, T. K (1993) A mathematical mod el of the finding of usability problems, (pp. 206 213) Proceedings of ACM INTERCHI'93 Conference (Amsterdam, The Netherlands, 24 29 April). Nugent, G. C. (1982) Pictures, audio, and print: Symboli c representation and effect on learning Educational Communication and Technology Journal 30, 163 174. Nunnally J. (1978). Psychometric theory. New York: McGraw Hill. Overstreet, M (1998) Text enhancement and content fam iliarity: The focus of learner attention Spanish Applied Linguistics, 2, 229 258. Overstreet, M. (2002). The effects of textual enhancement on second language learner reading comprehension and form recognition. Unpublished doctoral dissertation, University of Illinois, Urbana Champaign. Omaggio Hadley, A (2001) Teachin g Language in Context (third edition) Boston: Heinle & Heinle. Paivio, A (1971) Imagery and verbal processes New York: Holt, Rinehart & Winston. Paivio, A. (1975). Coding distinctions and repetition effects in memory. In G. H. Bower (Ed.), The psycho logy of learning and motivation vol. 9 (pp. 179 214). New York: Academic Press. Paivio, A (1986) Mental representations: A dual coding approa ch New York: Oxford University Press. Paivio, A (1991) Dual coding theory : Retrospect and current status Canadian Journal of Experimental Psychology 45, 255 287.

PAGE 344

330 Paivio, A., & Csapo, K (1969) Concrete image and verbal memory codes Journal of Experimental Psychology 80, 279 285. Paivio, A., & Csapo, K (1973) Picture superiority i n free recall: Imager y or dual coding? Cognitive Psychology 5, 176 206. Park, O., & Hopkins, R. (1993). Instructional conditions for using dynamic displays: A review. Instructional Science 21, 427 449. Paulston, C. B (1972) Structural pattern drills In H.B. Allen & R. N. Cambell (E ds.), Teaching English as a second language (pp. 129 138) New York: McGraw Hill. Pereira I. (1996). Markedness and instructed SLA: An experiment in teaching the Spanish subjunctive Unpublished doctoral dissertation, University of Illin ois, Urbana Champaign. Peters, A. M (1985) Language segmentation: Operating principles for the perception and analysis of language In D.I. Slobin (Ed.), The cross linguistic study of language acquisition vol 2 (pp. 1029 1067) H illsdale, NJ: Lawrence Erlbaum Associates. Pezdek, K., Lehrer, A., & Simon, S (1984) The re lationship between reading and cognitive processing of television and radio Child Development 55, 2072 2082 Pienemann, M (1989) Is language teachable? Applied Linguistics 10, 52 79. Pienemann, M (1998) Language processing an d second language development: Processability theory Philadelphia: John Benjamins. Pienemann, M., & Johnston, M. (1987) Factors in fluencing the development of language Proficiency In Nunan D. (Ed.) App ly ing second language acquisition research (pp. 45 141) Adelaide, Australia : National Curriculum Resource Centre, AMEP Posner, M (1994) Attention in cognitive neuroscience: An overview In M. Gazzaniga (Ed.), The cognitive neurosciences (pp. 615 624) Cambridge, MA: MIT Press. Posner, M., & Peterson, E (1990) The attentional systems of the human brain Annual Review of Neuroscience 13, 25 42. Qin, J (2008) The effect of processing instruction and dict ogloss tasks on acquisition of the English pa ssive voice Language Teaching Research 12, 61 82.

PAGE 345

331 Richards, J. C., & Rodgers, T. S (1986) Appr oaches and methods in language teaching Cambridge: Cambridge University Press. Rieber, L. P (1990) Animation in computer based instruction Educational Technology Research and Development 38, 77 86. Robinson, P (1995) . Language Learning, 45, 283 331. Robinson, P (1996) Learning simple and compl ex second language rules under implicit, incidental, r ule search, and instructed conditions, studies in second language acquisition, 18, 27 67. Robinson, P (1997) Generalizability and automaticity of second language learning under implicit, incidental, enhanced, and instructed conditions Studies in Second Language Acquistion 19, 223 247. Rosa, E (1999) A cognitive approach to task based rese arch: Explicitness, awareness, and L2 development Unpublished doctoral disser tation, Georgetown University, Washington DC. Rosa, E., & Leow, R (2004) Awareness, different l earning conditions, and second language development, Applied Psycholinguistics 25, 269 292. Explicitness, intak e, and the issue of awareness: Another piece to the puzzle S t udies in Second Language Acquisiti on 21, 511 556. Rutherford, W. & Sharwood Smith, M (1985) Consc iousness raising and universal grammar Applied Linguistics 6, 274 282 Salaberry, M. R (1997) The role of input and output practice in se cond language acquisition Canadian Modern Lan guage Journal 53, 422 451. Sanz, C (2004) Computer delivered explicit versus e xplicit feedback in processing instruction In B. VanPatten (Ed.), Processing inst ruction: Theory, research, and commentary (pp. 241 256) Mahwah, NJ: Erlbaum S anz, C. & M organ Short, K (2004) Positive evidence vs. explicit rule presentation and explicit negative fee d back: A computer assisted study Language Learning 54, 35 78.

PAGE 346

332 Schachter, J (1998) Recent research in language learning studies: Promises and problems. La nguage Learning 48, 557 583. Schmidt, R (1990) The role of consciousness in second language learning Applied Linguistics 11, 206 226. Schmidt, R (1993) Awareness and second language acquisition Annual Review of Applied Linguistics 13, 206 226. Schmidt, R. (1995) Consciousness and foreign language le arning: A tutorial on the role of attention and awareness in learning In R Schmidt (ED.), Attention and awareness in foreign languge learning (pp 1 63) Honolulu : University of Schm idt, R. (2001) Attention In P. Robinson (Ed.), Cognition and Second Language Instruction (pp. 3 32) New York: Cambridge University Press. Schmidt, R. & Frota, S. (1986) D eveloping Basic Conversational Ability in a Second Language: A Case Study of an A dult Learner of Portuguese In R. R. Day (Ed.) Talking to Learn: Conversation in Second Language Acquisition (pp. 237 326) Rowley, MA: Newberry. Selinker, L (1972) Interlanguage International Review of Applied Linguistics 10, 209 231. Severin, W. J (1967) The effectiveness of relevan t pictures in multiple channel communications Audio Visual Communication Review 15, 386 401. Sewell, E. H., Jr., & Moore, R. L (1980) Cartoon embellishments in informative presentations Educational Communication a nd Technology Journal 28, 39 46. Sharwood Smith, M (1981) Consciousness raising and the second language learner Applied Linguistics 2, 159 168. Sharwood Smith, M (1986) Comprehension v ersus acquisition: Two ways of processing input Applied Lingui stics 7, 239 274. Sharwood Smith, M (1991) Speaking to many minds: On the relevance of different types of language information for the L2 learner Second Language Research 7, 118 132. Sharwood Smith, M (1993) Input enhancement in instructed SLA: T heoretical bases Studies in Second Language Acquisition 15, 165 179.

PAGE 347

333 Sherman, J. L. (1976) Contextual information and prose comprehension Journal of Reading Behavior 8, 369 379. Shook, D. (1994) FL/L2 reading, grammatical inform ation, and the input to intake phenomenon Applied Language Learning 5, 57 93. Skehan, R (1998) A cognitive approach to language learning. Oxford: Oxford University Press. Skinner, B (1957) Verbal behavior New York: Appleton Century Crofts. Schneider W., & Shiffrin, R.M (1977) Controlled and automatic human informa tion processing: Detection, search, and attention Psychological Bulletin, 84, 1 66. Spada, N., & Lightbown, P. M (1993) Instruction and t he development of questions in L2 classrooms Studies in Second Language Acquisition 15, 205 224. Stafford, J. Y (1990) Effects of active learning with co mputer assisted or interactive video instruction Unpublished doctoral dissert ation, Wayne State University, Detroit, MI. Stevens, J (2002). Applied multivari ate statistics for the social sciences (fourth edition). Mahwah, New Jersey, Lawrence Erlbaum Associates. Stevenson, H. W., & Siegel, A (1969) Effects of instr uction and age on retention of filmed content Journal of Educational Psychology 19, 56 61. Stockwell, R., Bowen, J. & Martin, J (1965) The gra mmatical structures of English and Spanish University of Chicago Press. Stoneman, Z., & Brody, G. H (1983) Immediate and long term recognition and generalization of advertised products as a functio n of age and presentation mode Developmental Psychology 19, 56 61. Sutcliffe, A., & Namoune, A (2007) Getting the mes sage across: Visual attention, aesthetic design, and what users remember Pr oceedings from 11th IFIP TC 13 International Conference R io de Janeiro, Brazil, September 10 14, 2007. Swain, M (1985) Communicative Competence: Some Roles of Comprehensible Input and Comprehensible Output in its Development In S. Gass, & C. Madden (Eds.), Input and Second Language Acquisition Rowley, MA: Newbury House.

PAGE 348

334 Swain, M (1993) . The Canadian Modern Language Review, 50 158 164. Swain, M. (1995). Three functions of output in second l anguage learning. In G. Cook & B. Seidlhofer (Eds.), Principle and practice in a pplied linguistics: Studies in honour of H.G. Widdowson (pp. 125 144). Oxford: Oxford University Press. Swain, M (1998) Focus on form through conscious ref lection. In C. Doughty, C. & J. Williams (Eds.), Focus on Form in Clas sroom Second Language Acquisition (pp. 85 113) Cambridge: Cambridge University Press Swain, M. & Lapkin, S (1995) Problems in output an d the cognitive processes they generate: A step towards second language learning Applied Linguistics 16 371 391. Takagaki, T (1984) Subjunctive as a marker of subordination Hispania 57, 238 256. Terrell, T (1991) The role of grammar instruction in a communicative approach The Modern Language Journal 75, 52 63 Terrell, T., Baycroft, B., and Perrone, C (1 98 7). The Subjunctive in Spanish interlanguage: Accuracy and comprehensibility. In Foreign language learning: A research perspective edited by Bill VanPatten T. Dvorak and J. Lee, 23 48. Cambridge, MA: Cambridge University Press. Terrell, T., & Hooper, J (1974) A semantically based analysis of mood in Spanish Hispania, 57, 484 494. Tomlin, R., & Villa, V (1994) Attention in cognitive science and second language Acquisition Studies in Second Language Acquisiton 16, 183 203. Truscott, J (1998) Noticing in second language acquisition: A critical review Second Language Research 14, 103 135. VanPatten, B (1984) pronouns: More evidence for a word order strategy Hispanic Linguistics 1, 57 67. VanPatten, B. (1 985) The acquisition of ser and estar in adult second language learners: A preliminary investigation of transitional stages of competence. Hispania 68, 19 23.

PAGE 349

335 VanPatten, B. (1987) Classroom ser and estar : Accounting for de velopmental patterns. In B. VanPatten T. Dvorak, & J. F. Lee (Ed s .), Foreign language learning: A research perspective (pp. 61 75). Rowley, MA: Newbury House. VanPatten, B (1990) Attending to form and content in the input: An experim ent in conscio usness Studies in Second Language Acquisition 12, 287 301. VanPatten, B (1993) Grammar teaching for the acquisition rich classroom Foreign Language Annals 26, 435 450. VanPatten, B. (1994) Evaluating the role of consciousness in the second languag e Acquisition: Terms, linguistic features, and research methodology AILA Review 11, 27 36. VanPatten, B (1995) Cognitive aspects of input processing in second language acquisition In P. Hashemipour, R. Maldonado, & M. van Naerssen (Eds.), Studies in second language learning and Spanish li nguistics in honor of Tracy D. Terrell (pp. 170 183) New York: McGraw Hill VanPatten, B (1996) Input processing and grammar instruction: Theory and Research Norwood, NJ: Ablex. VanPatten, B (2002) Processing instruction: An update Language Learning 52, 755 803. VanPatten, B (2003) From input to output: A teac acquisition New York: McGraw Hill. VanPatten, B (2004) Input processing in SLA In B. VanPatten (Ed.), Processing Instruction: Theory, Research, and Commentary (pp. 5 31) Mahwah, NJ: Lawrence Erlbaum Associates. VanPatten, B., & Cadierno, T. (1993a) Explicit instruction and input processing Studies in Second Language Acquisition ,15, 225 243. VanPatten, B., & Cad ierno, T (1993b) Input processing and second language acquisition: A role for instruction Modern Language Journal 77, 45 57. VanPatten, B. & Fernn dez, C (2004) The long term effects of processing instruction In B. VanPatten (Ed.), Processing Inst ruction: Theory, Research, and Commentary (pp. 273 290) Mahwah, NJ: Lawrence Erlbaum Associates.

PAGE 350

336 VanPatten, B., & Oikkenon, S (1996) Exlpanation vs. structured input in pro cessing instruction Studies in Second Language Acquisition 18, 495 510. VanPa tten, B., & Sanz, C (1995) From input to outp ut: Processing instruction and communicative tasks In F. Eckman, D. High land, P. Lee, J. Mileham, & R. Rutkowski (Eds.), Second Language Acquisition: Theory and pedagogy (pp. 169 185) Mahwah, NJ: Erlbaum. V anPatten, B., & Wong, W (2004) Processing instructi on and the French causative: A replication In B. VanPatten (Ed.), Processing Instruction: Theory, Research, and Commentary (pp. 9 117) Mahwah, NJ: Lawrence Erlbaum Associates. Verano, M (1987) Achiv ement and retention of Span ish presented via videodisc in linear, segmented and interactive modes Unpu blished doctoral dissertation, University of Texas, Austin. Wetstone, H. S., & Friedlander, B. Z (1974) The effec t of live, TV, and audio story narrat Journal of Educational Research 68, 32 35. White, C (1999) Expectations and emergent beliefs of self i nstructed language learners In Metacognitive knowledge and beliefs in language learning Sys tem Special Issue 27, 443 457. White, C (2003) Language Learning in Distance Education Cambridge: Cambridge University Press. White, J. (1998) A t ypographical input enhancement study In Doughty, C. & Williams, J. (E ds.), Focus on Form in Classroom Second Language Acquisition (pp. 85 113) C ambridge: Cambridge University Press. White, L (1991) Adverb placement in second languag e acquisition: Some effects of positive and negative evidence in the classroom Second L anguage Research, 7, 133 161. Williams, J (1999) Memory, attention, and inductive learning Studies in Second Language Acquisition, 21, 1 48. Williams, J (2005) Learning without awareness Studies in Second Language Acquisition 27, 2, 269 304.

PAGE 351

337 Wong W (2002) Decreasing attentional demands in input processing: A textual enhancement study Paper presented at the annual meeting of the Second Language Research Forum (SLRF), Toronto, Canada. October 3 6, 2002. Wong, W (2003) Textual enhancement and s implified input: Effects on L2 comprehension and acquisition of non meaningful grammatical form Applied Language Learning 13, 2, 17 45. Wong, W (2004) Processing instruction in French: The roles of explicit information and structured input In B. Van Patten (Ed.), Processing instruction: T heory, research, and commentary (pp. 187 205) Mahwah, NJ: Erlbaum. Wong, W (2005) Input enhancement from theory and research to the classroom Boston: McGraw Hill Higher Education. Wong Fillmore, L (1976) The second time around Unpublished doctor al dissertation, University of California, Berkeley. Wolf, D. A (1993) Issues in reading comprehension as sessment: Implications for the development of research instruments and classroom tests Foreign Language Annal s 26, 3, 322 331. Wolfe Quintero, K (1992) Learnability and the acquisition of extraction in relative clauses and wh questions Studies in Second Language Acquisition 14, 39 71. Woodson, K. (1997). Learner centered input processing: Bridging the gap between foreign language teachers and SLA researchers Unpublished doctoral dissertation, Georgetown University.

PAGE 352

338 Appendix A Informed Consent Form

PAGE 353

339 Informed Consent to Participate in Research Information to Consider Before Taking Part in this Resea rch Study IRB Study # 107737 Researchers at the University of South Florida (USF) study many topics. To do this, we need the help of people who agree to take part in a research study. This form tells you about this research study. We are asking you t o take part in a research study that is called: The effects of processing instruction, structured input, and visual input enhancement on the acquisition of the subjunctive in adjectival clauses among intermediate level distance learners of Spanish The pe rson who is in charge of this research study is Victoria Russell. This person is called the Principal Investigator. However, other research staff may be involved and can act on behalf of the person in charge. The research will be done at the University of South Florida, College of Arts and Science, Hillsborough Community College, Brandon Campus, and at your homes as you work online. Purpose of the study The purpose of this study is to find out more about which methods are the most effective for we b based instruction of complex Spanish grammar Study Procedures If you take part in this study, you will be asked to Provide some background information on your native language and the languages that you have studied in the past, your age, and your gender (this information will be kept confidential). You will also be asked about your feelings regarding learning a language online.

PAGE 354

340 Take a pre test, a posttest, a delayed posttest and a reading comprehension test. Each test will take approximately 15 minutes of your time. Complete an instructional activity package online that replaces the assignments that you would normally do for your class to learn the grammatical form (workbook, lab manual, and online act ivities). The instructional activity package contains ten activities related to a grammatical form that you will need to learn for your course. It should take no more time than you would normally spend to complete course related activities when learning a new grammatical form or structure (about 2 hours). Complete an online reading activity while you take notes in a text box. This activity will take approximately 15 minutes. Complete a posttreatment questionnaire asking about your experiences while you w orked online. This activity will take approximately 15 minutes. You will fill out the pretreatment questionnaire here today. All other materials will be delivered online. You may begin your activity package any day next week after you complete the prete st today or tomorrow, and you may complete the activity package any time of the day or night that suits your schedule. You will take the posttest the same day that you complete your activity package. The reading activity, reading comprehension test, and posttreatment questionnaire will be completed online one to three days after you submit your activity package. Finally, you will be asked to take a delayed posttest two weeks after you submit your online activity package. This test will take no more than 15 minutes of your time. Alternatives You have the alternative to choose not to participate in this research study. Participants in the research study will receive bonus points added to their final average. If you choose not to participate in the rese arch study, but you would like to earn the bonus points, you have the option of completing an alternative assignment. The alternative assignment consists of completing a package of worksheets, and completing a reading and writing assignment in Spanish. T he alternative assignment will take approximately 3 hours of your time. Benefits The potential benefits to you are: It will help you to learn a particularly difficult aspect of Spanish grammar. In addition, participation in the study should help inc rease your performance on your final exam, as

PAGE 355

341 vocabulary and grammatical structures that are tested on your exam will be practiced in the study related materials. Risks or Discomfort This research is considered to be minimal risk. That means that the risk s associated with this study are the same as what you face every day. There are no known additional risks to those who take part in this study. Compensation I will not pay you for the time you volunteer while being in this study Confidentiality We must keep your study records as confidential as possible. To ensure that your records are kept confidential, your background questionnaires and informed consent forms will be stored in a locked filing cabinet for five years. After that time, they wil l be destroyed. After you complete the study related materials, they will be erased from Blackboard and your scores will be stored anonymously using an ID number rather than your name. However, certain people may need to see your study records. By law, anyone who looks at your records must keep them completely confidential. The only people who will be allowed to see these records are: The research team, including the Principal Investigator, study coordinator, and all other research staff. Certain gove rnment and university people who need to know more about the study. For example, individuals who provide oversight on this study may need to look at your records. This is done to make sure that we are doing the study in the right way. They also need to m ake sure that we are protecting your rights and your safety.) These include: o The University of South Florida Institutional Review Board (IRB) and the staff that work for the IRB. Other individuals who work for USF that provide other kinds of oversight ma y also need to look at your records. o The Department of Health and Human Services (DHHS). We may publish what we learn from this study. If we do, we will not let anyone know your name. We will not publish anything else that would let people know who yo u are.

PAGE 356

342 Voluntary Participation / Withdrawal You should only take part in this study if you want to volunteer. You should not feel that there is any pressure to take part in the study, to please the investigator or the research staff. You are free to p articipate in this research or withdraw at any time. There will be no penalty or loss of benefits you are entitled to receive if you stop taking part in this study. Your decision to participate or not to participate will not affect your student status, course grade, or job status. Questions, concerns, or complaints If you have any questions, concerns or complaints about this study, call Victoria Russell at 813 810 9885 If you have questions about your rights as a participant in this study, general q uestions, or have complaints, concerns or issues you want to discuss with someone outside the research, call the Division of Research Integrity and Compliance of the University of South Florida at (813) 974 9343. If you experience an unanticipated problem related to the research call Victoria Russell at 813 810 9885. Consent to Take Part in this Research Study It is up to you to decide whether you want to take part in this study. If you want to take part, please sign the form, if the following statements are true. I freely give my consent to take part in this study. I understand that by signing this form I am agreeing to take part in research. I have received a copy of this form to take with me. _____________________________________________ ___________ Signature of Person Taking Part in Study Date _____________________________________________ Printed Name of Person Taking Part in Study

PAGE 357

343 Statement of Person Obtaining Informed Consent I have carefully explained to the person taking part in the study wh at he or she can expect. I hereby certify that when this person signs this form, to the best of my knowledge, he or she understands: What the study is about. What procedures/interventions/investigational drugs or devices will be used. What the potential b enefits might be. What the known risks might be. ____________ Signature of Person Obtaining Informed Consent Date Printed Name of Person Obtaining Informed Consent

PAGE 358

344 Appendix B Pretreatment Questionnaire

PAGE 359

345 1. Name (fi rst and last): ___________________________________________________ 2. Age: __________ 3. Please check: Male____ Female____ 4. What language(s) did you grow up speaking? ________________________________ 5. What language is spoken in your home? ______ ______________________________ 6. Do you speak another language at least half of the time besides English? If so, which language? __________________________________________________ 7. How many semesters have you studied Spanish in college? ________ _____________ 8. Did you take Spanish in high school? If so, for how many years? ________________ ________________________________________________________________________ 9. Do you have daily or weekly contact with the Spanish language outside of class? If so, please explain. ________________________________________________________ 10. Why did you choose to learn a language online? Circle the response that best applies to you: a. Convenience b. I enjoy using computers and the Internet c. Other ____________________ 11. How would you rate your computer skills? a. Highly proficient b. Fairly proficient c. Not very proficient d. Using a computer is difficult for me 12. How easy is it to use Blackboard and Quia to acc ess and complete your course materials? a. Very Easy b. Easy c. Somewhat easy d. Difficult e. Very Difficult 13. Would you take another language course online ? a. Yes b. No c Maybe

PAGE 360

346 Appendix C Form s A, B, and C of the Subjunctive Knowledge Test

PAGE 361

347 Form A

PAGE 362

348

PAGE 363

349

PAGE 364

350

PAGE 365

351

PAGE 366

352

PAGE 367

353

PAGE 368

354

PAGE 369

355

PAGE 370

356 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 371

357 Form B

PAGE 372

358

PAGE 373

359

PAGE 374

360

PAGE 375

361

PAGE 376

362

PAGE 377

363

PAGE 378

364

PAGE 379

365

PAGE 380

366 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 381

367 Form C

PAGE 382

368

PAGE 383

369

PAGE 384

370

PAGE 385

371

PAGE 386

372

PAGE 387

373

PAGE 388

374

PAGE 389

375

PAGE 390

376 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 391

377 Appendix D Comprehension Test

PAGE 392

378

PAGE 393

379

PAGE 394

380

PAGE 395

381 Appendix E Note Sheet

PAGE 396

382 Read the following want ads that were taken off of two Spanish language websites on the Internet. As you read the ads please note in the text box all of the words (such as vocabulary items and verb forms) that are necessary for you to comprehend the text. D o not type in every single word, only the w ords that help you understand the meaning of the ad. Want Ad #1: BUSCO UNA CASA QUE TENGA UN JARDIN AMPLIO CERRADO Y POR LO MENOS 2 HABITACIONES, 2 BAOS EN TOLUCA O ALREDEDORES. QUE ACEPTEN MASCOTAS ES IMPRESCINDIBLE PARA M. ¡ME URGE! *The five want ads and note sheets (text boxes) were delivered online one at a time. Participants were not permitted to back track to a previous want ad once they filled in the text box.

PAGE 397

383 Appendix F Authentic Input Text

PAGE 398

384 Reading Activity Read the follo wing want ads that were taken off of two websites on the Internet. As you read each ad, please note all of the words (such as vocabulary items and verbs) in the text box that you focus on to understand the Spanish want ad. Please do not write down every single word. 1. BUSCO UNA CASA QUE TENGA UN JARDIN AMPLIO CERRADO Y POR LO MENOS 2 HABITACIONES, 2 BAOS EN TOLUCA O ALREDEDORES. QUE ACEPTEN MASCOTAS ES IMPRESCINDIBLE PARA M. ¡ME URGE! 2. Busco una casa que no est alejado de la ciudad, sin problem as de agua, recoleccin de basura, sin vecinos problemticos, casa o departamento que est en buen estado, que tenga lnea telefnica y enrejada. 3. Busco una casa que tenga 3 dormitorios y es muy urgente porque tengo que irme de donde vivo por problema de trabajo y el colegio. Lo nico que pido que sea tranquilo y los vecinos sean buenas personas y est un colegio cerca de la casa. 4. Busco un apartamento que me alquilen para pareja sin nios Busco que me alquilen un apartamento para pareja sin nios con cochera, entrada independiente, que tenga 1 o 2 habitaciones, en Guadalupe o alrededores. pago mximo 100.000/mes 5. BUSCO A ALGUIEN QUE TENGA APARTAMENTO Y DESEE COMPARTIRLO. NECESARIO QUE TENGA BAO PRIVADO PARA M, COCINA, COMEDOR, SALA Y SI ES POSIBLE 2 CUARTOS Y UN PATIOCITO. EL PRECIO MS O MENOS TENDRA QUE SER ENTRE 20.000 HASTA 50.000 *The ads in the Authentic Input Passage were taken from two web sites that post f ree classified ads in Spanish. They can be found at: www.MundoAnuncio.com and www.adoos.com.mx All ads were retrieved from the web on December 10, 2008.

PAGE 399

385 Appendix G Posttreatment Questionnaire

PAGE 400

386

PAGE 401

387

PAGE 402

388

PAGE 403

389

PAGE 404

390 Appendix H Satisfaction Survey

PAGE 405

391 Please indic ate the degree to which you agree or disagree with the following statements. Mark 1 for statements with which you strongly agree and mark 5 for statements with which you strongly disagree. 1. The direction s were clear and easy to follow. 5 4 3 2 1 2. I learned something from completing this activity package. 5 4 3 2 1 3. I preferred these types of activities to my regular classroom activities. 5 4 3 2 1 4. It was easy t o complete the web based grammar activities. 5 4 3 2 1 5. I enjoyed learning Spanish grammar using the materials. 5 4 3 2 1 Comments: ______________________________________________________________ ___________________________________ _____________________________________ ________________________________________________________________________

PAGE 406

392 Appendix I Vocabulary Practice Activity

PAGE 407

393

PAGE 408

394

PAGE 409

395

PAGE 410

396 Appendix J Traditional Instruction Explicit Grammar Explanation: Subjunctive Formation

PAGE 411

397 Formation of the Subjunctive Recall the formation of the present subjunctive for regular verbs : yo present indicative : Estudiar Estudio Comer Como Escribir Escribo Th en add the opposite ending : For ar verbs, add e, es, e, emos is, en For er and ir verbs, add a, as, a, amos, is, an Estudi o Estudi e Estudi es Estudi e Estudi emos Estudi is Estudi en Com o Com a Com as Com a Com amos Com is C om an Escrib o Escrib a Escrib a s, Escriba Escrib amos Escrib is Escrib an subjunctive: Tener Tengo Teng o Teng a Teng as Teng a Teng amos Teng is Teng an There are only five irregular subjunctive verbs ; here are the 3 rd person present subjunctive forms: Dar Estar Ir Saber Ser Dar : D, Des, D, Demos, Deis, Den Estar : Est, Ests, Est, Estemos, Estis, Esten Ir: Vaya, Vayas, Vaya Vayamos, Vayis, Vayan Saber : Sepa, Sepas, Sepa, Sepamos, Sepis, Sepan Ser: Sea, Seas, Sea, Seamos, Seis, Sean

PAGE 412

398 Some verbs take a spelling change in the subjunctive. Verbs that end in gar, zar, and car take the following changes: gar verbs pagar pa gu e buscar bus qu e empezar empie c e

PAGE 413

399 Appendix K Traditional Instruction Explicit Grammar E xplanation : Subjunctive Use

PAGE 414

400 To form the present subjunctive for regular verbs : yo present indicative : Estudiar Estudio Escribir Escribo Then add the opposite ending : For ar verbs, add e, es, e, emos is, en For er and ir verbs, add a, as, a, amos, is, an Estudi o Estudi e Estudi e s, Estudi e Estudi emos Estudi is Estudi en Escrib o Escrib a Escrib a s, Escriba Escrib amos Escrib is Escrib an e are regular in the present subjunctive: Tener Tengo Teng o Teng a Teng as Teng a Teng amos Teng is Teng an There are only five irregular subjunctive verbs ; here are the 3 rd person present subjunctive forms: Dar Estar Ir Saber Ser Dar : D, Des, D, Demos, Deis, Den Estar : Est, Ests, Est, Estemos, Estis, Esten Ir: Vaya, Vayas, Vaya, Vayamos, Vayis, Vayan Saber : Sepa, Sepas, Sepa, Sepamos, Sepis, Sepan Ser: Sea, Seas, Sea, Seamos, Seis, Sean In Leccin 13, you learne d that the subjunctive is used in adverbial clauses after certain conjunctions. You will now learn how the subjunctive can be used in adjective clauses to express that the existence of someone or something is uncertain or indefinite. The subjunctive is use d in an adjective (or subordinate) clause that refers to a person, place, thing, or idea that either does not exist or whose existence is uncertain or

PAGE 415

401 indefinite. In the examples below, compare the differences in meaning between the statements using the in dicative and those using the subjunctive. Necesito el libro que tiene informacin sobre Venezuela. I need the book that has information about Venezuela. Quiero vivir en esta casa que tiene jardn. I want to live in this house that has a garden. En mi barrio, hay una heladera que vende helado de mango. Necesito un libro que tenga informacin sobre Venezuela. I need a book that has information about Venezuela. Quiero vivir en una casa que tenga jardn. I want to live in a house that has a garden. En mi barrio no hay ninguna heladera que venda helado de mango. In my neighborhood, there are no ice cream stores that sell mango ice cream En mi b arrio, hay una heladera que vende helado de mango. When the adjective clause refers to a person, place, thing, or idea that is clearly known, certain, or definite, the indicative is used. Quiero ir al supermercado que vende productos venezolanos. I want to go to the supermarket that sells Venezuelan products. Conozco a alguien que va a esa peluquera. I know someone who goes to that beauty salon. Busco al profesor que ensea j apons. Tengo un amigo que vive cerca de mi casa. I have a friend who lives near my house

PAGE 416

402 The personal a is not used with direct objects that are hypothetical people. However, as you learned in Leccin 7, alguien and nadie are always preceded by the personal a when they function as direct objects. Necesitamos un empleado que sepa usar computadoras. We need an employee who knows how to use computers. Necesitamos al empleado que sabe usar computadoras. W e need the employee who knows how to use computers. Buscamos a alguien que pueda cocinar. No conocemos a nadie que pueda cocinar. can cook. The subjunctive is commonly used in questions with a djective clauses when the speaker is trying to find out information about which he or she is uncertain. However, if the person who responds to the question knows the information, the indicative is used. Hay un parque que est cerca de nuestro hotel? Is S, hay un parque que est muy cerca del hotel. ¡Atencin! Here are some verbs that are commonly followed by adjective clauses in the subjunctive buscar (no) conocer (no) haber necesitar querer Adjective clauses are subordinate clauses that modify a noun or pronoun in the main clause of a sentence. That noun or pronoun is called the antecedent. En Lnea 2.0 Spanish Language Course

PAGE 417

403 Appendix L Traditional Instruction Treatment Package

PAGE 418

404

PAGE 419

405

PAGE 420

406

PAGE 421

407

PAGE 422

408

PAGE 423

409

PAGE 424

410

PAGE 425

411

PAGE 426

412

PAGE 427

413

PAGE 428

414

PAGE 429

415

PAGE 430

416

PAGE 431

417

PAGE 432

418

PAGE 433

419

PAGE 434

420

PAGE 435

4 21 Appendix M Processing Instruction Explicit Grammar Explanation : Subjunctive Formation

PAGE 436

422 To form the present subjunctive for regular verbs: yo present indicative : Estudiar Estudio Escribir Escribo Beber Bebo Then add the opposite ending : For ar verbs add the following endings for the third person: e, en Estudi o 3 rd person singular (l, ella, usted) Estudi e Estudi o 3 rd person plural (ellos, ellas, ustedes) Estudi en For er and ir verbs add the following endings for the third person: a, an Escrib o 3 rd person singula r (l, ella, usted) Escrib a Escrib o 3 rd person plural (ellos, ellas, ustedes) Escrib an Beb o 3 rd person singular (l, ella, usted) Beb a Bebo 3 rd person plural (ellos, ellas, ustedes) Beb an m of the present tense are regular in the present subjunctive: Tener Tengo Teng o (3 rd person singular) Teng a (3rd person plural) Teng an Poder Puedo Puedo (3 rd person singular) Pued a (3rd person plural) Pued an There are only a few irr egular subjunctive verbs ; here are the 3 rd person singular and plural present subjunctive forms for the irregular verbs: Dar (D / Den) Saber (Sepa / Sepan) Estar (Est / Estn) Ser (Sea / Sean) Ir (Vaya / Vayan)

PAGE 437

423 Appen dix N Processing Instruction Explicit Grammar Explanation : Subjunctive Use

PAGE 438

424 WHAT DOES THE SUBJUNCTIVE LOOK LIKE? The following are some ar verbs conjugated in the third person singular of the present subjunctive. Notice that they take er endings : est udiar estudi e hablar habl e cantar cant e tocar toqu e The following are some er and ir verbs conjugated in the third person singular of the present subjunctive. Notice that they take ar endings : beber beb a tener teng a vivir viv a escribir escrib a Only a few verbs are irregular in the present subjunctive: Dar D Estar Est Ir Vay a Saber Sep a Ser Se a LOCATION OF THE SUBJUNCTIVE The subjunctive occurs in subordinate clauses. A subordinate clause must be preceded by a main clause. Subordinate clauses are generally introduced by the word que in Spanish. Example: Busco una persona que pueda trabjar los fines de semana. Busco una persona is the main clause in the previous sentence que pueda traba jar los fines de semana is the subordinate clause

PAGE 439

425 WHEN IS IT USED? 1) The subjunctive is used in adjectival clauses when the referent is unknown, uncertain, or hypothetical. servir = to serve Example 1: Busco un restaurante que sirva comida francesa en la ciudad. In the previous example, the present subjunctive is used in the subordinate clause because the speaker of the sentence is referring to a restaurant that is unknown, unc ertain, or hypothetical. Although the speaker is looking for a restaurant that serves French cuisine, he or she is unsure if such a restaurant exists in the city. Example 2: Busco un restaurante que sirve comida francesa en la ciudad. r a restaurant that serves French cuisine in the city. In example 2, the present indicative (present tense) is used in the subordinate clause because the speaker of the sentence is referring to a restaurant that is certain or known The speaker knows tha t there are restaurants in the city that serve French cuisine, but one of them has to be found. 2) The subjunctive is commonly used in questions with adjective clauses when the speaker is trying to find out information about which he or she is uncertain However, if the person who responds to the question knows the information, the indicative is used. Example 1: Hay un parque que est cerca de nuestro hotel? S, hay un parque que est muy cerca del hotel. If the person who responds answers with a negative expression such as ningn then the subjunctive is used in Spanish to express an element of uncertainty. Example 2: Hay un parque que est cerca de nu estro hotel? No, no hay ningn parque que est cerca del hotel

PAGE 440

426 BE CAREFUL! It is important to pay attention to verb endings in order to detect this difference in meaning in Spanish. Adjectival clauses that contain the subjunctive refer to people, places, and/or things that are uncertain, hypothetical, or unknown, and adjectival clauses that contain the indicative refer to people, places, and/or thin gs that are certain or known

PAGE 441

427 Appendix O Information on Processing Strategies

PAGE 442

428 Processing Strategies: How to Help Yourself Learn Spanish The subjunctive is a particularly difficult aspect of Spanish grammar to master for na tive speakers of English. One of the difficulties is that learners often have a hard time noticing the subjunctive forms in sentences because the present subjunctive endings are so similar to the present indicative endings: For example: the 3 rd person singular (l, ella, usted) form of the verb hablar in the present indicative is habl a and the 3 rd person singular (l, ella, usted) form of the verb hablar in the present subjunctive is habl e The subtle difference of the vowel switch from a to e is very difficult to notice when it occurs in sentences that contain other information. Spanish language learners, like yourself, tend to focus on vocabulary words rather than verb endings in order to comprehend the meaning of sentences. Sometimes however, it is important to focus on verb endings because a simple vowel shift can change the entire meaning of the sentence in Spanish. Take a look at the following two sentences: 1. Busco a un hombre que VENDE bocadillos en la calle (and) 1. Busco un hombre que VENDA bocadillos en la calle. When you read the two previous sentences, which words did you focus to help you understand them? Did you focus on the words hombre bocadillos and calle ? If so, you are not alone, as most Spanish lang uage learners will focus on content words such as these to extract meaning from sentences. However, the second verb in both of these sentences is the key to understanding the communicative intent of the speaker in Spanish. In the first sentence, the seco nd verb, vende, is in the present indicative, which indicates that the speaker of the sentence is referring to something or someone that is known In other words, the speaker of the sentence knows of a man who sells sandwiches in the street. In the secon d sentence, the second verb, venda, is in the subjunctive, which indicates that the speaker of the sentence is referring to something or someone that is unknown or hypothetical In other words, the speaker of the sentence does not know of a man that sells sandwiches in the street, although he is looking for such a man.

PAGE 443

429 the following sentence in English: In English, there is no way to tell if the speaker of the sentence is referring to a known or to an unknown referent. In other words, in English it is not clear if the speaker knows of a street vendor that sells sandwiches or not. In this instance, the Spanish lang uage is more precise than English! One thing you can do to help yourself understand Spanish sentences better is to closely examine the verbs in sentences. Try to pay attention to the ending of verbs to determine whether they are conjugated in the p resent subjunctive or in the present indicative. Another reason that the subjunctive is difficult to notice is that it usually occurs in the middle of sentences Information that occurs at the beginning of a sentence gets noticed and processed first, and interestingly, information that occurs at the end of sentences gets processed second. However, our brains process the information that occurs in the middle of sentences last! Take a look at the following sentenc e: Busco a alguien que hable espaol y chino. The subjunctive almost always occurs in the middle of sentences in Spanish. In the previous example, the second verb, hable is in the subjunctive. This verb indicates that the speaker of the sentence does not know of anyone who can speak both Spanish and Chinese. In other words, the speaker of the sentence is referring to someone who is unknown or hypothetical. Now that you are aware of this, you can try to pay more attention to verbs and verb endin gs that occur in the middle of sentences This strategy will help you notice the subjunctive more easily, which will enable you to interpret the meaning of Spanish sentences correctly.

PAGE 444

430 Appendix P Processing Instruction Treatment Package

PAGE 445

431

PAGE 446

432

PAGE 447

433

PAGE 448

434

PAGE 449

435

PAGE 450

436

PAGE 451

437

PAGE 452

438

PAGE 453

439

PAGE 454

440

PAGE 455

441

PAGE 456

442

PAGE 457

443

PAGE 458

444

PAGE 459

445

PAGE 460

446

PAGE 461

447

PAGE 462

448

PAGE 463

449

PAGE 464

450

PAGE 465

451 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 466

452 Appendix Q Structured Input Treatment Package

PAGE 467

453

PAGE 468

454

PAGE 469

455

PAGE 470

456

PAGE 471

457

PAGE 472

458

PAGE 473

459

PAGE 474

460

PAGE 475

461

PAGE 476

462

PAGE 477

463

PAGE 478

464

PAGE 479

465

PAGE 480

466

PAGE 481

467

PAGE 482

468

PAGE 483

469

PAGE 484

470

PAGE 485

471

PAGE 486

472 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 487

473 Appendix R Example of Computerized Visual Input Enhancement

PAGE 488

474

PAGE 489

475 Appendix S Processing Instruction w ith Visual Input Enhanceme nt Treatment Package

PAGE 490

476

PAGE 491

477

PAGE 492

478

PAGE 493

479

PAGE 494

480

PAGE 495

481

PAGE 496

482

PAGE 497

483

PAGE 498

484

PAGE 499

485

PAGE 500

486

PAGE 501

487

PAGE 502

488

PAGE 503

489

PAGE 504

490

PAGE 505

491

PAGE 506

492

PAGE 507

493

PAGE 508

494

PAGE 509

495

PAGE 510

496

PAGE 511

497

PAGE 512

498 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 513

499 Appendix T Structured Input with Visual Input Enhancement Treat ment Package

PAGE 514

500

PAGE 515

501

PAGE 516

502

PAGE 517

503

PAGE 518

504

PAGE 519

505

PAGE 520

506

PAGE 521

507

PAGE 522

508

PAGE 523

509

PAGE 524

510

PAGE 525

511

PAGE 526

512

PAGE 527

513

PAGE 528

514

PAGE 529

515

PAGE 530

516

PAGE 531

517

PAGE 532

518

PAGE 533

519 Note Structured input activities were adapted from Farley (2002) with permission.

PAGE 534

520 Appendix U Listening Scripts for the Instructional Treatments and the Subjuncti ve Knowledge Test: Forms A, B, & C

PAGE 535

521 Listening Script for the Traditional Instruction Grammar Activity Package : Activity 4: Listening You will hear five statements in Spanish. You will rewrite the sentences that you hear to make them negative. Use t he subjunctive where appropriate. You may listen to each recording as many times as necessary to answer each question. MODELO: (You hear) Hay un restaurante que sirve comida francesa ANSWER: (You type) No hay ningn restaurante que sirva comida fra ncesa. 4.1. Hay un restaurante que sirve comida japonesa 4.2. Hay una mujer que habla cuatro idiomas 4.3. Hay un hombre sabe programar computadoras. 4.4. Hay una tienda que vende tarjetas postales. 4.5. Hay una mujer que prepara bocadillos cuba nos. Activity 6b: Listening (Preguntas y Respuestas) After listening to the question, fill in the blank with the correct verb form (subjunctive or indicative). You may listen to the sound files as many times as necessary. 6.6. Conoces a alguien que viaje mucho a Venezuela? 6.7. Hay un banco que abra a las seis de la maana en la ciudad? 6.8. Conoces a alguien que hable cinco idiomas? 6.9. Tienes algn amigo que toque el piano? 6.10. Conoces a alguien que sabe programmar computadoras? N.B Students in the traditional instruction group also had an activity with five open ended speaking items. The students in the processing instruction and structured input groups had five additional listening items; however, they did not have any speaking activities.

PAGE 536

522 Listening Script for the Processing Instruction and Structured Input Grammar Packages with and without Computerized Visual Input Enhancement Activity 2: Looking for People and Places You will hear 10 sentences in Spanish. Listen carefully to determine if the speaker of each sentence is referring to something or someone that she knows, or if she is referring to something or someone of whose existence is unknown or uncertain to her. You may play the audio files as many times as necessary to answer each question. MODELO: (You hear) Busco a alguien que quiera viajar al extranjero cade mes. A. The person exists and is known to the speaker. B. The speaker does not know if such a person exists. Correct Answer: B 2.1. Busco una mujer que sepa hacer paella. 2.2. Quiero encontrar a un hombre que sabe reparar computadoras 2.3. Hay un restaurante por aqu que sirve comida francesa. 2.4. Hay un restaurante por aqu que sirve comida francesa. 2.5. Necesito encontrar a al guien que sabe hablar espaol. 2.6. Quiero encontrar una tienda que venda trajes de bao. 2.7. Busco un banco que cambia dinero. 2.8. Quiero encontrar un mercado que vende fruta fresca. 2.9. Busco una peluquera que no cargue tanto dinero. 2.10 Busco a alg uien que hable tres idiomas. Activity 5: Looking for People and Places Part 2 If the sentence you hear refers to a person, place, or thing that clearly exists or is known, not exist or times as necessary to answer the question. MODELO: (You hear) Hay un restaurante aqu que sirve comida venezolana. A. S B. No

PAGE 537

523 Correct Answer: S 5.1. Buscamos a alguien que hable italiano y francs. 5.2. Busco un restaurante que sirva comida japonesa. 5.3. Buscamos a un hombre que sabe programar computadoras. 5 .4. Buscamos una tienda que vende tarjetas postales. 5.5. Busco a una mujer que prepara bocadillos cubanos. Listening Scripts for the Subjunctive Knowledge Test (Forms A, B, & C) Students were only permitted to listen to each sound file twice. Test A. Listening Script 1. Busco una mujer que vende bocadillos en la calle. 2. Quiero ir a un restaurante que sirva comida francesa. 3. Quiero comprar una casa que tenga dos pisos. 4. Busco a alguien que pueda reparar mi computadora. 5. Neces ito un coche que no use mucha gasolina. 6. Quiero ir al restaurante que est en la esquina. 7. Necesito un empleado que hable italiano. 8. Busco un apartamento que est en el centro. 9. Busco a un hombre que trabaja con computadoras. 10. Busco una peluquera que no cargue tanto dinero. Test B. Listening Script 1. Busco a persona que sabe tocar el piano. 2. Quiero encontrar una tienda que venda gafas del sol. 3. Busco a alguien que quiera compartir el apartamento. 4. Vivo en una casa que tiene tres baos. 5. Busco una persona que ensee ingls y francs. 6. Quiero encontrar a alguien que diga la verdad. 7. Necesito encontrar un trabajo que ofrezca beneficios. 8. Tengo un amigo que es simptico y gracioso. 9 Quiero un jefe que sea inteligente y justo. 10. Necesito un empleado que haga buenas decisiones.

PAGE 538

524 Test C. Listening Script 1. Busco una mujer que vende bocadillos en la calle. 2. Quiero ir a un restaurante que sirva comida francesa. 3. Quiero comprar una casa que tenga dos pisos. 4. Busco a alguien que pueda reparar mi computadora. 5. Necesito un coche que no use mucha gasolina. 6. Quiero ir al restaurante que est en la esquina. 7. Necesito encontrar un trabajo que ofrez ca beneficios. 8. Tengo un amigo que es simptico y gracioso. 9. Quiero un jefe que sea inteligente y justo. 10. Necesito un empleado que haga buenas decisiones.

PAGE 539

A About the Author Victoria Russell has been a foreign language educator for the past twenty years; she has taught Spanish to students of every age from the elementary through post secondary levels. In the late 1990s, she lived abroad in Spain and the United Kingdom, where she taught Spanish at the regional college leve l. More recent ly, Victoria has taught ESOL courses and has supervised foreign language students in their final teaching internships At present, she is directing a large scale Spanish distance learning program that enrolls half of the basic language students at her uni versity, and she is also responsible for training graduate teaching assistants in the basic Spanish and French language programs. Victoria earned a B.S in Business Administration with a major in International Business and a Specialization in Spanish from Auburn University in 1990, and she earned a Master of Arts in Teaching Spanish Language and Literature in 1994 from Jacksonville University.


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 2200421Ka 4500
controlfield tag 001 002069309
005 20101007145827.0
007 cr bnu|||uuuuu
008 100421s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003189
035
(OCoLC)608202621
040
FHM
c FHM
049
FHMM
090
LB1601 (Online)
1 100
Russell, Victoria.
4 245
The effects of processing instruction, structured input, and visual input enhancement on the acquisition of the subjunctive in adjectival clauses by intermediate-level distance learners of Spanish
h [electronic resource] /
by Victoria Russell.
260
[Tampa, Fla] :
b University of South Florida,
2009.
500
Title from PDF of title page.
Document formatted into pages; contains 524 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2009.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
520
ABSTRACT: This study investigated the effects of processing instruction (PI) on the acquisition of the subjunctive in adjectival clauses by 92 intermediate-level distance learners of Spanish. PI is a novel instructional technique that is based on VanPatten's principles of input processing (1993, 1996, 2002, 2004), and it has three key components: (a) an explicit explanation of grammar that is not paradigmatic, (b) information on processing strategies, and (c) structured input tasks and activities. Structured input activities were isolated and combined with computerized visual input enhancement (VIE) in an attempt to increase the salience of targeted grammatical forms for web based delivery. VIE was operationalized as word animation of subjunctive forms through flash programming language.An experiment comparing four experimental groups with traditional instruction indicates that for interpretation and production tasks, there were no significant differences between PI and traditional instruction. However, learners who received PI combined with VIE outperformed learners who received structured input activities without VIE for interpretation tasks. In addition, the present study examined the effects of PI when learners encountered targeted forms that were embedded in an authentic input passage that was received following the experimental exposure. Thus far, studies in the PI strand have only examined how learners interact with structured, or manipulated, input.The results of the present study indicate that participants who received PI in combination with VIE noticed targeted forms in subsequent authentic input with metalinguistic awareness, and they demonstrated a significantly higher level of awareness than participants who received traditional instruction or structured input activities. Further, learners who received PI, with or without VIE, were better processors of targeted forms that were embedded in subsequent authentic input than learners who received structured input activities without VIE.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Co-advisor: Marcela van Olphen, Ph.D.
Co-advisor: Camilla Vsquez, Ph.D.
653
Complex L2 grammar
Distance language learning
Focus on form
Spanish as a foreign language
Web based instruction
Word animation
0 690
Dissertations, Academic
z USF
x Second Language Acquisition and Instructional Technology
Doctoral.
773
t USF Electronic Theses and Dissertations.
856
u http://digital.lib.usf.edu/?e14.3189