|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Exploring Writing o f English Language Learners i n Middle School: A Mixed Methods Study b y Robin L. Danzak A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Communication Scien ces and Disorders College of Behavioral and Community Sciences University of South Florida Major Professor: Elaine R. Silliman Ph.D. Ruth H. Bahr, Ph.D. Patricia Alvarez McHatton, Ph.D. Jacqueline Messing, Ph.D. Louise C. Wilkinson, Ed.D. Date of Approval: May 4, 2009 Keywords: Bilingual, academic language proficiency, identity, writing topic, writing assessment Copyright 2009 Robin L. Danzak
Dedication This dissertation is dedicated to the thousands of English language learners attending as they acquire a new language, culture and identity May they rise up like the phoenix.
Acknowledgements GRAT EFUL is the overwhelming feeling I have about my life and I would like to express that here to the many people who supported me throughout the dynamic journey that is the Ph.D nearly adequate t o convey my appreciation, but for now, on this page, it will have to do. I would first like to express my admiration and gratitude to m y academic advisor, Dr. Elaine Silliman. Over the last several years, Dr. Silliman has been both an inspiration and a dri ving force equally nurturing and challenging me through out the doctoral adventure and into the first hopeful steps of an academic career. I extend heartfelt thanks as well to the rest of my committee : Dr. Ruth Bahr, who helped me focus and remain calm, even in the face of nebulous statistical analyses; Dr. Patty McHatton whose passion for education and qualitative research reflects in her enthusiasm for teaching and mentoring students; Dr. Jacqu eline Messing, whose cours e in Educational A nthropology was both timely and influential in my career and research ; and Dr. Louise Wilkinson, who has been both inspiring and welcoming as a mentor and friend. Along the way I have been fortunate to cross paths with many excellent teachers and students who have had an impact on me as an educator and researcher. Some of these people made direct contributions to this investigation I extend many thanks to t he Bayview Middle School community and district for their hospitality and enthu siastic participation in this project. I deeply appreciate Cindi Garrett who dedicated many hours to transcribing, coding, and organizing data unstoppable, positive energy and bright spirit kept the project rolling, and I am grateful for her coll aboration and
friendship. I send m any t hanks to Flix Matas, Justine VanDyke, and Kyna Betancourt for check coding and providing feedback Finally, I appreciate Dr. Judith Bryant for providing methodological advice and serving as the outside chair at the dissertation defense. I am grateful for my wonderful family and friends, who have provided support and listening ears over the years My parents, Rosalie and Roger Stockseth, have always been open minded and supportive of all (or most) of my crazy ideas. Finally, I extend my love and gratitude to my dedicated husband, David, who has been loving patient, and understanding over the course of many weekends and events spent in front of the computer or at the librar y. Much love and appreciation go out to thes e people and the many others who have smiled upon this journey. A well known line from the Spanish poet, Antonio Machado provides some sense of closure to this adventure, as well as a look at what is to come: Caminante no hay camino, se hace camino al an dar. Walker, there is no path. The path is made by walking.
i Table of Contents List of Tables x List of Figures xii Abstract xiii CHAPTER 1: Introduction 1 Rese arch Problem 1 Purpose 3 Concept ual Framework 3 Review of the Literature 6 Theories of Writi ng Development 7 Cognitive F ramew orks of W riting 7 Social F ram eworks of W riting 11 Bilingual Wri ting Development 14 Research Findings 14 Interpretive Framework 17 Influence of Instruction on ELL Writing 18 Modification of I nstru ctional S trategies 18 ELL Curricula and Student E xpectations 26 L1 and L2 Interaction 29 Evidence for Cross Language Transfer 29
ii Some Caveats to Cross Language Transfer 30 L1 L2 Interaction in Writing 31 Assessment Measures for Bilingual Writing 37 Examination of Syntactic Structures 37 Comparison s across Languages, Genres, and Modalities 39 Summary and Conclusions 43 Research Questions 47 CHAPTER 2: Method 48 Ove rview of the Research Design 48 Mixed Methods Designs 48 Informal Interviews of Teachers of English as a Second Language (ESL) 51 Participants 52 Partici pants for Quantitative A nalysis 52 Focal P articipant s for Qualitative A nalysis 56 Data Sources and Collection Procedures 57 Participant W riting 58 Additional Data Sources 63 Participant Questionnaire 63 Participant I nterviews 64 Analysis 65
iii Quantitative Measures: Coding, Scoring, and Agreement 67 Segmentation of Texts into T Units 68 Lexical level: Noun Tiers and Number of Different W ords 69 Coding of Noun Tiers 69 Scoring of Noun Tiers 73 Number of Different Words (NDW) 73 Syntactic Level: Clausal Complexity Measure and Mean L ength of T Unit 73 Coding Clausal Complexity 73 Scoring Clausal Complexity 74 Mean Length of T Unit (MLT) 75 Discourse level: Analytic S cales 75 Inter Judge Agreement 78 Quantitative Measures: Statistical Analysis 79 Qualitative Analysis 80 Overview 80 Domain and Taxonomic Analyses 81 Domain Analysis 81 Taxonomic Analysis 83 Data Displays 84 CHAPTER 3: Results 85 Quantitative Results 85
iv Statistical Analysis 85 Lexical Level: Noun Tiers and NDW 86 Noun Tiers 86 NDW 89 Syntactic Level: Clausal Complexity and MLT 91 Clausal Complexity Measure 91 MLT 94 Discourse L evel: CSE S cales 96 Qualitative Results 99 Profiles 104 Diego: Futbolista Trilinge (Trilingual Soccer P layer) 105 Background 105 Coming to the U.S. 105 Language Le arning 106 Language Use and Bilingualism 106 107 Carolina: Modista en Pars (Fashion D esigner in Paris ) 108 Background 108 Coming to the U.S. 108 Language Learning 109 Language Use and Bilingualism 110 111
v ngl s No Me Gusta Para N (I Could Care Less a bout Englis h) 113 Background 113 Language Learning 113 Language Use and Bilingualism 114 114 Sara: Familia U nid a (United F amily) 115 B ackground 115 Coming to the U.S. 116 Language Learning 116 Language Use and Bilingualism 117 117 Manuel 118 Background 118 Coming to the U.S. 118 Language Learning 119 Language Use and Bilingualism 119 120 e next Michael Jordan) 121 Background 121 Coming to the U.S. 122
vi Language Learning 122 Language Use and Bilingualism 123 12 4 Summary of Profiles 126 Cross Case Analysis : Bilingual Perspectives 126 Bilingual Identity and Positive Views of Bilingualism 126 Monolingual Identity and Negative Views of Bilingualism 131 Results Summary 134 Summary of Quantitative Results 134 Lexical Level Results Summary 134 Syntactic Level Results Summary 134 Discourse Level Results Summary 135 Summary of Qualitative Results 135 Profiles Results Summary 135 Cross Case Results Summary 136 CHAPTER 4: Discussion 137 Discussion: Quantitative Patterns 138 Lexical Level Patterns 138 Noun Tiers 138 Factors Influencing Consistency of Coding 140 NDW 143 Implications 143
vii Syntactic Level Patterns 146 Clausal Complexity Measure 146 MLT 147 Factors In fluencing the Measurement of Syntactic Complexity 148 Implications 150 Discourse Level Patterns 154 Factors Affecting the Assessment of Overall Text Quality 156 Implications 158 Quantitative Conclusions 161 Impact of Topic 161 Transfer of Academic Language Proficiency 162 Knowledge Telling Orientation to Writing 162 Multiple Linguistic Levels of Text Composition 162 Discussion: Qualitative Findings 163 Patterns of Language Learning and Identity 163 Language Preferen ce 163 Bilingual and Monolingual Identities 164 Language Discrimination 165 Ethnic Differences 166 Factors Influencing Legitimation 166
viii Implications 167 General Discussion 169 A Multifaceted Look at Language Interaction 170 Language Identity and Academic Language Learning for ELLs 173 Diversity and Uniqueness 173 Regularity and Variance 174 The Non Emerging Bilingual Writer 176 The Dominant Emerging Bilingual Writer 177 The Balanced Emerging Bilingual Write r 178 Research Agenda 179 Follow Up Studies 180 Instructional/Intervention Strategies 183 Final Thoughts 184 References 186 Appendices 205 Appendix A: Summary of Data from Informal ESL Teacher Interviews 206 Appendix B: Summary of Res ults of Participant Questionnaire 209 Appendix C : Writing Prompts for Forma l Samples and Journals 21 1 Appendix D ingual Autobiographies 2 19
ix Appendix E : Partici pant Questionnaire 22 1 Appendix F : Interview Guide for Focal Participants 22 3 Appendix G : Data Displays for Qualitative Analysis 2 25 About the Author End Page
x List of Tables Table 1 Participa nts by Gender, Grade Level Family Origin, and Place of Birth 55 Table 2 The Six Focal Participants for the Qualitative Analysis 57 Ta ble 3 Topics for Formal Writing Samples 59 Table 4 Schedule of Data Collection: Participant Writing 63 Table 5 Research Questions, Assessment Measures, and Data Sources 66 Table 6 Noun Tiers ) Noun Scale 72 Table 7 Medians and Interquartile Ranges for Noun Tier Scores 87 Table 8 Medians and Interquartile Ranges for NDW 90 Table 9 Medians and Interquartile Ranges for Clausal Complexity Scores 92 Table 10 Medians and Interquartile Ranges for MLT 95 Table 11 Medians and Interquartile Ranges for CSE Global Writing Scores 97 Table 12 Journal Topics and Genres 101 Table 13 103
xi Table 14 Summary of ESOL Teacher Interview Data 206 T able 15 Personal and Educational Characteristics of Participants 209
xii List of Figures Figure 1 Graphic Interpretation of the Simple View of W riting (Adapted from Berninger & Hooper, 2003; Berninger et al., 2002) 9 Figure 2 Average Ranks within Genre T o pic for Noun T iers 88 Figure 3 Average Ranks within Language for Noun Tiers 89 Figure 4 Average Ranks within Genre T o pic for Clausal Complexity 92 Figure 5 Average Ranks within Language for Clausal Complexity 94 Figure 6 Average Ranks within G enre T o pic for CSE Global Scores 98 Figure 7 Average Ranks within Language for CSE Global Scores 99 Figu re 8 Data Display for Diego 225 Figure 9 Data Display for Carolina 226 Figur e 10 Data Display for Edgar 227 Figure 11 Data Display for Sara 228 Figure 12 Data Display for Manuel 229 Figu re 13 Data Display for Juan 230 Figure 14 Data Disp lay for Cross Case Analysis 231
xiii Exploring Writing of English Language Learners in Middle School: A Mixed Methods Study Robin L. Danzak ABSTRACT features of 20 Spanish speaking English language learners (ELLs) in middle school. Students came from Mexico, Puerto Rico, and the Dominican Republic. Participants wrote two expos itory and two narrative formal texts, each in Spanish and English, for a total of eight writing samples each Additionally, students developed 10 journal entries in their language of choice, and 6 randomly selected, focal participants were interviewed for the qualitative analysis. The quantitative analysis involved scoring formal texts at the lexical, syntactic, way ANOVA by ranks, and resulting ranks were compared across genre topic and langu age. A key outcome was that the text topic rather than genre or language, impacted on rank differences at all levels, possibly due to student engagement or influence of the prompt structure. P erformance at the three levels was essentially similar across both languages, revealing that participants were emerging writers in Spanish and English. Similar outcomes in Spanish and English also implied potential cross language transfer of
xiv academic language proficiency. Results further highlighted the interaction o f multiple linguistic levels in text composition. Finally, s tudents appeared to apply a knowledge telling strategy to writing, resulting in unsophisticated vocabulary and structures. w transcripts were analyzed with domain and taxonomic analyses to discern how their language learning experiences shaped their identities as bilinguals. Results showed that 1) Spanish was preferred for all focal participants; 2) students shared the experie nce of language discrimination; 3) bilingual and monolingual identities resulted in different attitudes toward language learning and varied writing performance ; and 4) Mexican and Puerto Rican students had diverse language learning experiences, leading to differences in identities and writing outcomes. Overall, the quantitative and qualitative fi ndings raise two questions: 1) which aspects of academic language proficiency are shared across both languages, and how might these be assessed with bilingual, int egrated language measures ? 2) How might integrated assessment in L1 and L2 aid in identifying adolescen t ELLs with language impairment?
1 CHAPTER 1 Introduction Research Problem Recently, there has been a national push to improve the literacy achievement of & Snow, 2004 ; Graham & Perin, 2007). According to the Alliance for Excellent Education (2007), less than one third of adolescent students achieve at grade level expectations in reading. In the case of writing, test scores across the country are equally as disturbing. On the last National Assessment of Educational Progress (NAEP) writing examination, given in 2007, the majority of 8 th graders (88%) scored at or above the basic level, but only 33% achieved a level of at or above proficient. For students in 12 th grade, 82% scored at or above basic, while only 24% achieved at or above proficient ( Salahu Din, Pers ky, & Miller, 2008) One group of adolescent students in need of greater attention in educational research and policy development is the growing, diverse population of English language learners (ELLs). Indeed, existing studies involving ELLs of all ages h ave documented that their literacy abilities often fall well below those of their native English speaking peers (Goldenberg, Rueda, & August, 2006; Gutirrez et al., 2002; Shanahan & Beck, 2006). According to Cummins and Schecter (2003), it takes the avera ge ELL at least five years to master the academic language skills necessary to catch up to native English speaking students. Similarly, Short and Fitzsimmons (2007) emphasized that adolescent ELLs in ) of their native English
2 speaking peers in that they are expected to acquire English simultaneously as they learn academic content through English. In addition, many of these students, in spite of having been schooled in the United States for several year s, are still developing academic being held to the same accountability standards as their native English (Short & Fitzsimmons, 2007, p. 1). These researc hers called for national scale improvements to the assessment and instruction of adolescent ELLs, as well as policy changes (e.g., NCLB definitions, school accountability criteria) and increased funding for research in this area. In spite of the literacy challenges apparent for adolescent English learners, much of the ELL and bilingual literacy research has focused on acquisition of early skills, such as phonological processing, word reading, and vocabulary development (Bialystok, 2007; Geva, 2006; Shanaha n & Beck, 2006). Hence, many of the studies on ELL literacy development have sampled students in the lower grades. In addition, there has been little investigation into the area of ELL writing, in spite of this being an area of concern due to the general p oor quality of ELL written work (Geva, 2006; Geva & Genesee, 2006). In addition, very few studies to date have examined micro level, specific linguistic features of ELL writing at the lexical, syntactic, and/or discourse levels. Clearly, there is a need fo r increased understanding of the development of second language writing of adolescent English learners.
3 Purpose The purpose of the present mixed methods study was to provide, through quantitative analyses, an in depth, cross language investigation of the linguistic features of the expository and narrative writing of 20, Spanish speaking ELLs attending a public middle school on the west coast of Florida. A secondary goal was to explore, through qualitative methods, how the language and literacy learning exp eriences of a randomly selected sub group of 6 focal participants may have developing identities as bilingual writers. Participants produced a series of expository and narrative written texts in Spanish and English for the q uantitative analysis, which involved the application of lexical, syntactic, and discourse assessment measures to investigate the variables of language (Spanish/English) and genre topic (expository/narrative, topic 1 or 2). Written journals and interviews o f the focal participants provided data for the qualitative analysis, which included domain and taxonomic analyses (Spradley, 1979) and the creation of data displays (Miles & Huberman, 1994). Conceptual Framework The present investigation was carried out ba sed on a sociocultural, constructivist view of identity, learning, and second language acquisition. Within this framework, identities are perceived as multiple and contextually or situationally motivated (DeFina, 2006). In addition, identity is largely con structed and expressed through discourse (Riley, 2006). For example, Sfard and Prusak (2005) described identity as a set of reifying,
4 structured through stories that people actively construct for themselves and others. In the case of second language (L2) learning in the context of school, there is relevant instructional practices may stren gthen social identity and increase student investment (Daisey & Jos Kampf n er; Meltzer & Hamann, 2004; Norton Peirce, 1995), as well as contribute positively to literacy achievement (Au & Mason, 1981; Cummins et al., 2005; Flores Dueas, 2004; Prez, 2004) In addition, the sociocultural perspective highlights their participation in the cultu ral practices of their homes and communities (Collins & Blot, 2003; DaSilva Iddings & Katz, 2007; Flores Dueas, 2004; Gee, 2004; Gee, 1996; Moll, Amanti, Neff, & Gonzlez, 1992). With regard to the incorporation of student resources as a bridge to learni ng, a key theoretical basis for the current study was the assumption that relationships exist between the first language (L1) and L2 that not only influence oral language production, but also play important roles in biliteracy learning and academic languag e acquisition (Bialystok, 2007; Bialystok, Luk, & Kwan, 2005; Durgunoglu, 2002; Francis, 2006; Geva, 2006; Geva & Genesee 2006). Research reporting evidence of cross language transfer of common underlying proficiencies by second language learners supports the notion that general cognitive and language skills, concepts, and metalinguistic awareness can be shared across languages (Bialystok, 2007; Cummins, 2000; Francis, 2006). That is,
5 common underlying proficiencies in L1 can provide scaffolds for the deve lopment of related knowledge and skills in L2, particularly in the case of well established, higher level cognitive processes and skills. These processes and skills include, for example, phonological awareness (Bialystok, 2007; Geva & Genesee 2006; Lafran ce & Gottardo, 2005), vocabulary depth (Ordez, Carlo, Snow, & McLaughlin, 2002), text production (Cummins, 1991), and knowledge of writing conventions and story grammar (Durgunoglu, 2002). In keeping with the sociocultural conceptual framework, the frame work of the cognitive resources to be accessed in the context of L2 language and literacy learning. Like funds of knowledge (Moll et al., 1992) and other cultural and exp eriential resources, strengthen writing development in L2. In the present study, participants developed various written texts in Spanish and English that were used cross L1 knowledge and skills, prior experiences, funds of knowledge, and cultural experiences and practices. In addition, implementation of a mixed methods design provided the opportunity for an in depth investigation of adolescent ELL writing within the context of a bilingual autobiography project that took place in the middle school English as a Second Language (ESL) classroom.
6 Review of the Literature As mentioned previously, there has been little scholarly inquiry into the area of ELL writing. Prior research has typically fallen into four categories: 1) bilingual writing development; 2) the influence of instruction on ELL writing; 3) the rel ationship between L1 and L2 writing; and 4) assessment measures of bilingual/second language writing. Studies reviewed here are organized based on these categories. The first area, bilingual writing development considers similarities and differences betwe en monolingual and bilingual writing acquisition, as well as characteristics of biliteracy development in general, particularly in the case of young children learning two languages simultaneously. The second area, influence of instruction on ELL writing c enters on effective ELL teaching strategies as well as challenges to instruction, such as curriculum reduction for language minority students. The third area, first and second language interaction in writing, explores the relationships between L1 and L2 wr iting, for example, the transfer of academic language skills, common underlying proficiencies, and use of L1 as a resource in second language literacy development. Indeed, studies focusing on writing development and writing instruction frequently highlight cross linguistic relationships and/or influences. Because this theme is central to the present study, findings related to the topic of cross language relationships will flow throughout the literature review. Finally, the fourth area, assessment measures f or (bilingual) writing, presents some of the various measures that have been applied to explore bilingual/second language as well as monolingual writing at the lexical, syntactic, and discourse levels. The review of the
7 literature on bilingual writing is p receded by a brief summary of predominant theories on the process of writing development in general. Theories of Writing Development In their meta analysis on strategies to improve adolescent writing in middle and high schools, Graham and Perin (2007) hi ghlighted two distinct yet interrelated roles played by writing in the school setting: 1) writing as a skill composed of a conglomerate of strategies that is used to accomplish communicative and academic goals; and 2) writing as a medium through which subj ect matter is learned. However, years of literacy learning goals. This section briefly describes writing development, highlighting predominant theories of writing as both a cognitive process and a social practice. Cognitive frameworks of writing. Writing is commonly defined as a complex, dynamic process that involves the integration of multiple levels of cognitive and motor skills and processes. This process oriented conceptualization of writing was initially writing, which replaced earlier linear models. Berninger and Hooper (2003) summarized l inear, recursive process in which the planning, translating of plans into written text, and the reviewing/revising processes continually been revised over the years to co nsider also the task environment. This includes elements of the physical and social contexts of writing, such as the composing medium,
8 collaborators, and audience and, in the case of student writers, McCutchen (2006) added the instructional context. Notwit hstanding the impact of the Hayes and Flower model on how writing is understood and instructed, the original model and its revisions have been criticized for their focus on the practices and processes of expert writers rather than novice writers or childre n (McCutchen, 2006). Certainly developing writers differ from experts in their processes of text planning, production, and revision. Bereiter and Scardamalia (1987) proposed an alternative to the Hayes and Flower model, confirming that children make use of different control processes in their writing than do more experienced producers of written text. Specifically, according to Bereiter a nd Scardamalia, young writers do not employ sophisticated planning, translating, and revising processes Instead, less ex perienced writers engage in a simplified, Scardamalia, 1987, p. 5) which involves recalling experiences or facts based on the assigned writing topic and telling about these in the text McCutchen (2006) confirmed th at the practice of knowledge telling has been evidenced by research on children (e.g. ages 10 12 years), which has shown that composition at this stage is dominated by text generation processes, while conceptual planning and revising are rarely observed un til later in adolescence. In search of a more holistic view of writing that would take multiple strategies and processes into account, Berninger and colleagues (2002) integrated cognitive, developmental, neuropsychological, and educational theoretical fram eworks into a proposed simple view of writing. In this triangular model, transcription processes of
9 handwriting and spelling along with executive functions such as attention and planning form the vertices. The goal, text generation, is represented by the apex of the triangle. (Berninger & Hooper, 2003, p. 6) occur in a working memory environment, represented by the inside of the triangle. Figure 1 summarizes this model. Figure 1 Graphic interpretation of the simple view of writing (adapted from Berninger & Hooper, 2003; Berninger et al., 2002) Working memory environment Transcription processes Text g eneration Executive function
10 With respect to the interactive nature of these processes, Berninger et al. (2002) stated: Deficits in transcription skills or neurodevelopmental processes related to transcription can interfere with development of the text generation e automatic low level transcriptions skills are, the more capacity limited, working memory resources are available for high level composing skills (p. 292). In this way, the simple view of writing accounts for competition among cognitive resources as text composition plays out. Along similar lines, Torrance and Galbraith (2006) discussed the processing constraints faced by developing writers. These constraints include dual task interference taneously), the transience of short term memory, and the processing demands of the strategies required for the writing task itself. These authors suggested that effective writing development required automaticity of low level components (spelling and handw riting) and efficient memory management strategies, such as the ability to shift attention among competing task demands. Indeed, Torrance and Galbraith (2006) highlighted the flexibility necessary for productive and successful composition of written texts: No matter how skilled we are at managing the writing process, there is an irreducible core of potential conflicts and writing will always be a struggle to reconcile competing demands. Writers motivationally have to accept this if they are to get the task done. (p. 78).
11 Often the demands of writing include those imposed by the social context of the writing event. The purpose, goals, audience, context of instruction, and the topic and expectations of the writing assignment itself are all influenced by s ociocultural factors. Taking this into account, Berninger and Hooper (2003) highlighted the need for writers to develop sensitivity to the social context of writing in addition to the cognitive processes involved in writing. Social frameworks for writing w ill be addressed in the section that follows. Social frameworks of writing. According to Prior (2006), the dominant writing research paradigm is currently driven by sociocultural theories. In contrast to cognitive theories of writing, sociocultural framew orks highlight the dialogic processes of text composition: because writers produce texts within a given sociocultural political historical context and utilize culturally appropriated tools, resources, and practices, all writing is viewed as socially mediat ed, distributed, and collaborative (Prior, 2006). For example, a white, monolingual English speaking child attending school in the suburban U.S. will learn and utilize literacies differently than an adult, Spanish speaking migrant worker studying English t hrough a community program. These individuals will also acquire different types of symbolic capital based on language status, power structures, and other contextual factors (Christian & Bloome, 2004). Similarly, as viewed through the sociocultural lens, wr iting may be conceptualized not only as a means of communication, but also as a medium for social action (Berdan et al., 2006; Collins & Blot, 2003).
12 In the case of writing in school, instructional context, power relations, the purposes and goals of writi ng, as well as audience and collaboration (with other students writing. Along these lines, Schultz and Fecho (2000) pointed out the following key characteristics of wri ting development: 1) it integrates social historical contexts; 2) it is nonlinear and varies across contexts; 3) it is influenced by social interactions, including classroom curriculum and instruction; and, 4) it shapes social identities of students. This being said, sociocultural frameworks of writing (and literacy, in general) often highlight the integration of literacy learning and the development of social identity in school stud ents to be readers and writers is as much a matter of language socialization, xvii). These issues take on a particularly important role in the case of language minority students. For example, Christian and Bloome (2004) pointed out that, for ELLs, the relationships between literacy learning and social identity formation in the cla ssroom minority status may result in their marginalization during literacy events, which diminishes their social capital, social status, and ultimately, academic perfo rmance. This pattern may also occur in the case of students with a language or learning disability as well (Brinton & Fujiki, 2004; Danzak & Silliman, 2005; Ruiz, 1995).
13 In relation to social identity development, Ball (2006) noted that sociocultural theo learning. Indeed, sociocultural frameworks view both the individual and the context as dynamic and multifaceted, and emphasize the complex interrelationships among peop le and contexts. Hence, the context of writing in school must also take into account ELLs, may not neatly parallel the behaviors and skills expected and valued at school. An the context of broader literate practices in the home, community and workpl ace, but also rather is connected at multiple levels to community practic es and identities. This statement highlights the sociocultural perspective of the writer as having multiple identities and participating in multiple contexts, again, a framework particularly relevant to ELLs. With regard to the ever changing sociocultural tablet upon which developing terms of writing development for students to bring their home cultures, peer cultures, xt in the process of being this type of question. The following sections begin to explore this question through an
14 overview of the literature on biliteracy: bilingual writing development, instruction, assessment, and the interaction of L1 and L2 throughout this process. Bilingual Writing Development Bialystok (2007) noted that the three prerequisites for literacy -oral language competence, understanding symbolic concep ts of print and metalinguistic awareness -are all differently influenced by bilingualism. For example, while bilingualism can enhance reduced vocabulary breadth or di versity in eac h system (Bialystok, 2007; Cobo Lewis, Pearson, Eilers, & Umbel, 2002). In addition, proficiency is likely to vary across the languages, which may lead to challenges for literacy acquisition in the less proficient system. For these reasons as well as others, it is important to remember that bilingual literacy development is unique; that is, a bilingual child may experience developmental patterns in each language that differ from those of monolingual speakers of each language. Research finding s. With this in mind and noting the scarcity of published research on early bilingual writing development, Rubin and Carlan (2005) explored the development of writing in Spanish and English for over 100 bilingual writers, ages 3 10 years, in low income ele mentary schools with bilingual programs on the US Mexico border. Children were asked to draw pictures of things they liked and describe them both developmental levels o f writing, created for monolingual English speakers, and Ferreiro
15 1982) comparable five stage system designed to explain writing development in Spanish. In their analyses of Spanish and English writing samples of the bilingual children, R ubin and Carlan (2005) determined that bilingual stages were generally similar to the monolingual stages across languages; however, there were also differences. For example, ding 3) often wrote the same symbols for both languages and read them differently in English and Spanish. This finding confirmed a similar observation by Moll, Saez, and Dworin (2001), who also found this pattern in the ir qualitative, multiple case study of two kindergarten students developing early biliteracy skills. As children advanced, for example, to the phonetic stage/level 4, Rubin and Carlan (2005) attributed writing errors to different letter sound relationship s in the two sentence structure became more complex in both languages. An additional finding was that, as early as age 6 years, the bilingual children used code switch ing and strategically applied their general knowledge of both languages in writing as communicative resources. These findings also support the general notion that, for a bilingual individual, first and second language competence and literacy skills have a mutual influence on each other in the case of biliteracy learning and development. Despite the value of these findings, it is notable that Rubin and Carlan (2005) stage
16 models that were compared. That is, writing samples were explored and described by general characteristics of text production and errors, for example, representation of each syllable as one vowel, omission of silent letters, and use of compound/complex sentences A more fine grained, micro analysis of developing biliterate writing is needed to better compare and contrast the lexical and syntactic features of ELL writing across both Spanish and English. Also in relation to biliteracy development, Edelsky (1982) e xamined changes over time in the writing of first through third grade students in Spanish and English, and how Participants attended a bilingual program at a school serving pri marily Spanish speaking children from migrant families, located in a semi rural area near Phoenix. To explore biliteracy development, Edelsky incorporated multiple sources of data, including samples ish, 49 in English), interviews with teachers and aides, classroom observations, a language situation survey, observation of parent events, and school records. She found examples of L1 (transfer) in segmentation, spelling, and personal sty le, as well as L1 (differences) in segmentation, spelling, use of tildes/accents, syntactic complexity, style, code switching, and handwriting. Edelsky (1982) concluded that the importance of these findings could be determined based on the perspective of the reader. Therefore, if writing was viewed as a hierarchical, linear set of skills applied invariantly across contexts (as many teachers at that time believed, according to the author), the findings would emphasize examples of
17 nonappl ication (non transfer). On the other hand, from the perspective that writing represents the context dependent orchestration of multiple cuing systems of global and highl ighted. Indeed, the author emphasized that children utilized linguistic input differently in each language depending on available resources and contextual constraints. Interpretive frameworks. contemporaries, spe cifically the cognitive process writing model of Hayes and Flower (i.e., interviews, observations, and document review) also highlighted the context of student writing i n the participating school, and included descriptions of both classroom driven by the cognitive processes framework at the moment of analysis, is framed within a broa through, not merely in, For Edelsky, there exist underlying L1 writing processes or common underlying proficiencies (Cummins 2000) that can be applied to L2 writing. Many researchers support the idea that these common underlying proficiencies can be incorporated as linguistic resources into language and literacy instruction of ELLs (Bialystok, 2007; Cummins et al., 2005; Dur gunoglu, 2002; Francis, 2006; Koda, 2007). A further step in this process would be to acknowledge and integrate other cultural resources that language minority students bring to the classroom, for example, skills and practices that they experience in their homes and communities (Ball, 2006; Moll et al., 1992; Moll, Saez, &
18 Dworin, 2001). This brings us to the second category of second language writing research, which deals with writing instruction for ELLs. Influence of Instruction on ELL Writing Two gener al themes emerge from the research regarding the influence of instruction on ELL writing development. One is that, although many instructional strategies found to be effective for ELLs are examples of simp ly good teaching (Rueda, August, & Goldenberg, 2006 ), these highly effective teaching strategies must be strategically modified to meet the unique needs of ELLs (Escamilla, 2006; Shanahan & Beck, 2006; Walqui, 2007). Another underlying theme in this body of research is that teachers of ELLs should take car e not to water down the curriculum or lower expectations for their students (Boyd & Brock, 2004; Moll, 1 986; Koelsch, 2006; Rueda et al., 2006; Walqui, 2007). Modification of instructional strategies. Regarding the first of these broad themes, Bialystok, Luk, and Kwan (2005) noted that biliteracy acquisition is different from monolingual literacy development in two key ways. First, a bilingual child brings different background skills to the literacy learning experience. For example, their oral proficiency, vocabulary, metalinguistic awareness (including phonological awareness), and overall cognitive development differ from those of monolinguals in each language. In addition, and noteworthy for the present investigation, a biliteracy learner has the ability to transfer skills from one language to the other. What remains under investigation, however, is the question of which strategies and skills are readily transferable.
19 Rueda and colleagues (2006) reviewed multiple studies on the sociocultural context of bi literacy learning and concluded that additive language learning environments include those in which experienced teachers emphasize cooperative learning, critical thinking, and whole language, and utilize technology, home school collaboration, and thematic, integrated curricula. However, these conclusions are based on descriptive investigations that did not report any outcome data. Hence, according to Rueda et al. factors instruction appears to vary widely across the literature reviewed. In contrast to Rueda et al. (2006), Shanahan and Beck (2006) provided a qualitative, narrative review of studie s of literacy instruction for ELLs in which they calculated effect sizes to compare results across investigations that met criteria for employing experimental, quasi experimental, or single subject designs. Three major findings were synthesized from the st udies reviewed: a) ELLs advanced more in their writing ability in instructional settings that used structured writing as opposed to free writing; b) no effects were found for use of cooperative learning groups to teach ELL writing; and c) revision training was generally helpful. Across the studies reviewed, effect sizes were smaller and more variable for ELLs than for monolinguals who participated as controls, indicating weaker practical significance and/or use of small sample sizes for ELL groups. This res ult may also be a consequence of the large inter subject variation that exists in the academic skills and literacy learning experiences among ELL students.
20 Shanahan and Beck (2006) concluded that teaching specific reading and writing elements can benefit E LLs, but teachers need to adjust instructional routines to meet the unique needs of these students. Similarly, Escamilla (2006) noted a caution: When it schooling of ELLs in general. For this author, these problems include a lack of research and an emphasis on an English ). One solution to this latter problem is that teachers of ELLs be knowledgeable to second language and literacy instruction. In this approach, students are provided w ith sufficient metalinguistic understanding about how the languages compare and contrast at various levels (Ball, 2006; Gutirrez Clellen, 1999). Along these lines, Rubin and Carlan (2005) recommended that teachers should recognize the unique features of b ilingual writing development, teach spelling patterns, and highlight similarities and differences teachers emphasize meaning, promote the use of bilingualism as a resour ce, support writing in both languages, encourage children to talk about how and why they write, and use writing to assess and plan further instruction. One caveat to these strategies is that teachers of ELLs may not have the knowledge, experience, or conf instruction. For example, in a discussion on the use of Spanish English cognates for
21 vocabulary, Snow and Kim (2007) advised that teaching about cognates may be too much to ask of educators who do not have sufficient knowledge of Spanish. Regarding the cross linguistic relations is unlikely to happen in classrooms where neither teacher nor p. 131). This statement is likely to generalize beyond cognate vocabulary to encompass other areas of language and literacy instruction for ELLs, such as comparison/contrast of phonetics/phonology, syntactic patterns, or discourse features across Spanish a nd English. In a review of research on teaching ELLs in the content areas (e.g., history, science, math, etc.), Janzen (2008) explored the linguistic, cognitive, and sociocultural features of academic literacy and the challenges these presented to ELLs. Ba sed on her findings, Janzen (2008) concluded that the language of academic texts includes distinctive features that differ across disciplines. For many students, especially ELLs, this academic language needs to be explicitly taught for students to successf ully interact with these texts. Janzen suggested that ELL students should be asked to engage with the material verbally, for example by thinking aloud, discussing in groups, or explaining learning strategies. In addition, this oral interaction with the con tent material was found explicit instruction of cognitive strategies, which may be improved with increase d teacher professional development, is critical to the succe ss of ELLs in the content areas.
22 (p. 1030). While this discussion was beyond the scope studies have considered the use of multicultural literature and/or the integration of culturally relevant resources, i.e., funds of knowled ge (Moll et al., 1992) as a pathway to engage ELLs in literacy instruction. In one example, Flores Dueas (2004 ) elicited written responses and group discussions about culturally relevant literature with four Mexican American students in grade 5. The participating students included two boys and two girls who attended an urban public sch ool in Texas, had been exited from transitional bilingual programs, and were considered by their teachers to be performing on average or above average in academic work in English. Flores Dueas (2004) framed her inves 1978) reader (text bound, reading for information) or aesthetic (experiencing text as primary, connection of text to self). The author found that the participants engaged more with literature written by Mexican American authors than literature from the dominant U.S. culture. In fact, the students were more likely to identify with the characters and experiences expressed in culturally familiar texts, and thus were able to provide an aesthetic reader response that included increased reflection on feelings, deeper interpretation of the text, and higher level writing. On the other hand, dominant culture literature evoked an efferent (i.e. text based) written response that was shallower, shorter in length, and offered little interpretation. The author concluded that the home school cultural mismatch experienced by many language minority students leads to lack of engagement because they cannot identify with
23 the literature of the dominant c ulture, which is typically included in the language arts curriculum. Although most literature is taught in school from an efferent perspective, when students are able to personally identify with literacy activities (aesthetic response), they are empowered to achieve ownership of literacy. Flores the course of one academic year. This qualitative study provides an interesting, albeit limited, example of the relationships between literac y learning and identity mentioned earlier. However, Flores legitimation (validity) of a qualitative design (Onwuegbuzie & Leech, 2007; Yin, 2003). For example, the author does not offer any criteria or rationale regarding participant selection and provides little information regarding data collection procedures or analysis. Specifically, it is stated that participants were interviewed, but no details are included about this process nor is an interview protocol provided. Also, the author briefly described the sessions in which participating students read, wrote, and discussed the literature; however, it is not clear how much time students had to write, how long the sessions lasted, or where they took pl ace. Finally, regarding data analysis, Flores Dueas (2004) videotapes of the gr distinction between efferent a nd aesthetic responses to particular texts. However, the categories or criteria employed to determine efferent or afferent response were not specified. Due to Flores it is difficult to under
24 Goldenberg et al. (2006) cautioned, it is difficult to substantiate research claims in the context of methodological shortcomings. Another qualitative study that offers an incorporation example o project with a bilingual class of Latino students in the 4 th grade. This project was based on a sociocultural conceptual framework and a funds of knowledge perspective (Moll et al., 1992). Participating students were requested to work with their parents and families to develop family stories that were then written, discussed, critiqued, and revised collaboratively in the classroom in a writing workshop format. In the workshops, English). Each student also prepared a translation of ano stories were published in books, one version for each language. Dworin (2006) found that 15 of 18 stories were originally written in Spanish, and ico, Easter in Guatemala, and childhood experiences of parents and grandparents. For this author, the collaborative nature of the workshops helped improve each story due to the inquiry of other students, who often demanded more details when stories were sh ared. Dworin also found an influence of English vocabulary in Spanish writing, for example, a stories was an effective way to encourage the development of metaling uistic awareness across languages, and also served as an individual and social accomplishment for the students.
25 including participant observation and direct involvement in fac ilitation of the family stories project at the school over the course of several months. The school, teacher, and students were described in detail, as was t he family stories project As a means to illustrate the writing produced by the participants, Dwori n (2006) highlighted the work of (2006) report. Similar to the other qualitative studies reported in this review, Dworin anecdotally described some salient features of these two compositions, but did not provide any systematic or in depth linguistic analysi s of any student writing. In addition, due to the descriptive nature of this study (the family writing project took place outside of the classroom and was not explored as an intervention), it is not possible to determine if the family stories project had a ny impact on the writing achievement of the participating students. Although there is something to be said about value of family involvement in (bi)literacy development (Goldenberg et al., 2006), more detailed, cross linguistic exploration of ELL writing i s necessary to learn how aspiring biliterate students develop skills across languages. With respect to the sociocultural influences on ELL literacy learning, Goldenberg ak evidence that sociocultural characteristics of students and teachers have an impact on
26 native language appeared to provide greater support for literacy acquisition o f language minority students than did culturally familiar literature written in L2. However, these authors cautioned that claims such as these are skeptical ones due to methodological problems, particularly in qualitative research. Goldenberg and colleague s identified the following four problems were noted by the authors: 1) insufficient triangulation of data; 2) lack of specific and/or sufficient information regarding participants, data collection procedures, and data analysis techniques; 3) lack of consid eration of competing hypotheses or interpretations; and 4) presentation of inferences or conclusions not examining the relationships between sociocultural factors and (Goldenberg et al., 2006, p. 253). ELL curricula and student expectations. With respect to the issue of curriculum reduction for ELLs, Moll (1986) observed through ethnographic research in bilingual communities that teachers of ELLs wer e likely to water down the curriculum to match participation in classroom ex periences and low levels of language and literacy instruction. Following a Vygotskian framework, Moll emphasize d that learning is a social process, and that early literacy practices should be about purposeful communication with others. Support for this pre mise was evidenced in home observations in a bilingual community in which most writing was functional and practical, involving such activities
27 contrast, in the class room Moll observed that writing was rarely used as a communicative tool. As a solution to this conflict, an attempt was made to increase the motivation and writing achievement of the participating ELLs. Moll (1986) and cooperating teachers implemented str uctured, supportive writing instruction that revolved around assignments involving parents and community members. The purpose of these assignments was to promote home literacy interactions and make writing more meaningful for the students. The author provi ded an example in which students were invited to interview adults and guidance) regarding their experiences with language learning, language use, and bilingualism. For M language difficulties while maximizing the use Moll, 1986, p. 107). study took place over 20 years ago, current research in ELL classrooms continues to report that teachers of ELLs hold low expectations for students h, 2006, p. 2). Koelsch argued that ELLs in high school should be encouraged to take challenging, college preparatory courses, as evidence has shown that academic achievement of adolescent ELLs is better predicted by track placement than English language p critical thinking and metacognitive skills as well as more complex literacy skills. Koelsch also noted that the quality of instruction in higher level classes may be an additional
28 factor in fluencing increased achievement for students in these courses. Overall, this author highlighted the need for educational systems to make structural changes in order to increase the educational opportunities of ELL students. Walqui (2007 ) agreed that expect ations, in tandem with structured support, need to be raised for ELL students. This researcher presented the objectives and research of Quality Teaching for English Learners (QTEL), a teacher professional development project sponsored by WestEd, a national nonprofit educational research agency. In her discussion, Walqui emphasized that ELLs require opportunities in which they can actively and legitimately participate in rich, high quality, well supported academic activities. These engaging, scaffolded oppor tunities provide students the support and intellectual challenges necessary to appropriate skills required to move from social to independent participation in related tasks. QTEL follows this same framework for the development and training of ELL teachers. Walqui summarized the major goals of QTEL as consisting of: academic rigor, high expectations, quality interactions, language focus, and quality curriculum. Although more research in this area is needed, it is clear that ELL students can benefit from a v ariety of quality instructional and learning strategies and experiences as they acquire literacy in a second language. Boyd and Brock (2004) summarized this well needs and experiences of ELLs. Instead, solid instructional strategies and appropriately
29 language and culture can create an environment that promotes literacy learning and school success for language minority students. L1 and L2 Interaction Evidence for cross language transfer. With respect to cross linguistic relationships in biliteracy, Francis (2006) asked a classic chicken egg question: "Does new competence develop in response to experience with academic literacy, or do children learn how to use previously acquired knowledge in a new way?" (p. 46). For Francis bilinguals depend on both shared (general) and language dependent (specific) mechanisms and structures as they develop secondary discourse abilities and metalinguistic awareness in both languages. Based on this hypothesis, bilinguals have a unified, underlying academic language proficiency as well as some language specific knowledge and skills, for example, grammatical re presentations. For this reason, Francis argued that a language general skill, such as phonological awareness, will transfer from L1 to L2, while language specific features, such as syntax, are less likely to transfer across languages. The research reviewe d here generally supports the notion that higher level competencies, potentially represented in a shared underlying system, are likely to transfer across languages. These competencies include: phonological awareness (Bialystok, 2007; Geva & Genes ee, 2006 ; Lafrance & Gottardo, 2005), vocabulary depth 1 (Ord ez et al., 1 Ordoez et al. (2002) described vocabulary depth as involving the quality of pho nological representations of words, knowledge of syntactic structures in which a given word is used, word class/es, morphological structure, richness of semantic representations, and pragmatic features of words.
30 2002), text production (Cummins, 1991), and knowledge of writing conventions and story grammar (Durgunoglu, 2002). In addition, there is evidence that, if students are proficient writers in the ir first language, these skills can transfer to a second language even if their basic oral communication skills are limited (Dressler & Kamil, 2006; Fitzgerald, 2006; Lanauze & Snow, 1989). Some caveats to cross language transfer. It is important to note that cross language transfer is not always an available or dependable route for ELLs to acquire or access skills in L2. For example, Francis (2006) noted that, in addition to what might eventually become shared features of academic discourse ability, bilin gual students must also master language specific features, for example, grammar. Additionally, although Harley and King (1989) suggested that lexical similarities, especially cognates, between (2007) cautioned that transfer of cognate vocabulary across languages. It is reasonable to understand how this statement can be generalized to highlight the impact of instruction (in both L1 and L2) on other academic language skills as well. For example, in a meta analysis of research findings for the effects of the language of instruction on literacy acquisition of ELLs, Francis, Lesaux, and August (2006) determined po sitive effect sizes in favor of bilingual education over English only education for ELLs. Certainly, a quality instructional context additive L2 acquisition framework in which ELLs can capitalize on the development of
31 common underlying proficiencies to support literacy in both languages (Cummins & Schecter, 2003; Gutirrez et al., 2002; Janzen, 2008; Rueda et al., 2006). L1 L2 interaction in writing. With regard to langua ge interaction specifically in bilingual writing, Fitzgerald (2006) reviewed 56 studies -both quantitative and qualitative -in a meta analysis of the last 15 years of research on multilingual writing in school. This author reported transfer from L1 L2 writ ing in spelling patterns and descriptive writing independent of L2 oral proficiency. In addition, the studies that Fitzgerald reviewed evidenced that, when L2 learners experienced difficulty with writing, problems were most likely due to faulty composing p rocesses rather than language factors. These findings show that text related core competencies may be present in writing across both languages, and also support the notion that higher level cognitive processes are more likely to transfer across languages. These processes include, for students in the primary grades, concepts of print and, for adolescent students, strategies for constructing meaning. author summarized four methodological issues that emerged from the meta proficiency and language learning conditions and experiences; 2) incomplete information regarding the data collection procedures used, measures applied, scoring systems, and reliability procedures including inter coder reliability; 3) insufficient detail regarding the methods of analysis; and, 4) the need to assess L2 writing with at least two writing samples for the purp oses of replication.
32 In addition to research regarding potential transfer of common underlying proficiencies (Cummins, 2000), there is also evidence that second language learners rely on their implicit knowledge of L1 to infer patterns of L2 at multiple l evels (e.g. pronunciation, spelling, and grammar), whether or not these inferences are correct. For example, Cronnell (1985) carried out an error analysis of the English writing of Mexican American ELLs in Grades 3 and 6, classifying errors as no influence influence of Spanish, influence of Chicano English, or interlanguage error. This author found that, although error types differed by age, English spelling errors included application of Spanish vowels and other Spanish influenced, pronunciation based err ors. More recently, Escamilla (2006) worked with teachers of Spanish speaking ELLs to assess student Spanish and English writing in Grades 4 and 5. For the purposes of the present study, it is notable that the writing samples were analyzed and scored ho listically on a 7 point scale using a rubric created by the participating school district. This type of scoring was selected as a means to involve teachers in professional development opportunities in which they would learn to apply the scoring system, ass writing, and determine inter rater reliability. English writing was Spanish interference, particularly in the areas of syntax, spelling, and 2006, p. 2344). Writing samples developed by the students imposition of Spanish vowels and other orthographic/phonetic influences on English spellings (e.g. ticher for t eacher attencion for attention ) and influence of Spanish syntax
33 on English construction (e.g., she get mad with us as a translation of ella se enjoja con nosotros omission of subject: because [ she ] is inteligent ). In addition, and in contrast to teachers (2006) also found that the teachers generally shared a belief that literacy education provided first in Spanish supported the development of English literacy. This belief was bas ed on the idea that Spanish literacy skills could transfer to English. In another study that explored the notion of cross language transfer, Lanauze and interdependence hypothesi s, which posited that academic language and skills in L1 influence and support these in L2. These authors used a picture description task to compare Spanish and English writing of bilingual children who had been rated as either achers in both languages. Participants attended grades 4 and rubric (and most of the other studies reviewed here), Lanauze and Snow (1989) applied various, fine grai several indicators in three overall areas: 1) linguistic complexity, which encompassed 9 indicators, such as number of words, number of T units, mean length of T unit, and number of noun/ verb phrases; 2) linguistic variety, which included type token ratios involving number of different verbs used and number of different colors mentioned; and, 3) semantic content, such as number of color words used and distribution of T units into categorie s, including general description, specific description, positional statement, and action statement.
34 Overall, the authors found poor quality writing across groups, including those students rated as good in both Spanish and English. However, this may have b een due to the elicitation procedure, which did not provide a highly engaging or personally meaningful task for the students. Students were provided with a color picture of a beach auze & Snow, 1989, p. 326). This activity took place in the course of the normal school day and was presented as an academic exercise. Students did not discuss the pictures or engage in other structured prewriting activities. Regarding language transfer, t he performance of students deemed as poor or good in both languages did not show significant correlations across languages. However, the performance of children rated poor in English but good in Spanish did show significant correlations on measures of ling uistic complexity, variety, and content. Lanauze and Snow (1989) interpreted this finding as evidence that children with good L1 skills could rely on these skills in their early acquisition of L2 writing. Similarly, the lack of correlations for students ra ted as good in both languages may imply that better developed skills in both languages become more independent of one another. On the other hand, it can be inferred that students rated as poor in both languages had not developed sufficient writing skills i n L1 to apply to their writing development in L2 effectively. which indicates that, in order to transfer academic language skills from L1 to L2, children need to have reached a certain level of competence in L1. Also in support of this notion, Escamilla (2006) discovered through her interviews with teachers of ELLs that they
35 h literacy without having had enough time to fully develop their instruction in L1 would provide ELL students a threshold in Spanish literacy upon which to base s ubsequent English literacy instruction. It is notable that the teachers participating in the Escamilla (2006) study worked in a district whose English Language Acquisition (ELA) program followed an early exit transitional bilingual model: ELL students were given literacy instruction in Spanish from Grades K 3, along with ESL instruction. By Grade 4, students were transitioned into all English instruction. Hence, even within a program model attempting to provide a foundation in first language literacy, these English and Spanish. The ELA program model just mentioned is based on the premise that L1 language and literacy skills can serve as a foundation upon which to build liter acy in L2. The research reviewed on cross language transfer has shown that, although L1 proficiency Moll et al. (2001) carried out a qualitative, multiple case study that compared incipient (emerging) versus instructed (more practiced) bilingual writing. The aim was to discover how bilingual children used their languages and cultural resources as tools f or biliteracy learning and practices in the classroom. These authors incorporated observation, field notes, and writing samples of two children in kindergarten (incipient biliterates) and one
36 student in grade 3 (instructed bilterate). Students attended a m agnet school with a diverse population and bilingual classes that emphasized cooperative, student driven, participatory learning. Based on their observations of the kindergarten children, Moll and colleagues determined that the two students understood the idea of a correspondence between oral words and written symbols; however, like the participants in the Rubin and Carlan (2005) stu dy, the two children wrote the same symbols for English and Spanish. For Moll and colleagues, this served as evidence that initial hypotheses about biliteracy are semantically driven; in other words, if it means the same in both languages, the written symbols must be the same. In contrast, Moll et al. (2001) reported that the student in grade 3 used biliteracy consciously as an academic resource. In addition, this participant was able to transfer writing abilities across languages, for examp le, application of their knowledge of genres in writing. In an additional research question, Moll et al. (2001) addressed how it might be possible to create additive bilingual conditions in a classroom so that Spanish could achieve a language status equiva Spanish, along with English, is an unmarked language; students could use either one or both to do their academic work and to obtain support to develop their These authors criticized the dogmatic view of knowledge in schooling, which often favors the dominant monolingual English culture and creates subtractive biliteracy environments in most ELL programs (Gutirrez et al., 2002; Nieto, 2002; Norton Peirce, 1995, Prez, 2004). For Moll and colleagues, schools will only be able to perpetuate an additive bilingual
37 and practices in support of the [bilingual] program, including developing the bilingual potential Moll et al., 2001, p. 446). The study by Moll and colleagues (2001) adheres to the same general conceptual framework as the present investigation. However, the Moll et al. project focused on emerging biliteracy skills of students i n the early grades. As evidenced by the literature review, this is the case for the majority of research performed in the area of ELL writing. In addition, Moll and colleagues did not systematically apply any formal linguistic measures to the analysis of t level about lexical, syntactic, and discursive levels of ELL writing proficiency. Assessment Measures for Bilingual Writing Examination of syntactic structures. Few studies have provided systematic analyses of the linguistic features of ELL writing. In two methodologically similar investigations, Kameen (1979) and Perkins (1980) addressed various syntactic structures in the second language writing of university students. Kameen (1979) examined the ELL university students. A total of 50 expository compositions written by ELL studen ts from multiple countries were assessed; 25 of these students had been rated as good writers and 25 as poor writers by their instructors. The measures applied, which broke down to 40 factors, included: 1) number and length of T unit; 2) number and length of clauses; 3) types of clauses (adverbial, noun, relative, coordinates); and 4) use of the
38 unit length, clause length, and incidence of the passive voice. Kameen noted that, while use of subordinate clauses did not distinguish good writers from poor ones, an increased number of words in a clause (rather than increased number of clauses) allowed good writers to produce longer T units. Kameen (1979) concluded that ELL stud ents could benefit from Perkins (1980) also investigated the ability of numerou s structures to predict the holistic writing scores of 29 advanced level, university ELL students on exposito ry texts. In this case, measures used included: 1) number of words, sentences, and T units per composition; 2) error free T units per composition; 3) number of words in error free T units per composition; 4) number of errors; 5) T unit length; and 6) two syntactic complexity indices. Predictors of the holistic writing scores were error free T units per composition, number of words in error free T uni ts per composition, errors per T unit, and total errors were significant predictors of the holistic writing scores. All other measures, including the syntactic complexity indices, were not significant. Perkins (1980) h do not take the absence of errors into account are of no use in discriminating among holistic evaluations at one advanced level of In contrast to the previously discussed studies, Harley and King (1989) explored verb usage in narra tive and expository (letters) texts written by 69 native English speakers in grade 6 who were learning French at an immersion school in Canada as
39 compared to 22 native French speaking peers. These authors employed a mixed methods design with quantitative a nalysis of the verbs used in student texts and qualitative analysis of the immersion classroom environment. Verb measures applied included: 1) lexical error rate (errors in verbs divided by total number of verbs); 2) lexical variety (type token ratio for v erbs); 3) lexical specificity (use of more frequent verbs); and 4) lexical sophistication (use of more infrequent verbs). The authors relied on three published frequency lists, including both an oral and a written corpus, to determine the frequency ratings of the verbs. It was found that the French immersion students, writing in their L2, relied more heavily on high frequency/utility verbs than did their native speaker peers. Derived verbs 2 which are less frequent, were less likely to be used by the secon d language learners. Harley and King also found evidence that the French learners made use of L1 L2 lexical similarities (i.e. cognates) in their writing. These results, along with classroom lexical variety of L2 learners through instruction of less frequent vocabulary items in context. An additional resources. Comparisons across languages, genres, and modalit ies. More recently, Berman and colleagues (Berman, 2008; Berman & Nir Sagiv, 2007; Berman & Verhoeven, 2002; 2 Harley and King (1989) described two types of derived verbs in French (parallel forms exist in Spanish and English): 1) verbs derived from other verbs by affixation (e.g. appear/disappear, take/retake); and 2) verbs derived from nouns or adjectives, with or without affixation (e.g. circle/encircle, flat/flatten).
40 Ravid, 2006; Ravid & Berman, 2006;) applied various, innovative linguistic measures to a large, international database of oral and written, narrati ve and expository, texts gathered from children and adults, monolingual speakers of seven different languages. Independent variables across this series of studies included language, age level (students in grade 4, grade 7, grade 11, and adult university gr aduate students), genre (narrative/expository), and modality (oral/written) (Berman & Verhoeven, 2002). The new linguistic measures applied to this data set included a noun scale (Ravid, 2006). This scale, adapted for use in the present study, rated nouns from 1 10 based on a noun usage became more abstract with age, especially in adolescence. Additionally, written expository texts were most likely to contain abstrac t nouns across the age levels when compared with spoken expository texts and spoken and written narrative texts. A second measure examined information density of the same corpus of texts (Ravid & Berman, 2006). This measure explored the use of narrative c ontent (eventive, descriptive, and interpretive) versus ancillary material in spoken and written narratives. Analysis found that increased narrative information was dependent on modality as spoken texts contained more ancillary material than written ones. The third measure involved clause packaging (Berman, 2008 ; Berman & Nir Sagiv, 2005 ). This measure compared oral and written expository texts and defined a clause package as units of text linked by syntactic or semantic relations. This measure found develo pmental differences again, particularly from adolescence on, for number of
41 clauses linked in a package as well as types of linkages used. Additionally, modality was a factor as more clauses were packaged together in oral than written texts. Finally, an innovative discourse measure (Berman & Nir Sagiv, 2007) was applied to the written texts. This measure explored the global organization of narratives versus expository written compositions. More sophisticated grammar and vocabulary were found in expository texts across all age groups. However, at the global level, the expository genre was not well developed until adolescence, while the narrative text structure was evidenced by children in the elementary school grades. These numerous studies confirmed that t ext genre, as well as modality, is a factor in the development and usage of various linguistic structures. As might be expected, the written modality offers more opportunities for sophistication than the oral modality, as does the expository genre as oppos ed to the narrative genre. Specifically, the expository genre (especially in its written form) is acquired later in development; however, at that point, it is more likely to contain abstract nouns and more complex syntactic structures. With respect to th e relationship between syntax and genre in another study of monolingual students, Beers and Nagy (2009) examined how classic measures of syntactic complexity, such as those utilized by Hunt (1965), functioned as predictors of uality. The participants included 41 English speaking students attending grades 7 8 at a suburban middle school. The participants wrote two persuasive essays (expository genre) and completed one previously started story (narrative genre).
42 The measures ap plied included number of clauses per T unit, number of words per clause, and number of words per T unit, along with a holistic score of text quality developed by the authors. Results of bivariate correlational analyses found that, for the expository genre, words per clause was positively correlated with text quality, while clauses per T unit was negatively correlated. In contrast, clauses per T unit was positively correlated with text quality in narrative texts. Additionally, it was notable that words per c lause and clauses per T unit were negatively correlated, and words per text was not significantly correlated with text quality. These findings support the previously described results of Berman and colleagues (summarized in Berman, 2008) and confirm the B eers and Nagy (2009) hypothesis that the relationships between syntactic complexity and text quality are influenced by text genre. It appeared that expository texts benefited from more efficient sentences with fewer clauses that made use of literate constr uctions such as adjectives and gerundial and infinitival phrases. On the other hand, narratives with longer, multiple clause constructions (e.g. descriptive, relative or adverbial clauses) were rated as higher quality. Of course, this study was not witho ut limitations. Similar to other studies of writing described here, Beers and Nagy (2009) allowed the participants a short time to develop their written texts (10 minutes). Also, the students did not physically compose at all, but rather dictated their tex ts to typists. While the authors argued that the students were trained in this procedure prior to data collection, it is possible that this format, as well as the very short composition time, did not provide a writing environment that would maximize the st
43 Summary and Conclusions This review has provided an overview of bilingual writing research explored through four categories: bilingual writing development influence of instruction on ELL writing first and second language interaction in writing, and assessment measures for (bilingual) writing In general, it can be said that a sociocultural perspective of writing served as the framework for a majority of these studies; in other words, the investigations reviewed considered not only the c ognitive processes of writing itself, but also the broader sociocultural contexts in which writing occurs. This perspective aligns with the conceptual frame of the present study, which explores both quantitative, linguistic features of the ELL participants as well as qualitative, experience and identity In the area of bilingual writing development, research has shown both similarities and differences between monolingual and bilingual literacy acquisit ion. For example, although bilingual children learning to write may advance along similar stages as their monolingual peers, bilinguals may differ with respect to vocabulary knowledge and types of errors that appear in their writing (Bialystok, 2007; Rubin & Carlan, 2005). In addition, emerging biliterate children may apply certain L1 strategies and skills to writing in L2 (Edelsky, 1982). It is important to note that studies in this area were descriptive in nature, with findings based on broad, holistic ex Therefore, this body of research would benefit from additional studies that apply fine grained analyses of bilingual writing systematically across languages. In addition, the research reviewed in this section cen tered on relatively young children (under age 10).
44 Writing development is a continuous and evolving process. More research in the area of adolescent writing is needed to better understand how older bilingual students advance in specific areas of their bili teracy development. In relation to the topic of writing instruction for ELLs, research highlighted the need to adapt effective literacy instruction strategies to meet the unique needs of ELL students, maintain challenging curricula, and hold high expectati ons for ELLs. Many teaching strategies have been found to support biliteracy development, including the use of structured writing practice and direct instruction of specific reading and writing elements. Results on the incorporation of culturally relevant literature in ELL classrooms have been mixed, and methodological issues associated with the (especially, qualitative) research in this area have made it difficult to make strong claims (Goldenberg et al., 2006). In spite of methodological shortcomings, th is research has provided some indication that the use of sociocultural strategies may be effective in increasing ELL Dueas, 2004). There is also evidence that bilingual classrooms that provide an equit able, additive, language learning context in which students have opportunities to interact with both L1 and L2 have the potential to enhance learning outcomes for ELL students (Rubin & Carlan, 2005; Rueda et al., 2006). However, successful incorporation of instruction may be much more easily said than done as ELL teachers often do not have the linguistic knowledge or confidence to include languages other than English effectively in their multilingual classrooms (Snow & Kim, 2007).
45 Regarding the area of L1 and L2 interaction in writing, there is evidence that bilinguals possess both shared and independent language mechanisms and structures. Additionally, shared linguistic structures may provide common underlying proficiencie s (Cummins, 2000) that can transfer across languages. These include phonological awareness (Bialystok, 2007), vocabulary depth (Ordoez et al., 2002), and knowledge of text level conventions (e.g. genre related text structure) in writing (Durgunoglu, 2002) On the other hand, syntax has been viewed as a language specific element (Francis, 2006). The present study integrates the concept of common underlying proficiencies as it g across Spanish and English. It is also of note tha t, with the exception of a few early studies (Kameen, 1979; Lanauze & Snow, 1989; Perkins, 1980), research in this area has generally explored bilingual writing in a holistic manner without the systemat ic application of micro analyses of linguistic features at the lexical, syntactic, or discourse levels. Some investigations that have explored these types of linguistic measures include research in both bilingual and monolingual writing. These studies have considered: 1) lexical sophistication (e.g. verb usage, Harley & King, 1989; noun scale, Ravid, 2006); 2) syntactic structures (e.g. clause and T unit measures, Beers & Nagy, 2009; Kameen, 1979; Perkins, 1980); and 3) discourse as it relates to overall text structure (information density, Ravid & Berman, 2006). As a group, the results of these investigations highlighted differences between text genres in that different measures predicted overall text quality for narrative versus expository texts. Similar ly, it was shown that written
46 expository texts generally evidenced more complex lexical and syntactic structures. The current study further explores genre differences in writing by contrasting lexical, syntactic, and discourse features across expository an d narrative genres. It is clear that more research is needed to better understand how ELLs acquire English writing and how the linguistic and cultural resources that these students bring to school can play a role in this process. This need is particularly great in the case of older ELL students who, due to more years of schooling in L1, may have a better foundation of academic language and literacy upon which to anchor their second language writing development than children in the elementary grades In addition, more specific detail regarding lexical, syntactic, and discourse features of biliterate writing would contribute to our understanding of how these features interact or do not interact across languages and genres. The present mixed methods investigation provides an in de pth, quantitative exploration of the linguistic features of Spanish and English writing, across expository and narrative written texts produced by ELLs in middle school. Additionally, with the goal of better understanding the variance in their writing perf ormance, this study provides a qualitative profile analysis of six individual students. These profiles begin to shed light on how the language and literacy learning experiences and behaviors of ELL students shape their attitudes and abilities as emerging b ilingual writers.
47 Research Questions The study addressed three quantitative research questions: 1. (Spanish/English) and genres (expository/narrative) as assessed by a noun tiers measure (adapted from Ravid, 2006) and number of different words? 2. and genres as evaluated by a clausal complexity measure (adapted from Ravid & Berman, 2006) and mean length o f clause? 3. How do discourse features of the p articipant s writing compare across languages and genres as examined by the and Narrative Writing Skills (Quellmalz & Burry, 1983)? Finally, a sociocultural que stion directed the qualitative portion of the study: 4. How do previous and current language and literacy learning experiences and/or attitudes, beliefs, and feelings about L2 writing?
48 CHAPTER 2 Method Overview of the Research Design The purpose of the present study was to obtain a deeper understanding of ELL writing at the middle school level. The quantitative focus concerned a comprehensive assessment of the linguistic f written in both Spanish and English. An additional aim of this investigation was to develop through qualitative methods an understanding of how the identities of the participating middle school EL Ls as writers have been shaped by their bilingual language and literacy learning experiences. Because of this dual purpose, and in keeping with a sociocultural framework, a mixed methods desi gn was selected to address the research questions. Mixed Methods Designs Creswell and Plano Clark (2007) defined mixed methods research as involving a (p. 4), and methods --that unite q uantitative and qualitative approaches throughout the research process. For these authors and qualitative approaches in combination provides for a better understanding of resea rch with this statement, Onwuegbuzie and Leech (2004) highlighted three benefits of the implementation of a mixed methods research paradigm. These include: 1) the opportun ity
49 to investigate phenomena in a flexible and holistic manner; 2) the ability to research both micro and macro aspects of a setting or phenomenon; and 3) the ability to validate qualitative data analysis with quantitative analysis and vice versa. In addit ion, according to the fundamental principle of mixed methods research, the combination of qualitative and quantitative methods provides the researcher with complementary strengths and non overlapping weaknesses (Johnson & Turner, 2003). For example, an ex perimental, quantitative study allows for large scale analysis of a behavior or phenomenon in which resulting findings may be generalized to the broader population, but will not provide detailed specifics regarding the individuals in this population. On th e other hand, a qualitative investigation offers an in depth, holistic examination of a phenomenon from the perspective of a unique individual or group. The ability to explore an issue with both methods affords the opportunity to achieve both a big picture understanding as well as more detailed view of the nuanced complexities of the phenomenon in question. In a mixed methods study, quantitative and qualitative methods may be mixed within and across stages of the research process. These stages broadly inc lude definition of the research objective, data collection, and data analysis and interpretation (Johnson & Onwuegbuzie, 2004). Onwuegbuzie and Teddlie (2003) further described mixed methods designs as occurring either in parallel/simultaneous or sequentia l fashion. The parallel and/or sequential use of qualitative and quantitative methods may also differ at the different stages of the research process. More specifically, Creswell and Plano Clark (2007) identified four major types of mixed methods designs, which also incorporate the
50 notion of simultaneous vs. sequential timing: 1) the triangulation design, a single phase (simultaneous) design in which the researcher merges quantitative and qualitative data sets in an integrated analysis; 2) the embedded desi gn, in which one set of data assumes a secondary or supportive role to the other set (timing may be either simultaneous or sequential); 3) the explanatory design, a two phased (sequential) structure in which qualitative data serve to explain or extend init ial quantitative results; and 4) the exploratory design, a two phase (sequential) model that begins qualitatively with the purpose of developing or testing a quantitative instrument. Based on these classifications, the current study can be defined as a tw o phased, embedded, mixed methods design, the second type that Creswell and Plano Clark (2007) identified. The design is two phased because, although the collection of quantitative and qualitative data occurred simultaneously, the data were analyzed sequen tially. That is, the quantitative analysis of the formal writing samples (N = 148) obtained from all of the participants (N =20) occurred first, followed by the qualitative analysis of the journal entries and interviews obtained from 6 focal participants. In addition, this investigation represented an embedded design due to the nature of the research questions, which distinguished between a primary purpose (addressed by a quantitative analysis of the rting, secondary purpose interviews.)
51 Informal Interviews of Teachers of English as a Second Language (ESL) Several decisions regarding the study design, participant selection, and data collection were determined based on informal interviews with three ESL teachers between July and September 2007. These interviews occurred as part of what Stake (1995) considered Conversations with the ESL teachers took place on the phon e and lasted approximately 30 45 minutes with each. Topics addressed included demographic classrooms, how their students developed English writing proficiency, and their recommendations for the study focus and data collection procedures. The teachers, who were all female, were selected on the basis that a large percentage of their students were Spanish speakers. A total of two ESL educators at the middle school leve l and one at the high school level participated. Each teacher instructe d EL L students of all grade levels and degrees of proficiency at her respective school. The ESL teacher interviews provided important insights regarding the background experiences and a bilities of potential participants, writing instruction in the ESL classroom, and feasibility of the investigation. One of the teachers interviewed, Ms. Brady 3 became the source of participants for the study because her classroom was 3 All name s of teachers, students, and the school have been changed to protect the confidentiality of participants.
52 selected as the study site. (See Appendix A for a summary of findings from these informal ESL teacher interviews.) Participants A major challenge for research involving bilingual populations is the difficulty in recruiting a homogenous sample with respect to language proficien cy, the frequency and conditions of language use, and social, educational, and linguistic experiences (Grosjean, 1998). Indeed, it was clear from the ESL teacher interviews and initial conversations with Ms. Brady that the ELLs attending local public schoo ls demonstrated extreme variability with regard to Spanish and English language abilities and behaviors, prior educational experiences, cultural and family backgrounds, etc. Because of this variability, it was not possible to recruit a large, homogenous gr oup of participants with the resources available. This being said, participant selection for this investigation occurred through what Maxwell (2005) termed purposeful selection (also known as purposeful sampling or criterion based selection). According to Maxwell, four possible goals exist for purposeful selection: 1) to achieve a representative or typical sample; 2) to capture the maximum variation in the sample; 3) to examine critical cases as they relate to theories being explored; or, 4) to compare or highlight differences among individuals or settings Participants for Quantitative Analysis The final group of participants in this study included 20 students attending middle school (grades 6 8; ages 11 14 years). All of these students we re Spanish English bilinguals whose families originated from Mexico and the Caribbean (Puerto Rico and
53 the Dominican Republic). Although all of the participants spoke Spanish at home, six of them were born in the United States. A ll of the partic ipants expe rienced varying years of schooling in the United States and collection, the participants attended the same middle school; therefore, they shared the same ESL teacher (Ms. Brady), ESL classroom, and ove rall school culture. All participants attended ESL classes at their school, Bayview Middle School, a public middle school on the West Coast of Florida. At the time of data collection, Bayview Middle had a diverse population of approximately 600 students. The of Mexican descent), 23% African American 6% Multiracial, and 3% Asian. It is clear from this distribution that the majority of ELL students at the school were from Hispanic backgrounds. Additionally, a pproximately 50% of students received free or reduced price lunch. To serve the needs of its second language learners, Bayview had one ESL class for the ELL students in grade 6 and another class that combine d the ELL students in grades 7 8. Each class met daily for two periods of 50 minutes each. Although as noted, Spanish speakers comprised the majority of the ELL population at Bayview, the rope, Southeast Asia, India, Africa, and the Middle East. The teacher, Ms. Brady, was an English Spanish bilingual of Puerto Rican heritage and the only ESL teacher on staff at Bayview at the time. Ms. Santos, a bilingual aide from Puerto Rico, assisted Ms Brady in the classroom.
54 At this point it should be mentioned that the researcher is also a fluent English Spanish bilingual. A total of six inclusion criteria were applied for participant selection: 1. Participants were Spanish speakers whose families ori ginated from Mexico or the report and teacher report, and confirmed by participant questionnaires. 2. Participants had received up to grade level instruction in their home country or the U.S. as confirmed through sc hool records and participant questionnaires. 3. Participants were not from migrant families, as determined by teacher report and consistent school attendance throughout the 2007 2008 academic year. 4. Participants qualified for ESL services for the 2007 2008 s chool year, as established by enrollment in ESL and qualifying test scores on the Language Assessment Scales (LAS R/W CTB/McGraw Hill 1988 ) from F all 2007. 5. Participants were able to write in English and Spanish, confirmed by evidence of student writing and ESL teacher report. 6. Participants had no previous diagnosis or record of disability or special education services including speech/language, behavioral services, etc., as documented by school records. The ESL teacher initially identified 29 students i n grades 6 8 who met the inclusion criteria. These students were invited to participate in the study and parental consent and student assent forms were distributed. The forms were si gned and returned for 24 of these students. After data collection for the original 24 participants was
55 complete, four of the male students, two in grade 6 and two in grade 8, were eliminated from the study because their writing did not meet the criteria for analysis. Of the four, two of these students could not write in Spanish one could not write in English, and one produced texts that consistently did not meet the productivity criteria for analysis (at least 10 T units and/or 75 words). Therefore, the writing samples of a final group of 20 students were included in the quanti tative analysis. Characteristics of t hese 20 participants are summarized in Table 1. (Additional information about the students, obtained from a participant questionnaire, is provided in Appendix B.) Table 1 Participants by Gender, Grade Level, Family Ori gin, and Place of Birth Grade level Gender Family from Mexico Family from Puerto Rico Family from Dominican Rep. Totals Born Mex Born U.S. Born P.R. Born U.S. Born D.R. 6 Female 2 0 0 0 0 2 Male 2 3 0 1 0 6 7 Female 1 1 1 0 0 3 Male 1 0 0 0 0 1 8 Female 2 1 0 0 1 4 Male 2 0 2 0 0 4 Totals 1 0 5 3 1 1 20
56 Focal Participants for Qualitative Analysis As previously stated, the present study followed a mixed methods methodology with an embedded design in which quantitative and qualitative data were collected simultaneously and analyzed sequentially (Creswell & Plano Clark, 2007). In line with this approach, during the data collection phase a subgroup of 6 focal participants was randomly selected for the qualitativ e data collection and analy sis from the original sample of 24 students. After all of the writing samples had been collected, the 6 focal participants were interviewed and their interview transcripts and journal entries were transcribed for the qualitative data analysis. One student in the focal group, Manuel, was one of the four participants dropped from the quantitative analysis because four of his eight formal writing samples did not meet the productivity criteria (at least 10 T units and/or 75 words). However, because he was a fo cal participant, his writing was coded and scored for the quantitative measures, but was not included in the statistical analyses. Table 2 provides demographic information about the focal participants. (The focal participants are described in detail in the qualitative results section.)
57 Table 2 The Six Focal Participants for the Qualitative Analysis Name Grade Age Place of Birth Schooling Outside the U.S. Schooling in the U.S. Diego 6 12 Florida Grades K 4 in Mexico Grades 5 6 Carolina 7 13 Puerto Rico Grades 2 6 in Puerto Rico Grades K 2 7 Edgar 7 13 Mexico Grades 1 5 in Mexico Grades 5 7 Sara 8 13 Mexico Grades 1 5 in Mexico Grades 6 8 Manuel 8 14 Mexico Grades K 5 in Mexico Grades 6 8 Juan 8 13 Puerto Rico Grades K 6 in Puerto Rico Grades 7 8 Data Sources and Collection Procedures One of the benefits of a mixed methods design is the ability to triangulate findings from multiple sources of evidence. For Yin (2003), the key advantage of triangulation of data sources is the multiple sources of evidence overlap to confirm findings. This study involved four data sources: 1) for all participants, eight formal writing samples, controlled for genre (expository/ narrative) and language of text (Spanish/English); 2) for all participants, 10 Grades are repeated to indicate when a student spent part of a school year in her or his home country and part of the year attending school in the U.S.
58 second language acquisition; and 4) for the 6 focal participants, an interview that provided a more in anguage experiences as well as attitudes and feelings toward these experiences and bil ingualism. All data were gathered during the spring semester of 2008. Each of these data sources and procedures for their collection are described in detail next. Participant Writing For the purposes of the present investigation, the primary source of data was the examine several aspects of academic language proficiency. The content of the focal partici to explore consistency and variability within and across the writing abilities of the participants, 8 formal samples and 10 journal entries were elicited from each student over a period of one month, from April May 2008. All writing samples were elicited through carefully constructed writing prompts provided to the students in English and Spanish. Participants produced two expository texts each in Spanish and English, as well as two narrative texts each in both languages for a total of 8 texts each or 160 total for all 20 participants. These formal samples were controlled for genre and language, and prompts were repeated so that students wrote on the same topic in both lan guages. The topics for the formal writing samples are outlined in Table 3.
59 Table 3 Topics for Formal Writing Samples (N =160) Formal writing sample Topic Expository 1 Family: A person I admire Expository 2 School: Letter to a new student Narrative 1 Family: Special or funny family memory Narrative 2 School: First day of school in the U.S. Students also created personal journals and developed 10 journal entries in their language of choice. Journal prompts, which were unique to each writing event and not repeated, were balanced for expository and narrative genres. Because journal entries were not controlled for language, they were not included in the quantitative analysis. However, the journals written by the focal participants (N = 60 texts ) were included in the qualitative analysis, which focused on the content r than linguistic factors (journals written by the remaining participants may be utilized for future analysis on this group of students). (The complete writing prompts for all formal samples and journal entries are provided in Appendix C.) All writing too k place in the context of ESL classroom instruction. The data bilingual autobiography project in which all the ELL students at Bayview Middle -not just those participat ing in this research study -took part (although all the ELL students
60 by the researcher for analysis). The bilingual autobiography project served to engage the stude nts, provided a goal and purpose for writing, and united all written texts under a common theme. The researcher and ESL teacher collaborated to facilitate the bilingual autobiography project and elicit the writing samples from the class. In addition, befor e data collection began, students constructed and decorated their journals and worked collaboratively to brainstorm topics for the autobiographies (see Appendix D). After the data collection was complete, with the support of Ms. Brady and her aide, the ELL students compiled, typed up, and edited their work, and presented the finished bilingual autobiographies to the school community at a book signing event. This culminating event ir identities as bilingual writers. One 50 minute class period was devoted to the production of each writing sample. In each session, the students were greeted individually upon their arrival and the researcher chatted with them as they found their seats a nd procured the necessary leaf paper and pencils). At this time, the ESL teacher also welcomed the students and announced any news or requests (e.g. reminder to return permission forms for a field tr ip). Once the students had settled into their seats and were prepared to work, the researcher presented the prompt, in both English and Spanish, projected on a screen at the front of the classroom. Student volunteers read the prompt aloud in both language s. The researcher then read it again for clarification, stated the genre (expository or narrative),
61 and noted the language (home language or English) required of that particular writing sample. The researcher and ESL teacher then took a few minutes to info rmally discuss the topic with the class, providing personal examples and encouraging students to make suggestions as to what they could write about. In addition, students were reminded to include not only eventive information (telling what happened, a chai n of events), but also descriptive (writing about the state of affairs, facts, or background information) and g was scaffolded to some degree by After presenting and discussing the prompt for approximately 10 minutes, the students were given 30 minutes to address the prompt in writing. During this time, classical musi c was played in the classroom and the researcher, teacher, and aide circled the room to encourage and monitor students. Students wrote by hand and worked independently, but were permitted to sit anywhere in the room -including on cushions on the floor, wri ting on clipboards -as long as they maintained focus on their work and did not distract others. As the students completed their texts, the researcher reviewed them to ensure that all aspects of the prompt were addressed. If they were not, students were enc ouraged to continue writing. At the close of each 30 minute writing period, the Word documents by the primary researcher and a research assistant. (See Coding and Scoring section below.)
62 In general, the teacher and researcher promoted productivity in student writing during the data collection phase in five ways. These included: 1) using clear and specific, bilingual prompts and instructions that outlined the multiple compo nents to be addressed in each writing sample; 2) discussing the prompts with the students and orally providing personal examples before writing as a warm up; 3) reminding students to use details, iting; 4) creating a calm, comfortable atmosphere with music an d flexible seating options; and 5) reviewing completeness. For the formal expository and narrative samples, student s were instructed to write either in English or Spanish Language of writing was alternated each week so that the students wrote the same topic in both languages with one week in between sessions (see Table 4 for the schedule of writing samples). On the se cond week of any given formal sample, students were permitted to review what they had written the previous week for approximately 5 minutes before writing the alternate language version. After briefly reviewing their previous texts, students returned their writing to the teacher and the researcher. When writing the alternate language version of a given prompt, students were
63 Table 4 Schedule of Data Collection : Participant Writing Monday Tuesday Thursday Friday Week 1 Journal 1 Narrative 1, English Expository 1, Spanish Journal 2 Week 2 Journal 3 Narrative 1, Spanish Expository 1, English Journal 4 Week 3 Journal 5 Narrative 2, Spanish Expository 2, English Journal 6 Week 4 Journal 7 Narrative 2, English Expository 2, Spanish Journal 8 Week 5 Journal 9 Journal 10 Additional Data Sources applied to collect background information about all of the participants, as well as to provide additional data for the qualitative analysis of the focal participants. These instruments included a participant questionnaire, and interviews of the focal participants 4 Participant questio nnaire. All participants completed a written questionnaire that was provided in both English and Spanish. The purpose of the questionnaire was to collect information regarding their family, heritage, place of birth, and places of schooling. Students comple ted the questionnaire during one class period after all writing samples had been completed. (The participant questionnaire, in English and Spanish, is provided in Appendix E.) 4 The ESL teacher, Ms. Brady, was also interviewed at the time of data collection. However, because the
64 Participant interviews. Interviews were conducted with each of the 6, randomly selected focal participants. All interviews were audio recorded with Audacity software (general public license) with permission of the students (and written consent of their parents). The interviews we re semi structured in nature in that they were organized around a pre designed interview guide (Bernard, 2002). The interview guide, designed for the purposes of the present study, included open ended, descriptive questions to be asked of each student. How ever, when conducting the interviews, the researcher assumed a flexible approach to the questioning order and incorporated additional, relevant interview instruments ut ilized here can be described as semi standardized and nonscheduled. The student interview was developed to provide insight into the qualitative research question that inquired about how previous and current language and literacy learning experiences and p and literacy learning experiences, language usage, and language and literacy practices for both Spani sh and English in the home, community, and school contexts. The interview language learning experiences and bilingualism in general. Interviews were conducted ind ividually with each focal participant during the ESL class period and after all writing samples had been collected. Interviews were approximately 20 minutes in length. It is important to note that, although interview
65 questions are presented here in English (see Appendix F for the student interview guide), the researcher immediately established a bilingual language mode (Grosjean, 1998) at the onset of each student interview, making it clear that students were welcome to respond in English or Spanish, and th at code switching was also permissible. This was accomplished by introducing and explaining the interview process in both languages and asking the participant which language s/he preferred to speak during the interview. Additionally, by the time the interv bilingual status and were comfortable speaking with her in either language. Interestingly, all 6 focal participants elected to conduct their interviews in Spanish. Analysis According t o Onwuegbuzie and Teddlie (2003), mixed methods data analysis enhances representation (ability to extract adequate information from data) by providing more opportunities to discover or construct meaning from the data. In addition, mixed methods designs inc rease legitimation (validity) by taking advantage of the strengths of both qualitative and quantitative analysis. For example, the researcher has the opportunity analyses can be applied (Onwuegbuzie & Teddlie, 2003). qualitative analyses were carried out on t he data. For the quantitative analysis, the 160 formal samples written by the 20 participants were evaluated at the lexical, syntactic, and discourse levels, and scores were compared within subjects across genre
66 (expository/narrative) and text language (Sp anish/English) of the text. For the qualitative portion of the study, interviews and journal entries of the 6 focal participants were analyzed using domain and taxonomic analyses to determine domains (categories), their included terms (members in each cate gory), and the relationships among them (Spradley, 1979). Both the quantitative and qualitative analyses are described in detail in the following section. Table 5 reviews the research questions, assessment measures applied, and data sources. Table 5 Resear ch Questions, Assessment Measures, and Data Sources Research question Assessment measure applied Data source 1. How do lexical features of compare across languages and genres? Noun tiers, adapted from analysis; number of different words (NDW) Formal writing samples, expository and narrative, in Spanish and English 2. How do syntactic features of compare across languages and genres? Clausal complexity measure, adapted from Berman & Nir Sagiv (2007); mean length of T unit (MLT) 3. How do discou rse features of the participant s writing compare across languages and genres? Analytic Scales for Assessing Narrative Writing Skills (Quellmalz & Burry, 1983).
67 Research question Assessment measure applied Data source 4. How do previous and current language and literacy learning experiences and/or practices influence the bilingual writers, including their attitudes, beliefs, and feelings about second language writing? Domain and taxonomic analysis (Spradley, 1979) Interviews and journal entries of the focal participants Quantitative Measures: Coding, Scoring, and Agreement and scoring for the quantitative analysis. The complete set of written samples (N = 160) were coded in the following order: 1) segmentation of the text into T units (Hunt, 19 65); 2) classification of the nouns into 10 noun scale categories (Ravid, 2006), which were later compressed into three noun tiers (see below); 3) classification of each T unit into type/s of clauses via the clausal complexity measure; and, 4) holistic and analytic scoring of the complete written texts with the and Narrative Writing Skills (Quellmalz & Burry, 1983). Coding the writing samples in this order provided opportunities to check previous coding at each new level of coding.
68 The primary researcher coded the texts written in English for all of the measures with the help of a research assistant who was a graduate student in speech language pathology and had some proficiency in Spanish. During this proc ess, both coders maintained a continuous dialogue regarding decisions about coding at all levels. The researcher reviewed texts coded by the assistant. The researcher also coded all texts written in Spanish. Details on coding procedures and scoring for eac h of the quantitative measures are explained in the following sections. Segmentation of Texts into T Units After the original, hand written texts had been transferred verbatim into Microsoft Word documents, each text was separated into minimal terminabl e units, known as T units. Hunt (1965) defined T (p. 21). T units are generally described in current literature as containing an independent clause and all of its subordinate clauses and modifiers (Gutirrez Clellen & Hofstetter, 1994). units for written English called for the separation of coordinate clauses into new T units only when the subject of the coordinate clause was explicit (e.g. I went to the store/ and I bought candy = 2 T units ; the / indicates a new T unit ). On the other hand, a coordinate clause with an ellipted subject is traditionally considered to be one T unit (e.g., I went to the store and bought candy = 1 T unit). In contrast to this approach a nd because participants wrote in both English and Spanish, which is a pro drop language, an alternative system was employed for the designation of coordinate clauses into T units. The criteria used by Gutirrez Clellen and
69 Hofstetter (1994) and Miller et a l. (2006) were applied. These authors designated all coordinate clauses as separate T units in both Spanish and English, whether or not the subject was explicit for segmenting Spanish T I admire my mom because she is n was coded as two T units in spite of the ellipted subject (she) in the coordinate clause. This separation is equivalent to Yo admiro a mi mam porque es buena /y (ella) trabaja mucho, Hence, both texts are separated before the coordinate conjunction, and or y This procedure allowed for consistency in the coding of T units in both languages. After the texts were divid ed into T units, they were entered into the Systematic Analysis of Language Transcripts software (SALT, University of Wisconsin, 2008) for further coding and analysis (discussed below). After the writing samples had been segmented into T units and entered into SALT, texts that did not meet productivity criteria (at least 10 T formal samples (N = 160), 12 individual texts did not meet the criteria and were eliminated, leaving 148 analyzable writ ing samples. Lexical Level: Noun Tiers and Number of Different Words Coding of noun tiers. Using SALT, the 148 writing samples that met the criteria design of this no un scale was based on previous research on the development of noun categories in children, s uch as: 1) e xploration of the acquisition of count nouns versus collective/mass nouns in young children (e.g., book, house, shirt vs. furniture, clothing,
70 stuff ); 2 authority, career, challenge ) and derived nouns (e.g., intervention, annoyance, underestimation ); and 3) use of simple/concrete to complex/abstract nouns in different types of spoken and written texts. Ravid note d that a systematic model for the development of noun categories had not yet second and third order nouns was also incorporated into the noun scale. For Lyons (1977), first order nouns i ncluded stable objects, people, and animals; second order nouns referred to processes, events, or states; and third order items were abstract nouns not defined by time or space. cla ssifying tool. As a ranking instrument, the noun scale scores nominals on a scale of 1 10, designating them as levels that range from concrete countable (level 1) to derived abstract (level 10). Additionally, the 10 levels can also be considered categories into which nouns may be classified based on their semantic pragmatic content. Ravid built the case that the noun scale dealt with universal concepts expressed as nouns, and, therefore, the categories transcended linguistic differences such as noun gender, which occurs in Spanish and not in English. Based on this argument, the noun scale measure appeared to be applicable across different languages, including English and Spanish. For the purposes of this analysis, the noun scale was used as a ranking tool, a nd SALT. After this coding was completed, the 10 levels were compressed into three categories, or tiers, similar to the procedure used by Berman & Nir Sagiv (2007), which al level noun scale into four categories. In this
71 case, the three noun tiers were created to provide fewer, but broader, categories for the purpose of statistical analysis. The nouns were classified into tiers based on the following procedure: 1) Nouns rated as categories 1 4 on the noun scale were reclassified as tier 1 nouns; 2) nouns coded as levels 5 7 on the noun scale were reclassified as tier 2 nouns; and, 3) nouns rated as categories 8 10 were reclassified as tie r 3 nouns. The noun tiers and examples of each level are summarized in Table 6.
72 Table 6 Noun tier categories Tier 1 1 Concrete, countable 2 Proper nouns 3 Collective, location 4 Role Ball, locker, cohetes (firecrackers), regalos (gifts), Ms. Brady, Ms. Santos, house, Mexico, Latinos, Busch Gardens, mom, principal, policia (police), payaso (clown) Tier 2 5 Generic 6 Temporal 7 Event Stuff, medicine, school supplies, actividades (activities), viernes (Friday), hour, year, un rato (a while), pelea (fight), soccer game, misa (mass), party, vacaciones (vacation) Tier 3 8 Imaginable abstract 9 Abstract 10 Derived abstract Report card, homework, techo (shelter), professional, money, bills, rent, love, error, reglas (rules), ganas (desire), encanto (enchantment), destreza (dexterity), alma gemela (soul mate)
73 Scoring of noun tiers. For scoring purposes, the nouns in each w riting sample were given point values based on their tier membership (tiers 1, 2, and 3 were valued at 1, 2, and 3 points, respectively). Points were then summed for each text, and the total number of points was divided by the total number of nouns in the writing sample. Resulting scores were recorded into an Excel spreadsheet for statistical analysis. Number of different words (NDW) writing samples were scored for number of different words used. NDW is a v ocabulary productivity measure commonly used in investigations of cross linguistic abilities of bilingual children (e.g., Miller et al., 2006; Paradis, Crago, Genesee, & Rice, 2003) as well as in research on monolingual children with language impairment (L I) (Fey, Catts, Proctor Williams, Tomblin, & Zhang, 2004; Hewit t, Hammer, Yont, & Tomblin, 2005 ; Nelson & Van Meter, 2007). In the present study, word root tables were provided by SALT for each writing sample These tables listed for each text all words in cluded and their frequencies. Based on these tables, the NDW was independently calculated by the researcher and research assistant by excluding repeated morphemes and totaling the remaining different words. Syntactic Level: Clausal Complexity Measure and Mean Length of T Unit Coding clausal complexity. In addition to coding and scoring at the lexical level, all texts were evaluated at the syntactic level with a clausal complexity measure. Many cross linguistic studies of language have used the clause as a unit of analysis (e.g., Berman & Nir Sagiv, 2007; Katzenberger, 2004; Gutirrez Clellen, 2002; Paradis & Crago, 2000; Ravid & Berman, 2006).
74 Ravid and Berman (2006) served as a starting point for the development of the clausal complexity measure applied to the present study. For Ravid and Berman, a clause was defined as a unit containing a subject and predicate, and text units were identified as containing less than a clause (here, a non clause), a single clause (independent clause), or a combination of c lauses (s ubordinate or coordinate clause ). Similar to the lexical coding previously described, the designation of clausal step process. Using SALT, the clauses in every T unit in each text were initially coded in a highly specific manner. F or example, subordinate clauses were categorized as nominal (subject or object), relative (differentiated for presence or absence of a relative pronoun), or adverbial clauses of various types (temporal, l ocative, causal, conditional, comparative, purpose, exception). Similarly, coordinate clauses were first classified as additive, temporal, causal, contrastive, or exclusionary c lauses (Alarcos Llorach, 1996; de l a Pea, 1999; Gili Gaya, 1972; Quirk, Greenb aum, Leech, & Svartvik, 1985). Based on this system and using SALT, the T units in each text were coded for type/s of clauses. In the second step similar to the collapse of the noun scale categories into the three tiers the numerous clause types were reclassified into broader categories for the purpose of the statistical analysis. These categories were non clause, independent clause, subordinate clause and coordinate clause. The se categories were then used to determine a clausal complexity score for e ach text. Scoring clausal complexity. For scoring purposes, each T unit was assigned a point value based on the number of embedded subordinate clauses. Non clauses were
75 valued at zero points, and independent and coordinate clauses received one point. T u nits with one subordinate clause received two points, two subordinates, 3 points; three subordinates, 4 points, and four or more subordinates, 5 points. Points were then summed for each writing sample, and the total number of points was divided by the tota l number of clauses in the text. Mean length of T unit (MLT). In addition to the clausal complexity measure, MLT for each text was determined as a measure of written syntactic productivity. The MLT for each writing sample was calculated by SALT. Discou rse Level: Analytic Scales Finally, the writing samples were scored at the discourse (text) level using the developed by the Center for the Study of Evaluation (CSE) at the Un iversity of California Los Angeles (Quellmalz & Burry, 1983; referred to hereafter as the CSE Scales). The CSE Scales were developed at a time when a common approach to writing evaluation ch as multiple choice tests that addressed text organization and sequencing and/or language usage and mechanics for written composition. Quellmalz (1979) argued for the need to assess actual written compositions with domain referenced assessments that prov ided clearly defined criteria from which to evaluate writing outcomes. For Quellmalz, a domain referenced assessment adhered to three principles: 1) ecological validity : the measure is applicable to real students i n the context of school writing ; 2) genera lizability : performance on the measure will predict performance on related tasks; and 3) diagnostic value : the measure
76 has instructional implications regarding student placement and/or teaching strategies. In addition, Quellmalz (1982) discussed the need f or an analytic (rather than holistic) writing assessment that could distinguish between mastered and non mastered skills for various elements of a text. Finally, the selected criteria should enable outcomes that provide opportunities to diagnose problems a nd implement appropriate instruction or intervention (Quellmalz, 1982). The CSE Scales (Quellmalz & Burry, 1983) were developed based on the and its assessment. As oppose d to a holistic scoring method whose outcome is a single, overall score for a given composition, the CSE Scales represented an analytic assessment holistic score for the overall quality of the composition. Rubrics describe the criteria to rate texts for each element on a scale of 1 to 6; scores of 1 3 are considered non mastered/not competent and scores of 4 6 are considered mastered/competent. This type of writin g assessment can also be described as primary trait scoring, a system in which raters are trained to evaluate various textual features based on a given set of criteria (Schriver, 1990). Schriver classified this method as a category of expert judgment focus ed evaluations of writing, and highlighted as a benefit the wealth of information that this type of measure can provide about a text. On the other hand, a disadvantage of primary trait scoring is that it is difficult for readers to reach agreement consiste of ratings has
77 However, Quellmalz and Burry (1983) attempted to minimize inter judge variability by p roviding explicit scoring criteria and training procedures for the CSE Scales. Regarding the theoretical and research base for the CSE Scales, Quellmalz and Burry (1983) distinguished between the structural development of expository and narrative writing. For these authors, expository writing was based on the logical development of ideas with explicit and specific support; therefore, the CSE Scales offered criteria for scoring each of five elements of expository compositions on the 6 point scale: 1) genera l competence or overall quality (which incorporates the scores on all other elements), 2) organization and coherence, 3) paragraph organization, 4) quality of support provided, and 5) mechanics, which includes spelling, grammar, and punctuation. In contras t to the logical development of expository writing, Quellmalz and Burry (1983) described narrative texts as a chronological development of events supported by descriptive detail. With this in mind, the CSE Scales further provided criteria to rate narrative compositions in four areas, using the same 6 point scale for each: 1) general impression or overall judgment, 2) focus/organization, 3) support, which involves use of descriptive detail, and 4) grammar/mechanics. For the present investigation, the CSE S cales were selected as the discourse level measure because they supplied similar decision rules and a scoring system that could be were scored for all elements usi ng the CSE criteria (rubrics) and the 6 point scoring system. The holistic scores (overall quality ratings), which incorporated the ratings on other elements, were utilized for statistical analyses.
78 Inter Judge Agreement According to Miles and Huberman ( is to support definitional clarity of codes and strengthen reliability of the data. For Bakeman and Gottman (1986), agreement ensures accuracy of observers and reliability of procedures. With these goals in mind 20% of the narrative and expository samples were selected at random (N = 4 participants, 32 writing samples) to be check coded for agreement by coders who did not participate in the original coding A doctoral student in Communication Sciences and Disord ers with a linguistics background check coded the writing samples produced in English, and a Spanish English bilingual clinical instructor in speech language pathology check coded the texts written in Spanish. Both check coders met with the researcher pri or to coding and were trained in how to segment the texts into T uni ts and to apply the noun tiers, clausal complexity measure, and CSE Scales In the case of the noun tiers, check coders were asked to identify and rate each noun as either a tier 1, 2, or 3. For clausal complexity, the check coders coded each T unit for the following categories: 1) non clause; 2) independent clause; 3) subordinate clause/s; and, 4) coordinate clause/s. The check s utilizing the following formula to compute a percentage of agreement (Bakeman & Gottman, 1986, p. 75): P A = N A / N A + N D 100 P A is the percentage of agreement, N A is the number of agreements, and N D is the number of disagreements.
79 Percentage of kappa, was utilized here based on the nature of the judgments required for the measures applied. For example, particularly in the cases of T units and clauses, there were not predetermined num bers of items to be coded. This situation led to both true disagreements in coding as well as disagreements based on error, i.e. omission of codes. This being said, the inter coder agreement for number of T units averaged 87% for the English texts and 82% for the Spanish texts. For noun tiers, inter coder agreement averaged 81% for texts written in English and 83% for texts written in Spanish. The averages for clausal complexity were 76% for English texts and 74% for Spanish texts. Finally, agreement percen tages on the CSE Scales were calculated based on Quellmalz mastered (scores of 1 3) versus mastered (scores of 4 6). Agreement for these ratings averaged 75% for English and 81% for Spanish texts. Miles a nd Huberman (1994), who referred to the same formula to calculate percentage of inter coder agreement for qualitative data, estimated an average agreement percentage around 70% for the first round of check coding. Based on this guideline, the inter judge a greements were interpreted as acceptable. Quantitative Measures: Statistical Analysis Because the scores from the various measures were not normally distributed, a nonparametric statistical procedure was selected for analysis. According to Siegel and Cast ellan (1988), nonparametric statistics are well suited for non normal distributions and research with small sample sizes, both of which are applicable to the present study;
80 way ANOVA by ranks (within subjects) and the correspondin g post hoc test (Siegel & Castellan, 1988) were applied for the analysis of the lexical, syntactic, and discourse level data. For each level, comparisons were made between languages within genre topics (all expository samples in both languages were compar ed; all narrative samples in both languages were compared), as well as between genre topics within languages (all texts of both genres written in Spanish were compared; all texts of both genres written in English were compared). SPSS Statistics 17.0 softwa re (SPSS, Inc., 2008) was used for the statistical analyses. Post hoc tests were conducted to determine which ranks differed significantly ( p < .05). Qualitative Analysis Overview In the present mixed methods design, the qualitative portion was subordinat e to and embedded within the quantitative investigation. In an embedded mixed methods pose of learning experiences had shaped their attitudes and identities as bilingual writers. Qualitative data were collected with the quantitative data in a simultaneous fa shion. However, qualitative data were analyzed sequentially, after the quantitative a nalysis (Creswell & Plano Clark, 2007).
81 As previously mentioned, a subgroup of 6, randomly selected, focal participants was included in the qualitative analysis. These st written transcribed for the qualitative analysis, resulting in 72 double spaced pages of interview transcripts and 30 double spaced pages of jo urnal texts. In addition, the qualitative analysis integrated to some extent the quantitative r esults of the focal participant s writing. Two levels of qualitative analysis were applied to the data: 1) a within case analysis (profile) of each focal partic ipant, and 2) a cross case analysis that explored the specific topic of attitudes toward bilingualism across the 6 focal participants. At each of support of XSight qualitat ive data analysis software (QSR International, 2006). Data displays (Miles & Huberman, 1994) were created in conjunction with these analyses for each case as well as for the cross case analysis. Domain and Taxonomic Analyses Domain analysis. The purpose of a domain analysis is to discover domains (categories) in the data as they are perceived and applied by the participants. For 100). A domain includes a co ver term (name of the category) and two or more included terms that are connected to the cover term by a single semantic relationship. Once the domains are defined, included terms are determined based on category membership. Hence, a domain analysis result s in a series of hierarchical relationships in which each
82 domain is comprised of various included terms based on semantic relationships such as attribution (X is an attribute of Y), cause effect (X is a cause/result of Y), strict inclusion (X is a kind of Y), sequence (X is a step/stage in Y), and rationale (X is a reason for doing Y) (Spradley, 1979, p. 111). In the present study, the domains were based on the topics of the journal entries and the interview questions, and thus were established apriori by the researcher (with input from the students in the case of the journal topics see Appendix D). Thus, cover terms for the domains, which were similar for all focal participants, included coming to the U.S., language learning, bilingualism, goals and wishes traditions, family and friends However, for each focal participant, included terms differed based on the content of their journal entries and interviews. For example, although two students may have shared the domain, coming to the U.S. included terms for one student may have been, among others, separation from family and friends, nervous about school, and goodbye party while for the other participant, moved in order to learn English, happy to be here, and making new friends may have emerged as include d terms. In addition to the assignment of included terms to the various domains, each entries and interviews. Thus, following the example above, the included term s eparation from family Una de los muchos momentos tristes de mi vida fu cuando dege a toda mi familia en Puerto Rico los dege a (One of the many sad moments of my life was when I left all my family i n Puerto Rico; I left them all; Juan, Journal 3).
83 Taxonomic analysis. A taxonomic analysis was employed in conjunction with and simultaneous to the domain analysis in the present study. Taxonomic analysis is an effective partner of domain analysis in that the taxonomic analysis represents a holistic review of the data in order to determine the relationships among domains and their included terms. Thus, established domains, included terms, and their supporting texts were connected through the taxonomic anal ysis. A simple illustration of this process, based on the previous examples, is the following: the included term moved in order to learn English as well as its supporting texts were linked to the domains coming to the U.S. and language learning; thus, th is included term linked the two domains. Similarly, a given supporting text might link to two or more included terms, which in turn would link their domains. For example, the 2 All estaban TO DOS friends were there; Carolina, Journal 3), would link to the included term, separation from family and friends (included under the domain coming to the U.S. ), as well as the included term, friends in Puerto Rico (included under the domain friends ). XSight qualitative data analysis software (QSR International, 2006) was used for both the domain and taxonomic analyses, which were carried out simultaneously for the individual profiles as well as the cross case analysis. This software facilitates the coding of qualitative data (e.g. documents, transcriptions) and allows the user to build analysis frameworks (including data displays see below) that, in this case, provided a medium
84 within which to develop and manage the domains, included terms, supporting texts, and the relationships among these. Data displays. Throughout the qualitative analysis, data displays were created for each profile as well as for the cross lines were used to create taxonomic diagrams that diagramed domains of interest their included terms, supporting texts, and the relationships among them. These data displays were initially created in conjunction with the domain and taxonomic analyses using XSight, and were later were more carefully developed using Inspiration 8 concept mapping software (Inspiration Software 2008). Because its purpose is to create concept maps, outlines, and other graphic organizers, Inspiration 8 provided numerous organizational tools and visual options to create the data displays for the present study. As they demonstrate the domains, rela ted terms, and how these are connected, the data displays provide graphic summaries of each individual profile as well as the cross case analysis.
85 CHAPTER 3 Results As a mixed methods design, the present study provides both quantitative and qualitative r way ANOVA by ranks and the corresponding post hoc test (Siegel & Castellan, 1988) were applied for the quantitative analysis. Domain and taxonomic analyses (Spradley, 1979) and data displays (Miles & Huberman, 1994) were utilized for the qualitative analysis. In this chapter, quantitative results will be presented first, followed by the qualitative results. Quantitative Results Statistical Analysis way ANOVA by ranks (within subjects) and t he corresponding post hoc test (Siegel & Castellan, 1988) were applied for the analysis of the lexical, syntactic, and discourse level scores. SPSS Statistics 17.0 software (SPSS, Inc., 2008) was used for the statistical analyses. The independent variables included language of text (Spanish/English), genre (expository/narrative), and topic (topic 1 or 2) o f text. The latter two variables were collapsed into one variable (genre topic). The genre topics for the different formal samples were: 1) Expository 1: A person I admire; 2) Expository 2: Letter to a new student; 3) Narrative 1: Special or funny family memory; and, 4) Narrative 2: My first day of school in the U.S. The dependent variables were the scores that the participants attained on each of the foll owing measures: 1) noun tiers and NDW for the lexical level; 2) the clausal complexity measure and MLT for the syntactic level; and 3) the global scores from the
86 CSE analytic writing scales (Quellmalz & Burry, 1987) for the discourse level. For each level, comparisons were made between languages within genre topic (all expository texts in both languages were compared; all narrative texts in both languages were compared), as well as among genre topics within language (all Spanish texts -expository and narrat ive -were compared; all English texts -expository and narrative -were compared). Post hoc tests were then conducted to determine which ranks differed significantly ( p < .05). The quantitative analyses were conducted to answer the following three research q uestions: 1. (Spanish/English) and genres (expository/narrative) as assessed by a noun tier analysis (adapted from Ravid, 2006) and NDW? 2. How do syntactic features of the partic and genres as evaluated by a clausal complexity measure and MLT? 3. How do discou rse features of the participant s writing compare across languages sitory and narrative writing skills (Quellmalz & Burry, 1983)? Lexical Level: Noun Tiers and NDW Noun Tiers. Scores on the noun tiers measure ranged from 1.00 to 2.50 (minimum possible score = 1.0; maximum possible score = 3.0) across all writing samples (N = 148). The distribution of noun scores was positively skewed, with the greatest number of
87 texts scoring in the 1.21 1.60 range. Table 7 provides the medians and interquartile ranges of the noun tier scores for each formal writing sample. Table 7 Media ns and Interquartile Ranges for Noun Tier Scores Genre topic Language Percentiles 25 th 50 th (Median) 75 th Expository 1 Spanish 1.32 1.56 1.68 English 1.30 1.46 1.68 Expository 2 Spanish 1.32 1.41 1.50 English 1.34 1.45 1.55 Narrative 1 Spanish 1.24 1.41 1.58 English 1.34 1.46 1.61 Narrative 2 Spanish 1.37 1.50 1.72 English 1.55 1.64 1.82 within genre 2 (3) = 11.03, p =.021. Post hoc testing confirmed that Narrative 2 in English (My first day of school in the U.S.) ranked significantly higher than Narrative 1 (Special or funny family memory) in both English and Spanish (see Figure 2). There were no significant differences within genre topic for noun tier scores in the expository texts.
88 These results suggest that, in the case of the narrative compositions, the participants used more abstract nouns in recounting their first day of school in the U.S. than they did narrating a family memor y. Hence, it appears that topic, rather than genre or (A person I admire, Letter to a new student) and languages. Figure 2 Average ranks within genre topic for noun tiers for English language comparisons, 2 (3) = 13.11, p = .004. Specifically, post hoc test ing determined that Narrative 2 in English was again ranked significantly higher than all other genre topics in English (see Figure 3) There were no significant differences among ployed more sophisticated nouns in the narratives about their first day of school in the U.S. when consistent across the texts written in Spanish. 0 0.5 1 1.5 2 2.5 3 3.5 4 Exp1 Exp2 Nar1 Nar2 Average Rank Genre Topic Span Eng
89 Figure 3. Average ranks within language for noun tiers. NDW. Scores on the NDW measure varied greatly, ranging from 29 to 146 across all writing samples (N = 148). Like the noun tier scores, the NDW distribution was positively skewed, with the greatest number of texts producing between 61 and 80 different words. Medians and interquartile ranges for NDW are provided in Table 8. within genre topic or within language for the NDW measure This result indicates that, as across the genres topics of the writing samples. 0 0.5 1 1.5 2 2.5 3 3.5 4 Span Eng Average Rank Language Exp1 Exp2 Nar1 Nar2
90 Table 8 Medians and Interquartile Ranges for NDW Genre topic Language Percentiles 25 th 50 th (Median) 75th Expository 1 Spanish 55.25 64.50 81.00 English 52.75 62.50 77.00 Expository 2 Spanish 64.50 72.00 86.75 English 53.75 72.00 86.75 Narrative 1 Spanish 51.50 65.00 77.00 English 54.50 67.50 79.75 Narrative 2 Spanish 64.25 84.50 94.50 English 62.00 75.50 87.00 In summary, on the noun tiers measure, the participants performed relatively consistently across both languages on the expository texts. However, in the case of the narrative texts, scores on Narrative 2 in English (My first day of school in the U.S.) were significantly higher than Narrative 1 in both Spanish and English (Special or funny family memory). Indeed, when all four English texts were examined within language (across genre d higher than the ranks of all other texts. In the case of texts written in Spanish, there were no significant rank differences in noun tier scores across genre topics. The participants also demonstrated consistent NDW scores across both genre topics and l anguages; no
91 significant rank differences were found for NDW. These results indicate that, with the exception of increased lexical sophistication in one writing sample (Narrative 2 English), similar irrespective of language or genre topic. Syntactic Level: Clausal Complexity and MLT inclusion (productivity) criteria for analysis were at least 10 T units and/o r 75 words. The total number of T units in the included texts (N = 148) ranged from 5 to 52 across all texts (median = 16). Similarly to the noun tiers and the NDW, the distribution for T units was positively skewed, with the majority of students producing texts containing 11 20 T units. Clausal complexity measure. The scores on the clausal complexity measure ranged from 1.05 to 1.62 (N = 148). Table 9 displays the medians and interquartile ranges for the clausal complexity measure across the different wri ting samples. within genre 2 (3) = 10.77; p = .013. Post hoc tests confirmed that Expository 1 in English (A person I admire) ranked significantly higher than ANOVA found significant differences among the ranks within genre topic for narrative 2 (3) = 9.00; p = .029. Post hoc testing determined that Narr ative 2 in Spanish (First day of school in the U.S.), ranked significantly higher than Narrative 1 in English (Special or funny family memory) (see Figure 4). The within genre topic results suggest
92 that, similar to the findings for noun tiers, topic played of more or less complex clausal structures. In this case, within the expository genre for English texts, the topic, A person I admire ranked higher than the topic, Letter to a new student. Similarly, for narrative, th e topic, First day of school in the U.S. written in Spanish ranked higher than the topic, Special or funny family memory written in English Table 9 Medians and Interquartile Ranges for Clausal Complexity Scores Genre topic Language Percentiles 25 th 50 th (Median) 75 th Expository 1 Spanish 1.25 1.35 1.51 English 1.26 1.38 1.48 Expository 2 Spanish 1.19 1.29 1.37 English 1.11 1.29 1.35 Narrative 1 Spanish 1.26 1.34 1.44 English 1.16 1.26 1.30 Narrative 2 Spanish 1.29 1.35 1.42 English 1.23 1.34 1.43
93 Figure 4. Average ranks within genre topic for clausal complexity. 2 (3) = 9.52; p = .023. Specifically, post hocs confirmed that Expository 1 (A person I admire) was ranked significantly higher than Narrative 1 (Special or funny family memory). There were no significant rank differences within language for texts written in Spanish. This result indicates that the participants utilized more complex syntactic structures in an expository text than a narrative one in English. However, their clausal complexity in Spanish was similar across both expository and narrative texts and topics (see Fi gure 5). 0 0.5 1 1.5 2 2.5 3 3.5 Exp1 Exp2 Nar1 Nar2 Average Rank Genre Topic Span Eng
94 Figure 5. Average ranks within language for clausal complexity. MLT. The MLT ranged from 4.77 21.60 across all texts (N = 148). The medians detected one significant finding for MLT for the English language texts, 2 (3) = 8.31, p = 0.4. Post hoc te sting confirmed that Narrative 1 (Special or funny family memory) was ranked significantly lower than both of the expository texts. Along with the results of the clausal complexity measure, this finding suggests that the participants wrote shorter T units as well as less complex syntactic structures -for Narrative 1 in English. This may imply that this particular topic was not very engaging for the students. 0 0.5 1 1.5 2 2.5 3 3.5 4 Span Eng Average Ramk Language Exp1 Exp2 Nar1 Nar2
95 Table 10 Medians and Interquartile Ranges for MLT Genre topic Language Percentiles 25 th 50 th (Median) 75 th Expository 1 Spanish 8.72 9.42 11.37 English 8.23 8.97 12.21 Expository 2 Spanish 7.17 8.47 9.44 English 7.53 9.02 10.72 Narrative 1 Spanish 7.18 8.78 9.89 English 7.07 7.95 9.28 Narrative 2 Spanish 8.18 8.74 10.01 English 7.81 9.00 10.44 To summarize the results on the syntactic level measures, significant rank differences were revealed for clausal complexity scores in both the expository and higher on Expository 1 (A person I admire) than Expository 2 (Letter to a new student). Additionally, scores on Narrative 1 in English (Special or funny family memory) were significantly lower than those on Narrative 2 in Spanish (My first day of school in the U .S.). Both of these results indicate that topic, yet again, had an influence on the
96 Within the English language texts, the participants earned higher clausal complexity scores on Expository 1 than they did on Narra tive 1, indicating that genre Narrative 1 in English ranked significantly lower on clausal complexity for both the variables of genre topic and language, although this was not the case for Spanish. Similarly, MLT within language was ranked significantly lower for Narrative 1 in English than for the expository texts in English. Overall, Narrative 1 in English stands out as consistently scoring lower than other writing sam ples on the syntactic level measures. Discourse L evel: CSE Scales Global scores on the CSE Scales ranged from 1 5 points; no text (N = 148) earned the maximum score of 6 points. Table 11 displays the medians and interquartile ranges for the CSE scores across the eight writing samples.
97 Table 11 Medians and Interquartile Ranges for CSE Global Writing Scores Genre topic Language Percentiles 25 th 50 th (Median) 75th Expository 1 Spanish 2.00 3.00 4.00 English 2.00 3.00 4.00 Expository 2 Spanish 2.00 3.00 4.00 English 2.00 2.00 3.00 Narrative 1 Spanish 2.00 3.00 4.00 English 2.50 3.00 4.00 Narrative 2 Spanish 2.00 3.00 3.50 English 3.00 3.00 4.00 within genre 2 (3) = 10.47; p = .015. Post hoc testing found that Expository 2 in English (Letter to a new student) was ranked significantly lower than the all other expository texts (see Figure 6). There were no significant differences in the ranks among the narrative texts. These fi ndings indicate that, again, topic played a role in e expository -but not narrative -writing samples.
98 Figure 6. A verage ranks within genre topic for CSE global scores. cant differences in the CSE global 2 (3) = 11.28; p = .01. Post hocs determined that Expository 2 (Letter to a new student) was ranked significantly lower than Narrative 1 (Special family memory). There were no signific ant differences among the ranks for Spanish. Figure 7 displays the average ranks on the CSE global scores within language. The within language results suggest tha t, in English, the participants composed a better constructed narrative (Special family memory ) than an expository (Letter to a new student) topics were of similar quality. 0 0.5 1 1.5 2 2.5 3 Exp1 Exp2 Nar1 Nar2 Average Rank Genre Topic Span Eng
99 Figure 7. A verage ranks within language for CSE global scores. To summarize the discourse level results it was found within genre topic that students performed significantly worse on Expository 2 in English (Letter to a new student) than on all other texts. Similarly, within the English language texts, Expository 2 scored significantly lower than Narrative 1 (Special or funny family memory). Thus, Expository 2 in English was ranked significantly lower than other texts both within genre topic and within language. Qualitative Results After the quantitative analyses were completed, the qualitative data analy sis, which was embedded within the quantitative method s (Creswell & Plano Clark, 2007), was carried out. Data for the qualitative analysis included the interview transcripts and written journal entries of the 6, randomly selected, focal participants. These students were: 1) Diego, grade 6, from Mexico; 2) Carolina, grade 7, from Puerto Rico; 3) Edgar, grade 7, from Mexico; 4) Sara, grade 8, from Mexico; 5) Manuel, grade 8, from Mexico; and 6) Juan, grade 8, from Puerto Rico (see Table 2). The students and t heir profiles are 0 0.5 1 1.5 2 2.5 3 3.5 Span Eng Average Rank Language Exp1 Exp2 Nar1 Nar2
100 presented by grade level (lowest to highest) and, within each grade level, gender (girls, then boys). The purpose of the qualitative portion of the study was to answer the research question: How do previous and current language and literacy learning experiences and/or attitudes, beliefs, and feelings about second language writing? To address this question, domain analysis was used in conjunction wi th taxonomic analysis (Spradley, 1979) and the development of data displays (Miles & Huberman, 1994). The qualitative results included two levels of analysis: 1) individual, within case analyses performed on each focal participant to serve as profiles of t hese bilingual adolescents; and 2) a cross case analysis that served to compare and contrast how the focal participants perceived their bilingual status. With respect to the domain analysis, as discussed in the Methods chapter, the domains were created apr iori to data collection and comprised the topics of the journal entries and interview questions. On the other hand, the journal entries and their responses to the int erview questions. Table 12 outlines the topics of the 10 journal entries, their genres, and, for the expository entries, their primary structure. (See Appendix C to review the journal prompts.)
101 Table 12 Journal Topics and Genres Journal entry Topic Genre and structure (expository) Journal 1 Intro to journal Expository, describe Journal 2 Happy moment Narrative Journal 3 Sad moment Narrative Journal 4 A problem or conflict Narrative Journal 5 The languages I speak Expository, compare contrast Journal 6 Sports and hobbies Expository, describe Journal 7 Goals Expository, cause effect Journal 8 Family/cultural traditions Expository, describe Journal 9 My dream vacation Narrative Journal 10 Three wishes Expository, explain Data displays, in the form of concept maps (Miles & Huberman, 1994), were created during the qualitative data analysis to organize and summarize the results. XSight qualitative data analysis software (QSR International, 2006) and Inspiration8 concept mappi ng software (Inspiration Software, 2008) aided in the analysis and the creation of the data displays. (The concept maps are presented in Appendix G.) In the following section, the individual profiles of each focal participant are presented first, followed by the cross case analysis. In order to integrate the qualitative analysis with the quantitative findings, the linguistic characteristics of the focal
102 entries. Table 13 providing th eir global scores from the CSE Scales (Quellmalz & Burry, 1983). This scale 6, with global scores representing the following levels of general w riting proficiency: 1) not at all competent; 2) not very competent; 3) almost competent; 4) adequately competent; 5) definitely competent; and, 6) very competent (p. 26). Dis course Level section in the Methods chapter.
103 Table 13 Global Scores on CSE Scales Genre topic Language Diego Mexico Carolina PR Edgar Mexico Sara Mexico Manuel Mexico Juan PR Expository 1 A person I admire Spanish 3 4 5 3 2 3 English 2 4 3 2 1 4 Expository 2 Letter to new student Spanish 3 4 3 2 1 4 English 2 4 2 1 1 3 Narrative 1 Special family memory Spanish 3 2 3 3 2 3 English 2 5 2 1 2 2 Narrative 2 First day of school U.S. Spanish 3 3 4 3 3 2 English 3 4 2 3 2 3 Average for all texts 2.6 3.8 3.0 2.3 1.8 3.0
104 Profiles The journals and interviews of the 6 focal participants provided a substantial amount of qualitative data to analyze, considering the embedded nature of the qualitative analysis in the present investigation. As examples, the transcripts amounted to 30 doub le spaced pages of journal text and 72 double spaced pages of interview text. In addition to the 10 domains generated apriori by the topics of the journal entries (see Table 12), the student interviews afforded additional domains common to all of the focal participants. These included background, language usage, language learning, and bilingualism, coming to the U.S., school (home country and U.S.), and family and friends among others The goal of the present qualitative analysis was to discover how the pa language and literacy practices and experiences contributed to their identities as bilingual writers. For this reason, the domains reported here -as well as their derived included terms and supporting texts for each focal participant -represent only those that specifically supported the research question 5 These three interrelated domains are: background, coming to the U.S. and language learning, language usage, and bilingualism (all of which included the content Notable quantitative findings (i.e., linguistic characteristics of some formal writing samples based on the lexical, syntactic, and discourse measures) are included in the profiles of each foca l participant. 5 Although domains of interest other than thos e discussed here (e.g., goals and dreams ) are included in the data displays (see Appendix G).
105 Diego: Futbolista Trilinge (Trilingual Soccer Player) Background. nio mexicano (Mexican kid; Journal 1) with a passion for soccer, was born in Florida. However, Diego moved with his family back to their hometown of Ixmiquilpan, in the state of Hidalgo, Mexico, at the age of 5 years. He attended kindergarten through grade 4 in Ixmiquilpan. When he was began, for the f irst time, to study English in school. At the time of data collection, Diego was 12 years old and a student in grade 6. Coming to the U.S. his journal 6 Yo me mobi para estados Unidos porque y o tenia asma y costaba muy caro las medicinas y siempre tenia que tomarlas pero nada, siempre me sentia lo mismo tosia en la noche y cuando corria me cansaba muy rapido y no podia respirar United States because I had asthma and the medicin es were very expensive and I always had to take them but nothing, I always felt the same and was always coughing at night Journal 3). Although in his journal Diego complained that his move to th Yo amprendi de esta experiencia que no es malo mudarse de un lado a otro porque puedes sabe or aprender otras cosas que no sabes Journal 3). 6 are interpreted in square brackets in cases where t he error is likely to interfere with comprehension.
106 Language learning. When asked in the interview how he felt about learning and speaking English at school, Diego had a mixe d response. He initially stated, about Fue fcil de aprenderlo, luego, luego 219 ) However, like the other focal participants, Diego noted the existence of language prejudice at school, where he claimed to speak both Spanish and English with his friends and teachers. Referring to some teachers at Bayview (although specifically not Ms. Brady, the ESL teacher) espaol. Que aqu es Amrica dicen, que aqu es los Estados Unidos United States; lines 252 253). When asked to offer some insight as to why people might cmo se dice? entender speaking, of not knowing what they are saying, like in Spanish, not understanding; lines 260 262) Language us e and bilingualism. Diego considered himself to be trilingual, because he spoke not only Spanish and English, but also Otomi, an indigenous language of Mexico. According to Diego and other focal partic ipants from the same area (Manual un poquito saba un montn; saba hablarlo I knew how to speak it; line 59). Diego used to speak Otomi with his grandparents, who still live in Mexico and predominantly
107 Tengo tambin un abuelo tambin que l no habla espaol, casi puro ese idioma that language; lines 75 76). At the time of data collection, Diego lived with his parents, brother, sister, aunt, uncle, and cousin. He reported that out of these relatives, his aunt was the most likely to speak O tomi, and she sometimes spoke it with his mother. Spanish Diego wrote almost all of his journal entries in Spanish, with the amples in both languages earned CSE global scores between 2 and 3 points, which is to say that his writing ranged from not very competent to almost competent on the rating scale. His Spanish writing samples generally outscored the English texts. For the mo st part, his writing was void of punctuation and paragraph separations, and contained numerous of lexical sophistication; however, some examples of complex sentences were observed Yo quiero conocer al ronaldinho en persona para ablar con el y que me ensee a dominar el balon como el (I want to meet ronaldinho in person to talk with him and so that he can teach me how to dominate the ball like him ). This sentence includes two subordinate clauses (adverbials of purpose). In Narrative 2 (My first day of school in the U.S.), Diego wrote: cuando yo iba al bao o a otro lado me lamaba Gonz alo i les dije que yo no soy Gonzalo ( when I went to the bathroom or somewhere they called me Gonzalo and I told them that I am not Gonzalo ) This sentence also contains two subordinate clauses, an
108 adverbial temporal (beginning with when ) and a nominal obj ect clause (beginning with that ). Overall, Diego appeared to have adjusted to middle school in the U.S. and had a positive attitude about learning English at school. When asked how he felt about it, he Yo aqu me siento bien, hablar ingls aqu (Here I feel good, speaking English here; line 224) This being said, although he self identified as trilingual and reported being comfortable speaking English, when given the choice Diego preferred to speak or write in Spanish. In addition, his formal writ ing samples generally scored higher in Spanish than English. Carolina: Modista en Pars (Fashion Designer in Paris) Background. Carolina, who dreamed of studying fashion design in Paris, w as born in Puerto Rico and moved with her family to Kentucky at the age of 3 years. She lived in Kentucky for 4 years, where she attended preschool, kindergarten, grade 1, and part of grade 2. Her family then returned to Puerto Rico, where Carolina attende d grades 2 6. At the time of data collection, Carolina was 13 years old and a student in grade 7. Because she attended the primary grades in Kentucky, Carolina learned literacy first in English. Additionally, she studied English in school while living in P uerto Rico. Coming to the U.S. According to Carolina, her family moved to Florida from Puerto Rico so that she could improve her English. Se me estaba olvidando [el ingls], lo que como ahora me acostumbr en Puerto Rico a hablar el espaol, pues se me estaba olvidando. Pero entonces cuando mami me dijo a m que se me estaba olvidando el ingls, yo me empec a preocupar. Porque mami me
109 dijo a m que el ingls es bueno (I was forgetting [English], since now I was used to speaking Spanish in Puerto Rico, so I was forgetting it But then when Mom told me that I was forgetting English, I started to worry. Because Mom told me that English is go od lines 195 198). Like many of the participants, Carolina described leaving Puerto Rico as a sad Mi momento triste fue cuando me fui de Puerto Rico. Fue bien triste porque tuve que dejar a todos mi familiares, amigos y personas que quiero mucho sad moment was when I left Puerto Rico. It was really sad because I had to leave all my relatives, friends and people that I love a lot; Journal 3). However, in the same journal entry she stated, Me alegro mucho que me vine para Florida. Ahora conoci muchas personas nuevas. Tengo amistades nuev as. Estoy muy contenta. Aveses extrao a Puerto Rico, pero fue lo mejor que ise! (I am really happy that I came to Florida. Now I met a lot of new people. I have new friends. I am very happy. Sometimes I miss Puerto Rico, but it was the best thing I ever did! Journal 3). Language learning. Carolina explained in her interview that she began learning English at age 3 when her family came from Puerto Rico to Kentucky. She spent four years in Kentucky where she attended grades K 2. Regarding the experience of learning to read in English, she stated that it was, Difcil y a la misma vez fcil, porque las palabras eran ms o menos como en espaol because the words were more or less like in Spanish; lines 130 131). Later, sh e described
110 her return to school in Puerto Rico and learning to read and write in Spanish: (There it was easy for me, because I already know Spanish and every day I am speaking it; lines 147 148). Regarding her return to Florida and learning English at Bayview Middle, aqu todo el mundo habla el espaol everyone speaks Spanish; line 185). Apparently, Carolina was aware that an advantage of having knowledge of two languages is that one can support the development of the other. Language use and bilingualism. Of the 6 focal participants, Carolina most clearly identified herself as a bilingual language user. In fact, when a ddressing the topic of El espaol yo lo hablo con mi familia y mis amistades latinos y el Ingls con los maestros y amistades que no hablan El Espaol. Cuando estoy en mi casa o hablo los dos lenguajes y lo mismo en mi escuela. I speak both languages everywhere I go cuz there are lots of bilengual people When I am at home I speak both languages and the same with school. I speak both languages everywhere I go cuz there are lots of bilingual people; Journal 5). Similarly, when asked which language she preferred for the Ambos, cualquiera whichever ; line 5). Indeed, both Both, Spanglish 34).
111 Carolina also expressed a positive view of b ilingualism and stated that she thought she was fortunate to be able to speak two languages. When asked how she felt Orgullosa! Similarly, although she recognized that language prejudic e exists, she was confident that envidiosa. people who think Carolina at all; in fact, her perception of other students at school was that they thought, ¡Ay, tu (Wow, you are lucky! Lines 232 233). Carolina wrote her journals mostly in Spanish with some English; in both languages she wrote well in comparison to the other participants. On the CSE Analytic Writing Scales, she earned a score of 4 (adequately competent) on all of her expository formal wr iting samples (in both Spanish and English), and a score of 5 (definitely competent) on Narrative 1 in English (Special or funny family memory). This organized, prov ided adequate support, and contained only a few mechanical errors was not notable. However, she did display clausal complexity in her writing in both English and Spani sh. For example, in Expository 1 (A person I admire), in English: The person I most admire is my mother Because she works really hard to give me and my
112 brother everything she can This sentence contains 4 subordinate clauses: two relatives, an adverbial ca usal, and an adverbial of purpose. It is notable that the two relative clauses in this example do not contain relative pronouns (i.e. that ). In English, the relative clause structure without the pronoun is considered less sophisticated (i.e., more aligned with oral language) than the clause containing a relative pronoun (Hunt, 1965 ) However, acquisition of a differentiated syntactic structure because, in Spanish, relative pronou ns in this context would be obligatory. In Spanish, in Expository 2 (Letter to a new student), Carolina wrote: Cuando vengas para ac a te aconsejo que te prepares para las peleas porque aqui se forman muchas peleas ( When you come here I advise you that you prepare for fights because here a lot of fights are started ). This sentence contains three subordinate clauses: an adverbial temporal (beginning with when ), a nominal object (beginning with that ), and an adverbial causal (beginning with because ). It was evident in her writing and interview that Carolina enjoyed school and learning. She was comfortable socially and made friends quickly when she came to Florida. Carolina also strongly self identified as a bilingual and considered this a positive attribute t hat others might envy. Overall, her writing, which scored relatively evenly both languages.
113 Background. Edgar was born in Mexico City and attended public school there for grades kindergarten through fifth. He continued grade 5 when his family moved to Houston, Texas when he was 11 years old At the time of data collection, Edgar was 13 years old and attended grade 7. Edgar had plans to return to live his adult life in Mexico where he aspired to be a professional soccer player or attorney. Language learning. Except for some instruction in kin dergarten, Edgar did not have any experience learning English in Mexico. Lo que pasa es que donde yo iba no era escuela privada. Era una escuela del gobierno y no nos s son las escuelas privadas en que te ensean todo. Las escuelas (What happened is that where I went was not a pri vate school. It was a government the private schools that teach you 29, 36 37). Regarding learning English at school in the U.S., Edgar expressed general feelings of negativity. In El ingls no me gusta para nada No s, pero no me llama mucho la atencin y entonces por eso no he aprendido yo tampoco, porque n o, no, algo que no me gusta, yo no le tomo importancia 344).
114 Later in the in S me interesan los idiomas, pero es que son muy difciles para aprender. Es lo que no me gusta: son bastante difciles 399 400). Hence, although Edgar recognized it might be important to learn English or other languages, he found language learning difficult and did not enjoy doing so. Lang uage use and bilingualism. Edgar expressed that he uses Spanish whenever possible, and did not consider himself a bilingual. El ingls lo tengo aprendido como 50%. Tengo otro 50% que me hace falta. No me considero biling lo puedo decir. Muchas veces no lo entiendo I have was said. A 408). Like the other focal participants, Edgar recognized the existence of language Yo hablo espaol y me gusta mi idioma. aunque a mucha gente no le [gusta] el espaol aqui en los estados unidos. con mis amigos yo hablo espaol y con mis papas hablo espaol my friends I speak Spanish and with my parents I speak Spanish; Journal 5). is journals and formal samples provide evidence that Edgar has the potential to be a persuasive writer, especially in Spanish. It is not surprising that Edgar received higher CSE global ratings
115 on the texts he wrote in Spanish (he received a 5, 4, and two were notable for the use of sophisticated, even metaphorical, lexical items. For example, in Expository 1 in English (A person I admire), Edgar included the Spanish words, ave fenix (phoenix) and polbo de estrellas (stardust) in describing a friend whom he admired. In Expository 1 in Spanish, Edgar used the words pasion (passion) destreza (dexterity), and triunfos (triumphs) in his descript ion of Maradona, the famous Argentinean soccer pla yer. This text also demonstrated details, for example in the complex sentence, Nacido en Argentina en un barrio muy pobre desde nio descubrio su gran pa cion por el futbol ( Born in Argentina in a very poor neighborhood since childhood he discovered his great passion for soccer). This sentence contains a fronted adverbial phrase, which can be viewed as a stylistic option used by more mature writers to creat e thematic variety in a text (Perera, 1984). Overall, Edgar seemed to have the lexical, syntactic, and text level skills to write proficiently in his first language, Spanish. However, he had not yet acquired enough proficiency to achieve the same level of writing in English, nor did he have the desire or confidence to Sara: Familia Unida (United Family) Background. Sara was born in Ixmiquilpan, in the state of Hidalgo, Mexico, where she attended kindergarten through grade 5. She entered grade 6 when she came to the U.S. with her mother and two sisters, prior to which she had never visited the U.S. or studied English in school. At the time of data collection, Sara was 13 years old and
116 attended grade 8. Sara had hopes to return to live in Mexico to attend college and have a career as a human rights attorney. Coming to the U.S. Sara repeatedly discussed how her father h ad abandoned the family to be with another woman, which prompted her mother to move with her and her from the family had a great effect on Sara, who repeatedly remin isced in her journal about the times when the whole family was together having fun, in the past. She also repeatedly wished for her whole family to be united again, including her father because, mi mama no puede ser mam y pap para mi both a mother and a father to me; Journal 10). At the same time, Sara expressed Su dedicacion es trabajar y sacarnos adelante tiene 4 aos que mi pap no vive con nosotros mi segundo deseo seria que toda mi familia podriamos regresa a mexico para est ar juntos (my second wish would be that all of my family could return to mexico to be together). Language learning. Like Diego, Sara studied Otomi in school in Ixmiquilpan and had family members who spoke it. Specifically, she explained in the interview t hat her father and his side of the family speak Otomi well. Regarding her own proficiency in (I understand it,
117 Regarding learning English at school, wh en asked how this experience has been Pues, ms o menos, no muy buen porque cuando llegu, pues, yo no le entenda nada, ni que me decan. Pues tengo primas aqu que tienen ocho aos aqu, pero ellas si saben dominar bien el ingl (Well, more or less, not very good because when I cousins here who have been here for 8 years, b ut they do know how to speak English three languages ; lines 125 128 ). Language use and bilingualism. In her journal, Sara clearly identified herself as a Spanish speaker, alt hough she ironically wrote the first sentence of this entry in English: espaol y yo ablo ingles con mi sister en mi casa casi no ablo ingles por que nadien abla ingles e never speak English because nobody speaks English at home at school I try to speak it bec ause the teachers speak English; Journal 5) When asked if she considered herself to be bilingual, Sara responded, laughing, g. With the exception of a few isolated words and phrases, all of
118 Spanish; one 3, some emerging complexity in English, for example, in Expository 1 (A person I admire), she wrote, The person I most admire is my mother Because she work really So hard every day Although this sentence contains errors (i.e. subject verb agreement), it includes two subordinate clauses: a relative clause (relative pronoun absent) and an adverbial causal Exp ository 1, which contains identical clausal structures: A la persona que yo admiro es ami mami por que ella nos a sacado adelante (The person that I admire is my mom because she has moved us ahead ). Overall, Sara appeared to be an emerging bilingual student who did not feel proficient enough in English to self identify a s a bilingual. Although her writing scores were higher in Spanish than English, her texts demonstrate similar syntactic structures across both languages. (See Figure 11, Appendix G, fo data display.) Background. Hidalgo, Mexico. Like Sara and Edgar, Manuel was born in Mexico and grew up there until his family moved to the U.S. and he entered grade 6, where he began to study English for the first time. At the time of the study, Manuel was 14 years old and attended grade 8. Manuel also desired to return to live his adult life in Mexico, where he hoped to contribute to the development of his hometown. Coming to the U.S. family came to the U.S. In his interview, he explained that he would never feel at home in
119 the U.S.: No me (I lived there Porque siento que no es mi pas, no Because I 111). Language learning. Like the other students from Ixmiquilpan, Manuel also had experience studying Otomi in school and being exposed to it through family members. Again, like the others, Manuel claimed to underst and it but not speak it. When asked in the interview to describe his experience learning English here, Manuel quickly No volviendo a nacer porque es otro idioma another language; lines 64 65). In addition to his frustration with learning English in school, Manuel also noted that language prejudice has had a negative effect on him (lines 75 Me he dado (I have realized that there are some people who say that we should not speak Spanish; lines 75). When asked how he felt No s (mumble), me Language use and bilingualism. Manuel stated in his interview that he prefers to use Spanish whenever possible, including speak ing with family and friends, watching T.V., and listening to music. Additionally, Manuel did not consider himself to be bilingual. In his interview, Manuel even went as far as stating that Mexicans do not speak
120 Spanish well: jico. Hay unos que piensan que [los (Spain arrived to conquer Mexico. 9 8). Hen ce, Manuel, an ELL student who wa s uncomfortable and unhap py learning English, also placed a low value on his own first language. Manuel struggled during data collection to produce all of the texts, and had difficulty writing in both Spanish and English. In fact, his formal samples were eliminated from the quantitative analysis because half of them did not meet the productivity criteria 7 His global scores on the CSE scale ge nerally stayed between 1 and 2 points, that is, not at all competent to not very competent. However, he scored a 3 on Narrative 2 in Spanish. In this text, Manuel provided details about his first day of school in the U.S. and developed a more elaborate tex t with some examples of clausal complexity. For example, Yo cuando llege a los EEUU y vine a la escuela el primer dia estaba muy nerbioso por que no conosia a nadien ( When I arrrived in the U.S and came to school the first day I was very nervous because I ). This sentence contains two subordinate clauses: an adverbial temporal (beginning with when ) and an adverbial causal (beginning with because ). Manuel also wrote, in Expository 2 in Spanish (Letter to a new student): yo 7 Recall that quantitative and qualitative data were collected simultaneously; therefore, the focal participants were randomly selected from the original group of 24 participants. It was discovered later
121 quisiera desirle a un estudiante de mi pais que este pais no es lo mismo porque I muchas cosa muy diferente aqui en los EEUU (I would like to say to a student from my country that this country is not the same because there are many different things here in the U.S. ). This sentence also contains two subordinate clauses, a nominal object clause (beginning with that ) and an adverbial causal clause (beginning with because ). Overall, Manuel was a struggling student in both his first and second languages. He was not happy livin g in the U.S. and felt that learning English was very difficult. scores on the CSE scale. Indeed, Manuel was the type of ELL student who may have slipped through the cracks w ith an unidentified language or learning impairment (Wagner, Francis, & Morris, 2005) display.) Background. rnal 7), was born in Puerto Rico, where he attended kindergarten through grade 6. Like Carolina, Juan received English instruction at school in Puerto Rico every year, and also had private English lessons to intensify his studies during the year before he moved to Florida. In addition to these experiences, Juan visited family members in the U.S. every summer for 2 month periods; however, he claimed that he only spoke Spanish with his family during these visits. Juan relocated to Florida as a student in grad e 7, and was 13 years old and in grade 8 at the time of data collection.
1 22 Coming to the U.S. Also like Carolina, Juan claimed it was his idea to come to me tuve que hir de p uerto rico para venirme asia los Estados Unidos para aprender Ingls. Yo me tube que ir porque En Puerto Rico los cosas estan malas y total este fu una desision mia de benirme asia los Estados Unidos I had to leave because in Puerto Rico things were bad and really it was my own decisin to come to the United States; Journal 3). Similarly to other participants, Juan described his De esta esperiencia yo aprendi que algunas decisiones no son lo que tu piensas y pueden doler mucho en el corazn you think and they can cause a lot of pain in your heart; Journal 3). Language learning. When asked about his feeling s regarding learning English at school, Juan reported positive experiences and a desire to learn. He expressed that knowing English would be important to his future career as a professional basketball player in the NBA. In addition, Me gusta porque Ms. Brady te ayuda. Ms. Brady si ve que si te si necesitas preguntas, ella antes, casi siempre antes de que t llegues al saln est me da un email si que te ayudo, si que necesita ayuda me l lama, me avisa. Te ayudo en lo que necesite s (I like it because Ms Brady helps you. Ms. Brady if she sees that you if you have questions, she before, she gives me her email if I
123 196). Teacher support was apparently a motivator for Juan as he adjusted to his new school and language environment. Language use and bilingualism. In s pite of his claim that he came to the U.S. to Yo hablo espaol bajo todas circunstancias. Porque espaol es el lenguage en el que yo me puedo deshaogar con todo el mundo. Cuando hay problemas cuando me molesta (I speak Spanish under all circumstances. Because Spanish is the language in which I can let go with everyone. When there are problems when something is bothering me whatever I speak Spanish; Journal 5). Although Spanish is his language of choice, when asked in the interview if he considered himself bilingual, he responded, con los morenos, yo hablo en ingls. Con los blancos hablo en ingls. Yo hablo ingls. Lo hablo, lo hablo con las personas que tengo que hablarlo with the Blacks, I speak English. With the Whites I speak English. I speak English. I speak it, I speak it with the people that I have to speak it with; lines 224 225) In addition, Juan expressed posit Me siento bien porque al, al mismo tiempo entiendo lo que dice mi gente, y tambin entiendo lo que dicen las otras personas. Y me defiendo.Me puedo defender cuando digan, cuando dicen cosas malas a mi, o cualquier cosa, porque me defiendo me I understand what my people are saying, and I can also understand what other people say. And I can defend myself. I can defend myself when they might say, when they say bad things to me, o r whatever, because I can defend myself; lines 227 229).
124 entries were written in Spanish. points on the CSE scale, and t both Spanish and English. That is, he did not consistently write better in one language or and syntactic com plexity. In describing his brother in Expository 1 in English (A person I admire), Juan used the words, [he is my] shadow and guardian In the Spanish version of the same text, Ju an described his brother as an alma hemela (soul mate) and a buena persona de corazon noble (good person of noble heart). With regard to syntax, Juan was one of the few students who mixed languages in his writing (although not in his journal), for example, he is with me en las buenas y en las malas Exposito ry 1, English); I like teachers to be gentle, and buenas because when a teachers yell at me I get angry with them and I am very tranquilo English). Overall, Juan appeared to be a confident young man with emerging writing abilit ies in both Spanish and English. Although he predominantly expressed himself in Spanish when given the opportunity, he appreciated the value of learning English as a second language, enjoyed his ESL class, and felt secure in his abilities. (See Figure 13, Summary of Profiles The individual profiles that resulted from the qualitative analysis uncovered much variety among the focal participants with regard to educational backgrounds, transitions
125 to the United States, language learning experiences, and perceptions of bilingualism As across Spanish and English languages. Diego, Edgar, Sara, and Manuel came to the U.S. from Mexico, and none of them had received any English language instruction prior to attending school here. Among these four students, varying strengths emerged. Diego had a positive outlook on learning and using English in s chool; however, his writing in Spanish generally scored slightly higher on the CSE Scales than in English. Edgar and Sara demonstrated a similar pattern, with CSE Scale scores consistently higher in Spanish than English. Edgar displayed some sophisticated vocabulary choices in his writing in Spanish; however, his scores on the CSE Scales in English were, for t he most part, in the range of not very competent. Sara, on the other hand, expressed a desire to learn English and utilized similar structures in her writing in both languages. Finally, Manuel, who was not happy in the U.S. and felt that learning English w competent to not very competent in both Spanish and English. Unlike the students from Mexico, Carolina and Juan, who grew up in Puerto Rico, had experienced English language instruction prior to moving to Florida. In fact, Carolina had attended grades K 2 in Kentucky. Carolina felt comfortable using both Spanish and English and demonstrated this in her writing, which, for the most part, scored in the adequately competent range for both S panish and English texts. Juan, who also expressed that he felt proficient as a Spanish English bilingual, achieved similar scores on the CSE
126 Scales in both languages. This brings us to the cross case analysis regarding the focal of bilingualism. Cross Case Analysis: Bilingual Perspectives In addition to the case studies, a cross case analysis was carried out to compare perceptions and feelings regarding bilingualism. The students varied co nsiderably in their language learning experiences as well as in their opinions about learning English and bilingualism. Out of the six focal participants, only Carolina and Juan, both from Puerto Rico, had studied English in school prior to arriving in Flo rida. The students from Mexico, on the other hand, had not received any previous instruction in English; however, Diego, Sara, and Manuel studied Otomi in school in Mexico and experienced some degree of its usage with their families. Overall, those student s who considered themselves bilinguals had positive attitudes about learning English and bilingualism in general. On the other hand, students who self identified as Spanish monolinguals did not enjoy learning English or feel proficient in English as a seco nd language. Bilingual Identity and Positive Views of Bilingualism The focal participants who considered themselves bilinguals included Diego, Carolina, and Juan. Diego came from Mexico and had not studied English prior to arriving in Florida in grade 4; however, he did have some experience with learning a second language, Otomi. Both Carolina and Juan grew up in Puerto Rico where they studied English at school. In addition, Carolina attended grades pre K 2 in Kentucky, and Juan visited the U.S. often as a child and had private tutoring in English for a year before
127 he moved to Florida. All of these students claimed to regularly speak both languages, enjoy and/or value both languages, have proficiency in both languages, and feel happy living in the U.S. Reg arding the regular usage of both languages, Diego, Carolina, and Juan claimed to make use of both Spanish and English when watching television, listening to music, and reading for fun. Juan was particularly explicit in explaining his bilingual T.V. viewing habits: bisbol y a m me gustan mucho los deportes, y que es en ingls. Y ahora los programas de msica porque a m me gusta mucho la msica, como el MTV Tres, ah eso lo veo en espaol. Este, T elemundo lo veo en espaol, y yo veo como tres o cuatro canales porque Y HBO, que es en really like sports, like music, like MTV Tres, that one I watch in Spanish. Like, Telemundo I watch in Spanish, and I watch like three or four channels because the rest are in English because I watch a lot of 113). Diego, Carolina, and Juan also claimed to use a combination of Spanish and (interview of Carolina, line 34), when speaking with friends and family. When asked to describe espanglish Le pongo s esas cosas
128 all the friends 103). For these students, learning English at school has been a generally positive experience. For example, Juan expressed that he liked English class and felt supported by the ESL teacher, and Carolina and Diego shared their experiences with first friends who helped them by interpreting when they first arrived. At th e same time, unlike the students who self identified as monolinguals, these participants did not emphasize the difficulties or challenges of learning a second language. In fact, in his interview Diego described Aqu cuando yo lle (Here when I arrived I already, like I 217). In addition, Carolina and Juan expressed the ide a that English was a valuable skill or tool. Porque mami me dijo a m que el (Because Mom told me that English is good for the future; lines 197 198). Juan elaborated on how Englis h would be important to his future as a Me siento bien porque el ingls es como, es una de las, es una de las segundas lenguas, es la segunda lengua ms importante del mundo. y mi sueo es llegar a, a l a NBA a jugar baloncesto profesionalmente. Y ah tengo que hablar el espaol [ingls] porque no me gusta que veces l dice un error, y la gente te entiende mal. Por e so no me gusta tener a alguien [i.e., un int rprete]
129 o d because and my dream is to g et to the NBA to play basketball sometimes he makes a mistake, and people mi to have anyone translating for me; lines 204 209) In spite of their general positive feelings about bilingualism, Diego, Carolina, and Juan all expressed having experienced language prejudice for speaking Spanish i n school. Cuando yo ablo espaol en la escuela dicen que able ingles que aqui es america yo pienso que ellos esta como celoso porque nada mas saben un idioma y nosotros sabemos dos idiomas o mas yo ablo 3 idiomas k Spanish at school they say that I should speak English that this is America I think that they are like jealous because they only know one language and we know two languages or more I speak three languages; Journal 5). These three focal participants shar ed the explanation that people who responded negatively toward Spanish speakers did so out of jealousy or envy. In his interview, Juan went beyond this explanation to discuss his perception of the social dynamic of the various racial groups at his school: En esta escuela yo pienso que algunos, porque los morenos, tienen muchos problemas con los mexicanos y a veces, yo pienso que los Que yo siempre veo que los morenos se llevan mejor con los boricuas y lo s y los dominicanos, y todas esas personas as (In this school I think that some, because the B lacks, they have a lot of problems with
130 the Mexicans and at times, I think that the B always see that the B lacks get along better with Boricuas [Puerto Ricans] and the and the Dominicans, and all those types of people; lines 239 245). Although these students recognized the existence of language prejudice, they also resisted letting it bring them down. For exa mple, Carolina, who stated that she felt proud to be bilingual, also explained that most other students consider her lucky to know more than one language. Similarly, Juan discussed how bilingualism could benefit not only him, but also the monolingual Engli sh speakers around him. es algo ms fcil p ellos. Porque ah no tiene que estar explicndole a la persona y haciendo y pasando tanto trabajo en explicarte. Por eso yo pienso que es mejor para ellos for them. for them; lines 232 234). Overall, Diego, Carolina, and Juan identified themselves as bilinguals and expressed a general sense of satisfaction with their English language learning experiences. These students could be considered emerging bilingual writers who are capable of composing in English; however, when given the choice, all three preferred to speak or write in Spanish. For example, in spite of their perceived bilingual status, Diego and Juan elected to have their interviews in Spanish. Carolina, who initially stated that the language of the interview did not matter, also spoke predominantly Spanish throughout the interview.
131 Similarly, all three participants wrote the majority of their journals in Spanish, although Diego and Carolina included some English in a few of the entries. Diego also generally received higher CSE ratings on Spanish texts than English on es (see Table 13). Carolina and Juan scored similarly across both languages, although in the case of narrative texts, Carolina received higher ratings in English. Hence, as can be expected for still irregular. Monolingual Identity and Negative Views of Bilingualism The focal participants who did not consider themselves bilinguals were Edgar, Sara, and Manuel. These students all came from Mexico and had not studied English before coming to the U .S.; however, Sara and Manuel had some experience learning and speaking Otomi as a second language in Mexico. In addition, all three students expressed the desire to move back to Mexico to attend college and/or to live as adults. Edgar, Sara, and Manuel di d not consider themselves to be bilingual because they felt that they did not speak enough English, expressed that English was difficult, did not want to learn English, and/or were not happy or comfortable living in the U.S. Edgar and Manuel were particul arly decisive about using Spanish whenever possible, in conversation as well as when watching television or listening to music. Edgar Yo hablo espaol y me gusta mi idioma. aunque a mucha gente no le [gusta] el espaol aqui en los es tados unidos. con mis amigos yo hablo espaol y con mis papas hablo espaol Even though many and with my parents I speak Spanis h; Journal 5).
132 Although Sara expressed an interest and desire to learn English, she felt that she did not have the proficiency to be considered bilingual. On the other hand, Edgar alluded to some proficiency in English, but insisted in his interview that he simply did not like it: (My lines 377 378). Manuel in particular expressed unhapp iness with living in the U.S. He claimed to feel depressed about language prejudice and stated that he would never feel at (because I feel that it is not my country; line 111). Like Diego, Carolina, and Juan, th e students who self identified as monolinguals shared the experience of language prejudice in their school and community. Sara (Well, the re are a lot of people who are racists, 189). (grade 5) called and sent notes to her moth er complaining that the girl was speaking En la escuela de mi hermana, ahorita, hace poco, ella hablaba espaol, as, con sus amigas. Entonces la otra vez que llaman a mi mam y le dicen na que su hija habla puro espaol. Siempre, toda la agenda, llena de notas y dice, na, que speaking Spanish, like this, with her friends. Then the other time they called my mom and
133 they said that her daughter is speak ing only Spanish. Always, all of her planner, full of 194). In spite of their overall negative feelings toward learning English, Edgar and Sara did express the understanding that it might be helpful for them to become bilingual. In particular Sara noted repeatedly her attempts to practice English at school and at home, where she described mixing languages with her two sisters: Es que empezamos hablando en espaol. Hay unas palabras q ingls, lo decimos en espaol. Y pues tratamos de hablar ingls how to say them in English, we say them in Spanish. And well, we try to speak English; lines 95 100). Sara also claimed to read in both English and Spanish although reading in A la vez en ingls para que aprenda. Y luego (At the same time [I read] in know how to read; lines 117 118). Yo s que el o aprender. Y ahora s le pienso echar un poco ms ganas al ingls. Pero ya cuando salga de aqu ya yo sepa ingls, aqu de la secundaria And now here I already know English, here out of secondary school; lines 346 348). In summary, Edgar, Sara, and Manuel identified themselves as monolingual, Spanish speakers and expressed a general lack of confidence and/or desire to learn
134 English, or felt that learning English was difficult. All of these students elected to have urnal, they wrote all of their journal entries in Spanish. Edgar, Sara, and Manuel also consistently received higher CSE global ratings on the formal samples produced in Spanish than those written in English. Unfortunately, this is not to say that these st udents are consistently demonstrating competence in their writing Spanish. Manuel, in particular, is a struggling writer in both languages. (See Figure 14, Appendix G, for the data display of the cross case analysis.) Results Summary Summary of Quantitativ e Results Lexical level results summary. For the noun tiers measure, it was evidenced that topic -more than language or genre -their formal written texts. In particular, Narrative 2 (My first day of schoo l in the U.S.) in English resulted in more abstract noun use than the other genre topics in both languages. differences for either language or genre topic comparisons, in dicating that lexical variety and usage were similar in both Spanish and English as well as for both expository and narrative texts. Syntax level results summary. For the clausal complexity measure, higher scores were produced on Expository 1 (A person I admire) in English than on Narrative 1 in structures in their writing. Additionally, Narrative 1 (Special or funny family memory) in
135 English ranked significantly lo wer on clausal complexity both for genre topic and language comparisons. This topic also ranked significantly lower for MLT for the English language comparisons. This result may imply that this particular topic was not engaging for the students. Discours e level results summary. Topic again appeared to play a role in the English was ranked significantly lower than the other texts both within genre topic and within language. This may be an indication that the letter format of Expository 2 was more challenging for students to compose in English than the narratives or other types of expository texts. Summary of Qualitative Results Profiles results summary. The qual itative profile analysis provided an inside view of the 6 focal participants that explored beyond their writing scores to encompass the following three interrelated domains : 1) background; 2) coming to the U.S.; and 3) language learning, language usage, an d bilingualism. The domain and taxonomic language experiences, and self perceptions of themselves as more or less proficient bilingual learners. In particular, the students from Puerto Rico, who had more experience studying English at school, emerged as having more similar writing scores in both languages. On the other hand, the students from Mexico, regardless of their self identification as bilingual or monolingua l, generally scored higher on texts written in Spanish than English.
136 Cross case results summary. Finally, the cross case analysis of the focal two patterns in language identity. The first pattern, biling ual identity and positive views of bilingualism, was exhibited by Carolina and Juan (from Puerto Rico), as well as Diego (from Mexico). These three students reported that they regularly spoke and felt proficient in both languages, enjoyed/valued both langu ages, and felt happy living in the U.S. Notwithstanding their self identification as bilinguals, these participants elected to write or speak in Spanish when given the option (journal/interview). Also, while the quality of their formal writing samples (CSE Scales) varied across languages, these students (particularly Carolina and Juan), tended to receive more consistent scores across Spanish and English The second pattern, monolingual (Spanish speaking) identity and negative views of bilingualism, was demo nstrated by Edgar, Sara, and Manuel (from Mexico). These three students felt that they did not speak enough English, English was difficult, they did not want to learn English, and they were not happy or comfortable living in the U.S. Additionally, all thre e of these students expressed a desire to return to live in Mexico. Like the self identified bilingual students, Edgar, Sara, and Manuel elected to write and speak Spanish when given the option (journal/interview). However, in contrast to the bilingual gro up, the quality of their formal writing samples (CSE Scales) was consistently rated as higher in Spanish than English.
137 CHAPTER 4 Discussion The present study explored, through an embedded, mixed methods design (Creswell & Plano Clark, 2007), expository a nd narrative writing in Spanish and English of 20 ELL students in middle school. All of the students produced 8 formal writing samples controlled for language and genre, as well as 10 journal entries in the language of their choice. The quantitative portio n, which was the primary focus, examined the languages (Spanish/English) and genre topics (Expository/Narrative, Topic 1 or 2). For the qualitative aspect, domain a nd taxonomic analyses (Spradley, 1979) were applied to the journal entries along with interviews of 6 randomly selected focal participants to their identities as bilin gual writers. The study addressed three quantitative research questions: 1. (Spanish/English) and genres (expository/narrative) as assessed by a noun tiers measure (adapted from Ra vid, 2006) and NDW? 2. and genres as evaluated by a clausal complexity measure (adapted from Ravid & Berman, 2006) and MLT? 3. compare across languages and genres as examined by the CSE Scal es (Quellmalz & Burry, 1983)?
138 The embedded, qualitative portion of the study was driven by a fourth question: 4. How do previous and current language and literacy learning experiences and/or prac attitudes, beliefs, and feelings about L2 writing? The discussion that follows is organized by these four research questions. For each question, three areas will be address ed: 1) interpretation of the patterns of findings in light of ELL research; 2) factors influencing the assessment of ELL writing including challenges presented by the measures that were applied and recommendations for future research; and 3) instructional and clinical implications for ELL literacy development. Following the discussion of the research questions, a general discussion presents the overall conclusions of the study and directions for future research in this area. Discussion: Quantitative Patterns Lexical Level Patterns two measures were used: a noun tiers measure (adapted from Ravid, 2006) and NDW. Differences in vocabulary use across languages (Spanish/English) and/or genres (expository/narrative) were expected. Noun tiers. In the case of the noun tiers measure, the scores ranged from 1 .00 to formal writing samples (N = 148). However, the distribution was positively skewed, with esult indicated a relative dependence on tier 1 (more concrete) nouns with less frequent use of tier 2 and 3
139 (more abstract) nouns. This pattern is comparable to the findings of Harley and King (1989), whose French language learners in grade 6 utilized mor e high frequency/utility verbs than less frequent, derived verbs in comparison to native French speakers on two narrative and three expository (letters) writing tasks. The participants in this study may have applied a similar pattern of noun usage in their compositions in both Spanish and English, selecting more high frequency, concrete nouns in their writing. noun usage; instead, some of these meanings were present, particularly in the Spanish texts (e.g. felicidades (congratulations), encanto (enchantment), destreza (dexterity), a lma gemela (soul mate), s eguridad (security), and vergenza (embarrassment). Even so, expectations, the statistical outcomes did not reveal significant rank differen ces in noun tier scores between the Spanish and English texts. On the other hand, a significant result on the noun tiers measure for the genre topic comparisons was that Narrative 2 in English (First day of school in the U.S.) ranked significantly higher than Narrative 1 in both languages (Special or funny family memory). For within English language comparisons, Narrative 2 also ranked significantly higher than the other three genre topics. These results imply that the topic of Narrative 2, rather than the language or genre, may have afforded the students a more productive framework within which to explore their lexical abilities in English. Perhaps, because Narrative 2 dealt with the subject of school in the U.S., the participants found it easier to includ e related English vocabulary that was more frequent in their day to day activities.
140 Indeed, nouns such as locker and office (tier 1), homeroom, period and tour (tier 2), and schedule, problem, and fear (tier 3), etc. were common in these English writing s amples. The fact that a narrative text ranked higher than the expository compositions on the noun tiers measure contrasts with the findings of Ravid (2006), whose noun scale analysis of spoken and written narrative and expository texts determined that wri tten age level (students in grades 4, 7, 11, and university graduate level). A note of caution is (2006) findings. The speakers of Hebrew who were from middle to upper middle socioeconomic status (SES) in Israel. Hence, linguistic as well as sociocultural differences -su ch as quality of education -may explain this discrepancy in results. Factors influencing the consistency of coding The noun tier application was not without problems, however. Ravid (2006) briefly noted that inter judge agreement for the 10 level noun sc ale reached 91% once all of the nouns in the 320 text sample had been process of de termining inter coder agreement. For example, it was unclear how the check coders were trained or what agreement formula was applied to the results (Bakeman & Gottman, 1986). Based on the challenges in coding that were experienced with this measure in the report may affect the overall reliability of the noun scale.
141 For the current study, at least three procedural challenges arose regarding noun scale coding in general as well as cross language coding differences specific to Spanish or English. One such procedural challenge for coding consistency was the fact that the same noun could be coded at more than one level in different linguistic contexts. For example, to class the noun, class was considered a level 3 (location); in contrast, this noun was class In a third possible scenario class would be considered a lev el 9 (abstract) item. A similar example was the word time This time time to grained 796). Hence, although the noun scale was designed to consider noun tokens in context, in practice, consistently coding contextually variable nouns on a 10 level scale was difficult. level noun scale into 3 noun tiers likely minimized some of the inconsistencies noted above. However, when coding for agreement on the noun tiers, a second issue arose. The coders did not always agree on the words to be regarded as nouns. For example, although the researcher did not consider indef inite pronouns (e.g., everything, someone, nothing) as nouns, one of the check coders rated these words as tier 2 nouns (generic). Also, certain nouns formed parts of commonly used phrases in Spanish, such as le ech muchas ganas (I put a lot of heart into it) or isla del
142 encanto (island of enchantment, motto for Puerto Rico). In both of these situations, the nouns ganas and encanto were initially coded as tier 3 (abstract); however, the check coder did not judge these nouns at all because they were deemed to be elements of the check coders with texts in which the previously determined nouns to be coded had been identified, thus eliminating the possibility that certain nouns would not be coded at all. This strategy could also have been applied to a final discrepancy that emerged during the Spanish check coding, namely, that compound nouns such as arroz con frijoles (rice and beans) could be regarded as either one or two sepa rate tier 1 nouns. The decision to treat compounds as one or two words influenced overall coding in that these disagreements resulted in different total numbers of nouns in the sample. This issue was resolved by asking the check coders to rate any missing or overlooked nouns, which were identified by the researcher. Consequently, both elements of Spanish compound nouns were coded as individual nouns. Shifting to cross language challenges, differences in the forms and meanings used to express physical and e motional states (e.g., hunger, thirst, fear, shame) resulted in a group of level 9 (abstract) nouns that existed only in the Spanish equivalents. For l state, the Spanish translation (I had hunger) instead utilizes the verb to have and includes an abstract noun as the direct object of this verb. The same is true for expressions containing thirsty/thirst, sleepy/sleepiness, afraid/fear,
143 em barrassed/embarrassment, etc. This cross language difference did not appear to influence the results in favor of the Spanish texts. Nonetheless, this issue should be considered for future applications of the noun scale when comparing across different langu ages. NDW No significant differences were found for NDW across the languages or genre Spanish and English as well as across the genre topi cs of the writing samples. This finding is not particularly surprising based on the res ults of the noun tiers measure. That is, the results of the noun tiers measure did not demonstrate consistent rank differences between the Spanish and English texts, nor between the expository and narrative genres. It is also noteworthy that NDW has not been shown to be a robust predictor of differences in writing quality (Beers & Nagy, 2009; Perkins, 1980). The noun tiers and NDW patterns indicated that the participants used similar, yet rather un sophisticated, lexical items in their written texts across the languages and genre was unexpected, it is notable that these ELL adolescents wrote with similar levels of lexical sophistication or simplicity across all texts. Implications. These patterns of findings are consistent with the consensus that acquisition of more literate, academic vocabulary is both a challenge and a critical factor in the school success of EL L students (Janzen, 2008; Snow & Kim, 2007; Wilkinson & Silliman, 2008). One aspect of academic vocabulary development is vocabulary depth. Ordez and colleagues (2002) described vocabulary depth as including quality of a
144 morphological structure, and related syntactic structures. Vocabulary depth is also conceived as a metalinguistic, cognitive skill that is language independent. Along these lines, Ordez et al. indicated that vo cabulary depth, expressed through use of to transfer across languages for Spanish English bilingual children in grades 4 and 5. In order to develop academic vocabu lary, including vocabulary depth, in either language or both, ELLs must have access to explicit instruction and frequent opportunities to apply this specialized vocabulary in varied and meaningful literacy experiences (Janzen, 2008; Wilkinson & Silliman, 2 benefit from increased (or more effective) vocabulary instruction and practice to improve their lexical diversity and depth in writing. Additionally, with respect to cross language semantic relationships, Harley and Ki ng (1989) found that English Speaking, French language learners in grade 6 made use of lexical similarities, especially cognates, to maximize their vocabulary resources when writing in L2. Systematic, comparative analysis of word roots and derivational mor level vocabulary in both languages. Higher level vocabulary includes what Beck, McKeown, and Kucan (2008) designated as Tier 2 and Tier 3 words, as opposed to Tier 1 words that comprise the basic oral language devious is a Tier 2 word that has a similar meaning as the more basic words, tricky and sneaky Tier 2 words
145 may als o be derived (e.g., poverty, impoverished; context, contextual, decontextualize) and are considered high frequency words for experienced language users. On the other hand, Tier 3 words, which are used infrequently, include vocabulary specific to academic d isciplines such as mathematics, science, or history (e.g., parallel, photosynthesis, unconstitutional ). Since many of these Tier 2 and 3 words are derived from Latin or Greek roots, they often share cognates in both English and Spanish (e.g., context/conte xto, parallel/paralelo, photosynthesis/fotosntesis ). Also, derivational morphemes share commonalities in English and Spanish as well ( able/able, tion/cin, dis/des, ab/a ). These forms could be taught in a comparative/contrastive manner to increase ELL stu languages. A cautionary note to this recommendation is that successful teaching about cognates may not be an available strategy for educators without sufficient knowledge of Spanish (Snow & Kim, 2007). Finally, the challenges in coding along with the unexpected cross language differences presented by the noun scale/noun tiers measure suggest that, for future examinations of ELL vocabulary in writing, other measures may be more effective to e to consider aspects of vocabulary depth in writing through a morphological analysis of derived words and/or content specific, academic vocabulary (Wilkinson & Silliman, 2008). Alternately, it might be useful to consider frequency of nouns and/or verbs on a low mid high frequency scale, similar to Harley and King (1989).
146 In addition to these measures, teachers of ELLs would benefit from practical, efficient assessments of lexical proficiency that can be applied in the real life context of the classroom An example of this might be an assessment tool that quantifies use of elaborated noun phrases 8 in writing. This type of measure would offer insight into lexical a s well as morphosyntactic development and could be applied in English and Spanish These types of investigations and assessments have the potential to provide more information about the extent to which phonology, orthography, and morphology are coordinate d into a unified representation for the expression of meaning in both L1 and L2. Syntactic Level Patterns choice of syntactic structures in their formal writing samples. Differences in syntactic structures across languages (Spanish/English) and/or genres (expository/narrative) were expected outcomes. Clausal complexity measure The scores on the clausal complexity measure ranged from 1.05 1.62 across all 148 formal writing samples, which indicated that the 8 Examples of I remember something funny and very special with my f amily (8M5, Narr1Eng); my beautiful mother (8M8, Expos1Eng) Examples in Spanish include Yo admiro a Maradona un exelente jugador de futbol (I admire Maradona, an excellent soccer player ; 7M2, Expos1Span); mi hermana era elejida para candidata a reyna de la feria de Pozuelos Cardonal Hidalgo Mexico (my sister was chosen as a candidate for queen of the fair of Pozuelos Cardonal, Hidalgo, Mexico ; 8M2, Narr1Span).
147 participants, for the most part, wrote with independent clauses and single subordinate clauses rather than sentences with multiple, embedded subordinate clauses. Combined with the results of the noun tiers measure, the cl ausal complexity measure further paints an overall picture of relatively unsophisticated structures employed by the students in their formal writing samples across both languages and genres. However, for the adolescents in this study, topic again played a role in determining significant differences in clausal complexity. In this case, for the within genre topic comparisons, Expository 1 in English (A person I admire) ranked significantly higher than Expository 2 in English (Letter to a new student) Additi onally, Narrative 2 in Spanish (First day of school in the U.S.) ranked significantly higher than Narrative 1 in English (Special or funny family memory). In the case of the English language comparisons, Expository 1 ranked significantly higher than Narrat ive 1. The within genre topic results again point to the topic of a writing sample as a more important influence on syntactic complexity in this study than the language of expression; namely, for both expository and narrative texts, significant rank diffe rences increased syntactic complexity on an expository text than a narrative one better aligns with previous research findings of increased syntactic sophistication in expository texts than narrative texts (Berman and Nir Sagiv, 2007). MLT In the case of MLT, Narrative 1 in English (Special or funny family memory) was ranked significantly lower than both of the expository texts in English. Along with the results of the clausal complexity measure, this finding suggests that the
148 participants wrote shorter T units, as well as less complex syntactic structures, for Narrative 1 in English than for the other writing samples. This may imply that this particular topic was not v ery engaging for the students. As adolescents in middle school, the participants in this study were likely spending more time with friends than their families. Therefore, the Narrative 1 topic may not have succeeded in evoking the prior experience/memories necessary for the students to compose more productive texts. Additionally, Narrative 1 in English was the first writing sample elicited; hence, a novelty task effect or task expectations may have influenced how students approached the writing activity. F actors impacting the measurement of syntactic complexity. ELL research has employed numerous measures to explore syntactic language proficiency with mixed results regarding their ecological validity. The present study is no exception, as it may be question ed whether the most appropriate syntactic categories were examined to compare The clausal complexity measure utilized number of subordinate clauses as a measure of T unit complexity; however, Kameen (1979) fou nd no significant differences based on number of embedded clauses or types of clauses. On the other hand, Kameen did find that significant predictors of writing skill were length of T unit (MLT) -also applied here -and use of the passive voice. Based on these results, Kameen recommended that the ELLs should be provided with instruction and practice in sentence combining, in
149 clauses to preposi tional, infinitival, and participial phrases In contrast, Perkins (1980) found that MLT, among several other syntactic writing scores. Instead, Perkins determined that only measures that considered errors (error free T units per text, number of words in error free T units, and total errors) emerged as significant predictors of writing outcomes. Finally, Beers and Nagy (2009) found that words per c lause, rather than words per T unit, was positively correlated with text quality for monolingual, English speaking middle school students. It is possible that these micro level measures of syntactic ability have not afforded consistent patterns of finding s due to their narrow scope. That is, tools that attempt to isolate syntax from the discourse features of its context might overlook important relationships among the various levels of language present in a text. Ariel (2009) argued that grammar and discou rse simultaneously complement and impose different constraints system of linguistic behavior, and just like horse and carriage, th ey definitely go l, 2009, p. 6). From this perspective, it becomes impractical to attempt to establish a means to measure grammar without considering the discourse in which the syntactic structures are situated. In their report on developing academic literacy for ELLs, Short and Fitz simmons (2007) also discussed acquisition of English academic vocabulary and sentence structure in the context of discourse and text knowledge. For these authors, academic literacy begins with exposure to and understanding of multiple types of texts
150 (inclu ding different genres and text media) for different purposes. Vocabulary and grammar are developed as a consequence of these interactions with texts. Regarding teaching strategies for ELL writing, both grammar and discourse level structures must be addres sed, and these may be instructed both in isolation and in conjunction with one another. Research on cross language transfer has found that, while syntactic level structures appear to be language specific, text level structures may form part of a bilingual support for the use of some direct instructional strategies that are specific to syntax and text structure. These issues are further discussed in the next section. Implications. Just as ELL students face the challenge of acquiring the specialized vocabulary of academic language proficiency in a second language, they also must learn to comprehend and produce more literate sentence structures in English to achieve grade level expect ations (Janzen, 2008; Short & Fitzsimmons, 2007). For the students in this study, the degree of syntactic complexity they displayed appeared to depend on their complexity in w riting also appeared to be relatively equivalent across languages. This finding conflicts with expectations as well as previous research on cross language transfer. For example, Francis (2006) argued that, although bilinguals possess a unified, underlying language proficiency, certain skills, including grammatical rules and syntactic (1982) study in which, for ELLs in grades 1 3, syntactic complexity appeared not to tran sfer across writing in Spanish and English. These findings emphasize that, for
151 bilinguals, language specific knowledge, as well as common underlying proficiencies, are equally important contributors to (academic) text production. With respect to bilingual that is the basis for text is acquired individually as a function of experience in each language. Children must establish the literary basis of their linguistic competen ce With this in mind, the results of the present study suggest two possible explanations for the apparent consistency in syntactic level expression across Spanish and English texts. The first explanatio n offers a glass half empty, or deficit, perspective. Namely, the participants had not achieved a high level of syntactic complexity in writing their emerging writin g in L2, English. For some of the participants, this scenario could be due to differences in educational experiences in their home countries. The second explanation provides a glass half full, or strengths, focus. The participants had achieved a sufficien t level of English language proficiency to produce syntactic structures in L2 that were on par with, if not more advanced than, their already established capabilities in L1. In this light, the students can be viewed as growing in their English academic lan guage proficiency, which, due to the effects of schooling in L2, should eventually surpass their academic language skills in L1 (Kohnert, 2008). These explanations are truly two sides of the same coin and offer some insights into the ideas of common under lying proficiency and the threshold hypothesis (Cummins, 2000). Based on research supporting the non transfer of syntactic skills (Francis, 2006),
152 the ELL students would not necessarily need advanced capabilities in L1 to develop them in L2. However, as pr eviously emphasized, syntax cannot be separated from discourse (Ariel, 2009). It is certainly possible that the nature of the writing task itself (i.e., the assigned genre and language, the use of a highly structured prompt, the purpose for writing) also m writing. For example, in Expository 1 (A person I admire), the last point in the prompt Invariably, the stu and included an adverbial conditional clause followed by a main clause in the conditional Similarl y, prompts that included why inevitably resulted in the production of T was a fu n day because my whole family was together These patterns provide an alternative explanation for why certain topics may have been more effective in eliciting more productive or more sophisticated writing across the measures. The first explanation, discu ssed previously, is that certain topics may have been more engaging for the participants in that they were better able to identify with the topic (e.g., Narrative 2, My first day of school in the U.S., which ranked high on the noun tiers measure). A second possible reason is that, beyond the topics themselves, certain vocabulary), affecting the final outcomes of the statistical analysis.
153 Keeping in mind the mutual i nfluence of syntax and discourse, this effect might be minimized in future examinations of the data by considering larger units of text (i.e., beyond the clause and T unit) for analysis. Indeed, the results of this study, along with the diverse research fi ndings with regard to the measurement of ELL syntactic ability and shared versus language specific skills, suggest a need to develop new measures of written syntactic complexity for ELL students that move beyond unitary surface structures to include larger units of text that take into account the overlap of sy n tax and discourse. For example, Berman (2008) explored this possibility by breaking texts down into clause packages, which encompass multiple T units or sentences and are linked by syntactic or seman writing for analysis, it may be useful to see how students can reconstruct basic written texts by combining simple sentences into more complex ones. Along these lines, Scott and Nelson (2009) explored how sentence combining tasks might provide a useful writing assessment, in particular, to aid in the identification of children with language learning disability (LLD). Likewise, Saddler and Preschern (2007) offered instructional strategies combining. Indeed, sentence combining could be used as both an instructional strategy and an proficienc y. Because writers essentially rebuild stories through the sentence combining strategy, this task not only serves to highlight the relationships between syntactic and discourse features of a text, but also allows for the assessment of these features in an integrated fashion.
154 Ariel (2009) emphasized the interconnectedness between discourse and grammatical structures and highlighted the bidirectional influences between them. Deeper yet broader explorations of these relationships may serve to increase our cr oss linguistic understanding of the written syntax of ELL students. Attempting to isolate syntax from multiple levels of text. Discourse Level Patterns To examine the p discourse level, the CSE Scales (Quellmalz & Burry, 1987) were applied. Differences in global text level scores across languages (Spanish/English) and/or genres (expository/narrative) were expected results. At the discourse level, the CSE global scores also showed significant differences for the genre and topic, although not for the language, of the writing samples. In this case, for the English language comparisons, although Narrative 1 (Sp ecial or funny family memory) ranked low on the clausal complexity measure, it ranked significantly higher than Expository 2 (Letter to a new student). In addition, for within genre topic comparisons, Expository 2 in English ranked significantly lower than all other expository texts in both languages. This may be an indication that the letter format of Expository 2 was more difficult for these ELL students to produce in English than was the narrative structure or the more traditional expository style text o f Expository 1 (A person I admire). Additionally, this result might be an indication that the topic of Expository 2 (Letter to a
155 new student) was not as engaging, or t he prompt was not as supportive, for the students as the alternative Expository 1 (A person I admire). The finding that a narrative text ranked higher than an expository text on the CSE Scales (in English) provides some support for the Berman and Nir Sagi v (2007) finding that, although more advanced grammar and vocabulary may be observed in expository texts than narratives (which was not necessarily the case here), young writers first achieve command over narrative structure in the elementary grades and ar e only able to successfully formulate well organized expository texts beginning in the middle high school years. It should be noted that the Berman and Nir Sagiv (2007) study included 160 writing samples (half narrative, half expository) produced by 80 mo nolingual English speakers in grades 4, 7, 11, and university level. Again, this sample was considered middle to upper middle SES in the United States. Despite these procedural and sample size differences, Berman and Nir ding narrative and expository texts may afford a useful explanation of the current result. It is possible that, in the case of ELLs, students would be more likely to capitalize on their strengths and would maximize their proficiency with the better establi shed, more comfortable narrative structure for writing in L2. The explanation that students relied on better developed narrative skills across the languages, as well as the lack of significant differences among texts written in Spanish and English, suppor ts evidence that text level writing skills, in contrast to grammatical structures, may be transferable across languages irrespective of L2 oral proficiency
156 (Dressler & Kamil, 2006; Fitzgerald, 2006, Lanauze & Snow, 1989). The explanation for this finding l ies in the distinction between syntax and discourse as linguistic processes. First, syntax is relatively language specific (Francis, 2006); hence, particularly in cases of languages from different families (e.g., Spanish, a Romance languages versus English a Germanic language), bilingual students must acquire different sets of grammatical rules for each system 9 In contrast, knowledge of genres, text structure, and composition processes can be considered language general skills; therefore a student who dev elops skills related to the writing process in one language may be able to apply them in another (Dressler & Kamil, 2006). Indeed, text level skills (e.g., knowledge of genre characteristics, such as story grammar, appear to form part of the common underly ing proficiency accessible across L1 and L2 (Cummins, 1991; Durgunoglu, 2002). Factors affecting the assessment of overall text quality Although Quellmalz and agree ment among different raters, this type of measure relies on expert judgment focused evaluation which, according to Schriver (1990), may present challenges for inter rater agreement. On the other hand, an analytic writing assessment such as the CSE scales c an provide a wealth of information about a text. In the case of the present analysis, the CSE 9 However, this may not be the case for lear ners who acquire a second language in the same family as their L1 (e.g., Spanish and Portuguese, both Romance languages). While there is little research on this topic, it would appear that metalinguistically savvy speakers of one system could transfer know n grammatical rules and structures to the other. However, even in the case of similar languages, what works in one may not always function in the other, resulting in errors (Carvalho & da Silva, 2008).
157 Scales did prove to be a challenge for inter rater agreement. Two main factors may have contributed to the difficulties in this area. First, neither the research er nor the check coders experienced the CSE training sequence in the way it was initially prescribed by Quellmalz and Burry (1983). The CSE Scales were originally developed to use in school districts, and the creators recommended a multi step district trai ning involving several levels of practice for scoring, checking, and testing of raters. Clearly, this type of training was not appropriate for the purposes of this study; however, increased training and discussion among the researcher and the check coders throughout the rating process may have raised the agreement percentages. A second methodological issue was the subjective nature of this type of scoring broad area of ju dgment such as mechanics, several elements were included in the overall rating for that domain (i.e., sentence construction, usage, spelling, and punctuation/capitalization). In this case, one rater may have placed more emphasis on one element than another causing a difference in scoring for that area. Finally, regarding the readin g of compositions progresses (Quellmalz, 1980, p. 7). Therefore, scores may have been inflated in the case of students who wrote above average texts as compared with the rest of the participants. These two issues call for improved ways to rate ELL writt en texts at the level of discourse. Perhaps alternative discourse level features, such as coreferential structures or
158 the other hand, the value of analytic or holistic scoring cannot be dismissed, as this type of assessment can provide teachers and speech language pathologists with important development over time. Implications. At the middle school level and beyond, adolescent students are expected to demonstrate their knowledge of content through writing (Graham & Perin, 2007; Janzen, 2008). Additionally, as Short and Fitzsimmons (2006) noted, students at this age level have both in s chool and out of school literacies, including household responsibilities such as reading or translating bills or communicating with landlords, physicians, etc. Based on the results of the CSE Scales, it was evidenced that, as a group, the ELL students in t his study were producing written texts that could be considered marginally competent in both Spanish and English. In fact, the majority of the global scores fell in the 3 to 4 point range, which is at the top of the non mastered level (1 3 points) and the bottom of the mastered level (4 6 points). Global scores of 3 and 4 were adequate exa there were not significant discourse level differences across languages, it could be said that these middle school ELL students were emerging bilingual writers in both Spanish and E nglish.
159 These outcomes indicate that the students, rather than composing and revising ldren writer takes cues from the writing assignment (prompt) and genre to activate related topic knowledge. The writer then generates a text by simply writing down everything s/he recalls about the given topic. Bereiter and Scardamalia (1987) contrasted the knowledge writers engage in a recursive, back and forth writing practice that involves developing and recreating thoughts and text in an interrelated process. Due to its linear nature, the knowledge telling strategy can often result in more primitive lexical and syntactic choices, as well as a general lack of global planning and t 2006, p. 16). Also, McCutchin (2006) noted that, because the knowledge telling strategy relies heavily on topic cues, students produce better texts when they have more si tuational knowledge of the topic at hand. This could explain why topics like Expository 1 (A person I admire) or Narrative 2 (My first day of school in the U.S.) proved to be more productive for the participants than others on these particular measures. I n addition, as recently addressed, the structure of the prompts, as well as their organization for their texts, examples of possible content, and even specific grammatical structures to include (see examples in the Syntactic Patterns section). This elicitation
160 procedure -which emphasized the prompt -may actually have primed the students to activate the knowledge telling strategy, more so for those who did not readily identif y knowledge telling framework (Bereiter & Scardamalia, 1987), which resulted in written compositions that were fairly straightforward (following the prompts step by step) and lexically, syntactically, and textually basic, yet consistent, across languages. It appears that, as inexperienced writers, the participants applied a simple yet efficient knowledge telling strategy that resulted in less sophisticated textual or ganization (as well as vocabulary and syntactic structures) in their writing. The implication of this conclusion, along with the evidence of language transfer of text level proficiencies, is that ELL students could benefit from writing instruction that aim s to further develop their skills in both L1 and L2. Indeed, Fitzgerald (2006) suggested that difficulties with ELL writing are not generally caused by lack of proficiency in L2 but rather problems with the composing process in general. For the bilingual w riter these difficulties would affect both languages. Shanahan and Beck (2006) recommended structured writing instruction (rather than free writing) as well as explicit instruction in revision to improve the writing achievement of ELLs. Further, Short an d Fitzsimmons (2006) made the point that ELL writing development also requires support outside the language arts classroom. Content area teachers can contribute to the development of ELL literacy by integrating reading and writing activities into their ins truction, for example, by using strategies that build s ( Beck et al. 2008). Along with
161 ELL specialists, teachers in the content areas can also support academic writing development through direct inst ruction and practice of genre specific characteristics and structures (Beck & Jeffery, 2009). For example, Beck and Jeff ery suggested that History and English classes are ideal contexts for students to engage with the genre of analytic exposition because these subject areas require analytic interpretation of historical events and literature. Similarly, teachers of science can promote the development of science information texts, another type of expository genre invo lving analysis, interpretation, and synthesis (Wilkinson & Silliman, 2008). In closing, it should be noted that text level writing skills cannot flourish without continuous development of lexical and syntactic proficiencies in a learning context that is ch allenging, yet supportiv e, of ELL students (Walqui, 2007 ). Quantitative Conclusions The quantitative findings of this study can be summarized by four overall patterns that surfaced in the formal writing samples of the participating ELLs. Impact of topic. First, although some genre related distinctions emerged, significant differences in lexical, syntactic, and discourse rankings were generally based identifica tion and engagement with writing topics, and also suggests the possibility that vocabulary or syntactic structures. These findings indicate that topic choice and elicita tion procedures may play a more important role in writing assessment than previously thought.
162 Transfer of academic language proficiency A second key finding was that across the lexical, syntactic, and discourse measures, the participants consistently disp layed similar skills in both Spanish and English texts. This finding supports previous research findings regarding language transfer and common underlying proficiencies (Cummins, 2000). Additionally, this pattern suggests that academic language proficiency may cross linguistic boundaries and hence, its development in either L1 or L2 would support literacy skills in the other language as well. Knowledge telling orientation to writing The third conclusion is that the results in the lexical, syntactic, and d iscourse domains identified the participants as emerging writers in both Spanish and English. It appeared that the ELL students in this study depended on a basic knowledge telling strategy typical of children and immature writers (Bereiter & Scardamalia, 1 987). Consequently, although the writing skills of this group of students may appear marginal or even adequate, these ELLs still have quite a bit of work to do in order to increase their literate vocabularies, acquire more sophisticated syntactic structure s, and shift from a knowledge telling to a knowledge transforming (Bereiter & Scardamalia, 1987) model of composition. Multiple linguistic levels of text composition A final consideration highlighted by the overall quantitative findings is the need to ac knowledge the relationships among lexical, syntactic, and discourse features of written text composition. The implication here is that educators and speech language pathologists in the area of ELL literacy must provide meaningful opportunities for students to develop these skills in an integrated fashion, also providing direct instruction on specific structures in isolation when
163 appropriate. Certainly, as developmental models of writing attest, a hallmark of the mature, successful writer is the ability to i ntegrate multiple language elements with composition skills in an automatic and recursive process (Berninger & Hooper, 2003; Torrance and Galbraith, 2006). Discussion: Qualitative Findings Patterns of Language Learning and Identity The qualitative interp retation applied domain and taxonomic analyses (Spradley, 1979) in conjunction with data displays (Miles & Huberman, 1994) to construct individual profiles that explored the language and literacy learning experiences of the 6 focal participants. The profil e analysis centered on three interrelated domains : 1) background; 2) coming to the U.S.; and 3) language learning, language usage, and bilingualism. Simultaneously, a cross case analysis was completed to compare and contrast the focal participants regardin g their self perceptions and feelings toward bilingualism. The results of these analyses highlighted differences as well as similarities among the focal participants and supported a more in depth understanding of their writing performance based on their la nguage experiences and practices. Overall, four general patterns emerged from the qualitative analyses. Language preference. The first pattern relates to the language choices of the focal participants. Notably, all of these students opted to hold their in terviews in Spanish with predominantly Spanish. The 6 focal students also wrote the majority of their journal continued preference to use L1
164 when given the choice, regardless of their perceived status as a bilingual or monolingual. This finding is similar to the experience of Dworin (2006), whose Latino students in grades 4 and 5, in collaboration with their pare nts, wrote the majority of their family stories (15 of 18 texts) in Spanish when given the option. participants had arrived in the U.S. within 2 years prior to data collec tion. Hence, most likely, they felt more confident using Spanish in an academic context (Cummins & Schecter, 2003). Yet, in spite of their relatively recent integration into schooling in English, these students demonstrated varying levels of proficiency in their writing in both Spanish and English. These differences were probably influenced by many factors, including previous educational experiences and, as discussed next, their identities and perceptions of L2 learning and bilingualism. Bilingual and mo nolingual identities. The second pattern that emerged from the cross bilingualism. Namely, the three students who self identified as bilinguals, Diego, Carolina, and Juan, regularly spoke and felt proficient in both languages, enjoyed/valued both languages, and felt happy living in the U.S. On the other hand, those students who considered themselves to be monolingual, Edgar, Sara, and Manuel, did not feel they had achieved an acceptab le level of English proficiency, stated that English was difficult, and expressed that they were not motivated to learn English and were not comfortable living in the U.S.
165 In general, the quantitative analyses revealed that the total group of participants did not consistently display statistically significant rank differences between Spanish and English on any of the writing assessment measures. When individual profiles of the focal participants were created, however, a somewhat different pattern emerged. Of the 6 focal participants, three students identified themselves as bilinguals and three identified themselves as monolinguals. On the CSE global measure, the self identified monolingual students consistently scored higher on the texts written in Spanish. In contrast, two of the students who identified themselves as bilinguals, Carolina and Juan, displayed a more consistent performance across the two languages. Language discrimination. A third pattern that arose from the qualitative analyses relates to th Spanish at school or in the community. This outcome that is consistent with the findings of other qualitative research on adolescent Latino students in the U.S. (Bejarano, 2005; McHatto n et al., 2007). and was also discussed in the interviews. All of the focal participants revealed that they had at some point heard someone comment that they should speak English in this count ry; however, their responses to this evidence of language prejudice varied. Although most of the students, perhaps influenced by their ESL teacher, rationalized these co mments made him feel depressed. Juan further described language prejudice as a part of the racial conflicts he observed at school, including gangs involving African American and Mexican students.
166 Ethnic differences. Finally, and related to the previous th emes, the fourth pattern distinguished the Puerto Rican students from those who came from Mexico. The two students from Puerto Rico, Carolina and Juan, had studied English and had experiences living in the U.S. before coming to Florida. Both of these stude nts considered themselves to be proficient bilinguals and had a positive view of bilingualism. These students also received similar CSE global scores on their writing in both languages. On the other hand, the four students from Mexico, Diego, Edgar, Sara, and Manuel, had not studied English prior to their arrival in the U.S. Of the four, only Diego self identified as a bilingual (trilingual); the others considered themselves monolingual Spanish speakers. Of the Mexican students, Diego Sara, and Manuel als o had experience using an indigenous language, Otomi, with their families and in school. Finally, regardless of their self identification as bilingual or monolingual, the students from Mexico scored consistently higher on texts written in Spanish than Engl ish. Factors Influencing Legitimation Some obstacles to legitimation ( Onwuegbuzie & Leech, 2007 ) arose in the qualitative portion of the study. For example, in the case of the interviews, it would have been desirable to conduct member checks 10 (Guba & Lincoln, 1989) with the interviewees; however, due to time constraints, i.e., the end of the school year, this was 10 The purpose of the member check is to solicit partic interpretation/coding of interview data. This technique has been recommended by numerous authors as a means to decrease threats to validity in qualitative research (Anfara, Brown & Mangione, 2002; Bran tlinger et al., 2005; Maxwell, 2005; Onwuegbuzie & Leech, 2007).
167 not possible. Also due to timing issues, it was not feasible to collect additional qualitative data (e.g., through classroom observations or document analysis beyond the writing samples). These data would have further served to triangulate and enhance the qualitative analysis. Finally, for future qualitative analyses of the present data set, check coding is recommended. In spite of the aforem entioned limitations, the qualitative findings of this study bring to light the methodological and clinical potentials of moving beyond statistical analyses of group data in order to explore the linguistic profiles (and social identities) of individual stu dents. Indeed, for the present study, the qualitative profile analysis offered enhanced insight into the language and literacy backgrounds, attitudes, and writing outcomes of the participants. Implications For an ELL student, both the perceived proficien cy in L2 and the self determined purposes for using L1 or L2 contribute to investment in learning and using English (Norton Peirce, 1995). Regardless of their status as bilingual or monolingual, Mexican or Puerto Rican, all of the focal participants identi fied with their Latino heritage, home language/s, and previous educational experiences in their countries of origin. As Bloome et al. (2005) stressed, literacy entails much more than learning to decode a system of symbols; rather it is a complex sociocultu ral process that involves socialization, enculturation, power relations, identity production, and situated interaction. Therefore, effective literacy instruction for ELL students must take into account their background
168 experiences as well as the sociocultu ral and linguistic resources they bring to the classroom (Cummins et al., 2005; Moll et al., 1992). With this in mind, Kohnert (2008) suggested that even monolingual speech language pathologists can achieve increased cultural understanding of their ELL cl ients and families through the effective use of three tools: 1) ethnographic interviews, which use open circumstances, perspectives, and goals; 2) skilled dialogue, a process that creates r espectful, reciprocal, and responsive interpersonal interactions including techniques for conflict resolution; and 3) collaboration with interpreters and translators, who can bridge remaining gaps in communication and understanding between the professional and the client/family. These practices can contribute to the creation of individual profiles of ELL students that extend beyond scores on standardized tests or language performance measures to shed light on the sociocultural and experiential factors that identities as bilingual or monolingual readers and writers. In the current study, the profile analysis delved deeper than the lexical, syntactic, and discourse features evident on the r their diverse writing abilities. From this perspective, a case that stands out is that of Manuel, the Mexican student in grade 8 who was excluded from the quantitative analysis because his writing
169 did not meet the productivity criteria 11 Notwithstanding his exclusion from the richness of the profile analysis and the final outcomes of the study (see General Discussion). Because Manuel struggled to write in both Spanish and English, he surfaced as a student with a possible, undiagnosed LI In this way, Manuel represented an outlier, or an extreme/unique case (Yin, 2003), which highlighted him both quantitatively and qualitatively as a student to further investigate. Per haps then, in addition to an examination of how students write, an exploration of what they write can provide educators and speech language pathologists with valuable entryways through which to better engage ELL students and meet their unique needs for dev elopment of academic language proficiency. Additionally, for ELL research, individual profiles may serve as a tool for differentiating sociocultural variables that facilitate or hinder L2 language and literacy learning. General Discussion This mixed metho ds investigation leads to several broad conclusions, as well as questions, regarding the bilingual writing of adolescent ELL students. Gutirrez and Orellana (2006) argued for the abandonment of deficit and difference frameworks for describing ELL students 11 The focal participants were randomly selected and interviewed during the data collection phase, so it was were too short to include in the quantitative analysis (criteria for analysis were at least 10 T units and/or 75 words).
170 provides a small scale attempt to broaden this dominant paradigm to a more inclusive framework and methodology that not only explores trends within and across ELL participants, but also widens and deepens our collective understanding of individual students through qualitative profile analysis. This methodology has the potential to enrich our understanding of bilingual language proficiency, as well as to aid in the development of more effective assessments and instructional strategies for ELL students. In addition, t With this in mind, the general discussion that follows focuses on two overarching themes that emerged from the quantitative and qualitative findings: 1) bilingual language proficiency is a dynamic, interactive system with relationships occurring at multiple levels both with in and across L1 and L2; and 2) how students construct and perceive their identities also impacts on how they approach L2 learning and related literacy practices. The presentation of these themes is followed by a discussion of future directions for research with adolescent ELLs. A Multifacet ed Look at Language Interaction The overall results of this study suggest that the ELL participants were able to apply skills from their however limited or expansive bilingual repertoires of academic (writing) skills to produce texts that were generally similar in both Spanish and Englis h. That is, the participants may have drawn from a more or less developed repertoire of
171 cross linguistic academic resources to employ in either Spanish or English writing. These resources might include both general cognitive and linguistic strategies as we ll as skills that incorporated vocabulary, syntax, and text related structures. Hence, students who had acquired the skills required to write in school in either Spanish or English, such as the ability to use more abstract nouns, construct syntactically co mplex sentences, and organize a genre appropriate text, were able to apply these skills in the other language as well. In contrast, and similar to the findings of Lanauze and Snow (1989), the students who wrote poorly in one language also did so in the ot her. The question that arises from these patterns for the adolescent ELL student is which aspects of academic language proficiency are shared across both languages, and, how can these be assessed and taught? Specifically, how can we provide adolescent ELLs with the metalinguistic awareness and strategies necessary to develop and access these skills across both L1 and L2 for reading and writing tasks? Finally, how might an integrated, cross language understanding of academic language proficiency support a mo re reliable identification of adolescent ELLs with a n undiagnosed LI? The results of this study point to interactive relationships among the lexical, linguistic re lationships among texts written in Spanish and English. These outcomes support a general interactive processing theory of language that recognizes bidirectional interfaces among areas of language (e.g., semantics, syntax, and discourse), cognitive
172 linguist ic domains (attention and memory), and, for bilingual individuals, L1 and L2 (Kohnert, 2008). In light of this theoretical framework, the findings additionally call for reflection on the classic, structuralist distinction of langue/parole (la nguage/speech ; de Saussure, 1959 ) better conceptualized in French as langue/langage or in Spanish, lenguaje/lengua (English does not have different words for these concepts). The first of these terms represents the human capacity for language and its cognitive underpin nings and the second refers to the manifestation of this capacity in a specific language such as English, Spanish, or French. In relation to language proficiency, Kohnert (2008) described this distinction as a language and proficiency in l ability in any specific lexical, syntactic, pragmatic, and discourse levels and their efficient use for varied and numerous communicative functions. On the o ther hand, proficiency in language, in motor, cognitive, and social abilities (Kohnert, 2008, p. 22). As an ELL student acquires L2, s/he will necessarily experience the challenges that result from developing proficiency in a second language, in this case, English. However, if a bilingual child has LI, general and systemic lags in coordinating various language levels (Silliman & Mody, 2008) will manifest themselves in both languages (i.e., Spanish and English). For both typically developing ELLs and those with language
173 learning difficulties, awareness of and access to general langu age abilities may serve as a key to unlock cross language aspects of academic language proficiency. Additionally, it is important to keep in mind that, for adolescent ELLs, aspects of academic language proficiency may vary across specific content areas. F or example, specialized vocabulary and morphosyntax characterize the discourse of science, social studies, mathematics, etc. (Bai ley & Butler, 2007; Janzen, 2008 ). This implies that language features unique to specific disciplines will require attention fo r the purposes of both assessment and instruction/intervention. The content challenge is yet another layer of the multileveled and integrated academic language system students must learn to manage to succeed in school. Language Identity and Academic Langu age Learning for ELLs Diversity and uniqueness. The adolescent ELL participants of this study varied widely in their language proficiencies as well as their language and literacy acquisition experiences and patterns of usage in Spanish and English. This di versity was explored more deeply through the individual profiles of the 6 focal participants who comprised the qualitative analysis. Additionally, as was demonstrated by the qualitative results, ELL bout language learning and may also impact on their writing performance. An example of this variation across all participants can be appreciated in that, while many of the students were born and educated outside of the U.S. and received no English languag e instruction through the elementary grades, others attended the primary grades (or even all grades) in the U.S. and acquired literacy first in English. Others still
174 grew up in Puerto Rico, where Spanish was their primary language, but they received Englis h language instruction throughout their elementary school years. Clearly, these students bring different language and literacy strengths and needs to the ELL classroom. All of these students had been designated as ELLs by the school system and received sp ecialized instruction to develop their English language proficiency. However, based on this variation, it is difficult to identify, for the total group of students, which language is the L1 and which is the L2. Similarly, when examining the content and str of potential language transfer is not clear: students may equally utilize known vocabulary, syntax, or text structure in Spanish to support their writing in Englis h, or vice versa. These differences also highlight the challenges faced by educators and other professionals who serve ELL students. The question becomes one of individual differences, emphasizing the need to determine effective and efficient methods to ca pture individual portraits of ELL language proficiency and more successfully support literacy development. Regularity and variance. The qualitative profiles illuminated the focal English) or monolingual (lacked the skills and/or interest to communicate in English). These identities, shaped by th eir previous language and literacy learning experiences and attitudes, also influenced their current language and literacy learning experiences and attitudes. The relationship between identity construction and literacy learning is further illustrated throu gh the emergence of patterns of regularity and variance (Gutirrez &
175 Orellana, 2006) among the adolescent ELL participants of this study. Generally speaking, these students might be viewed as emerging bilingual writers whose strengths and abilities varied depending on their experiences as well as the topic of the writing sample, taking into account the structure of the prompt itself. It has also been shown that the participants made use of cross linguistic, academic resources in their efforts to compose ex pository and narrative texts in Spanish and writing strategy (Bereiter & Scardamalia, 1987), resulted, for the most part, in a relatively regular performance across lang uages and genres in spite of the individual variance in the group. Considering these patterns of regularity and variance, and based on their scores on the various measures, the participants this study can be considered emerging bilingual writers whose wr iting proficiencies in both languages varied across a continuum from struggling in both languages to competent in Spanish and English. Along this continuum, three portraits of adolescent ELL writers were developed based on the quantitative and qualitative findings: 1) non emerging: ELL students who struggled and wrote poorly in both languages; 2) dominant emerging: participants whose writing was at least marginally proficient in either Spanish or English, but not both; and, 3) balanced emerging: students wh o demonstrated marginal to adequate proficiency equally across both languages. Each of these portraits has instructional implications related to the unique linguistic, social, and academic needs of the students who fit these profiles.
176 The non emerging bi lingual writer. The first portrait, non emerging, is embodied by students like Manuel, from Mexico, who self identified as a monolingual and claimed he would never get used to living in the United States. Manuel, a student in grade 8, expressed that learni ng English was very difficult and felt that it was like being born again because it was a different language. Manuel was not only was frustrated by the experience of learning English as a second language, but also lacked a solid foundation of academic lang uage skills in his first language, Spanish. As a result, Manuel was a struggling student and wrote poorly in both languages. This situation might also depict the experience of an ELL student with (undetected) LI. Students like Manuel, including ELLs with l anguage or learning disabilities, require extensive academic support to strengthen their general language and literacy abilities as well as to develop their English language proficiency. The former may be done in both languages to maximize success across b oth languages and in varied contexts (Kohnert, 2008). Regarding English language instruction, vocabulary development is the content areas (Janzen, 2008 ; Snow & Kim, 2007; Wilkinson & Silliman, 2008). Overall, ELL students who lack academic proficiency in both their first and second languages would benefit from explicit literacy instruction in both languages, including contrastive analysis experiences (Kohnert, 2008) to buil d metalinguistic awareness, emphasize connections between the languages, and strengthen common underlying proficiencies that can be applied to literacy tasks in either the first or second language.
177 The dominant emerging bilingual writer. The second portr ait, dominant emerging, is illustrated well by Edgar, from Mexico. Edgar was a student in grade 7 who self identified as a monolingual Spanish speaker. He expressed his disinterest and dislike of learning English at school, and had plans to return to Mexic o to attend university and law school. In contrast to Manuel, Edgar wrote relatively well in his first language, Spanish. He utilized abstract, metaphorical vocabulary and complex sentence constructions, particularly in his expository compositions. However due to his inexperience with -and perhaps also his negative feelings toward -English, Edgar was unable to transfer these skills to his writing in English. Students like Edgar, who come to the ELL classroom with some level of academic language proficienc y in their home language, require, in addition to acquisition of L2 vocabulary and sentence structure, metalinguistic strategies that will aid them in applying the skills they already possess to literacy tasks in their new language. Further, sociocultural factors will be important in encouraging dominant emerging ELL students to take risks and build confidence in their second language. Students like Edgar strongly identify with their home language and culture and may resist the second language and culture. For these ELLs, as others have recommended (e.g., Ball, 2006; Dworin, 2006; Flores Dueas, 2004; Moll et al., 1992; Moll et al. 2001), the incorporation of culturally meaningful opportunities to engage in literacy activities in a supportive, additive language learning environment.
178 The balanced emerging bilingual writer. Finally, the third writer portrait, balanced emerging, describes a student like Carolina, in grade 7. Carolin a attended the primary grades (K 2) in Kentucky and then continued her schooling in Puerto Rico before returning to Florida in grade 7. She also experienced continuous English language instruction at school in Puerto Rico. Carolina considered herself to be bilingual and felt proud of her skills. She demonstrated her bilingual identity through her attempts to mix both languages in her interview and journal entries (although Spanish was still dominant). In her formal writing samples, Carolina demonstrated pro ficiency across both Spanish and English texts, consistently scoring in the competent range on the CSE analytic measure. With regard to ELL instruction, students like Carolina are ready to be challenged with higher level academic language and literacy tas ks. In the classroom, these students can be encouraged to continue their expansion of both languages through development of more complex, literate vocabulary, sentence and text structures. A student like Carolina may also serve as a resource to other ELLs who are non emerging and dominant emerging, for example as a collaborator in a bilingual identity text (Cummins et al., 2005) or bilingual autobiography similar to what the participants developed for this study. These three portraits, which summarize the integration of the quantitative and qualitative findings of this research, offer an additional window through which to view the diverse identities and abilities of ELL students in middle school. Much work is still needed to better understand and meet the needs of adolescent ELL students, in particular
179 those who may have an undiagnosed language or learning impairment. As previously and create new lexical, morphosyntactic, and discourse measures to increase our understanding of ELL writing, as well as to advance our success in discovering the linguistic characteristics of ELL students who are in need of special education services. Research Agenda It is clear that there is a continued need for investigation in the area of ELL literacy, particularly with adolescent students, who have been largely overlooked in previous research (Short & Fitzsimmons, 2007). Gutirrez and Orellana (2006) called for a shift in ELL frameworks fro m those that paint these students with broad brush strokes to deeper and richer examinations of the diversity of ELL literacy practices in a variety of contexts. The conclusions of the present study align with this approach, highlighting the value of quali tative profile analysis of ELLs in conjunction with linguistic measures that take into account the numerous, interactive levels of a bilingual language system. More specifically, the results of this study bring attention to an u nresolved theoretical issue with clinical and educational implications; that is, which aspects of bilingual academic language proficiency are shared across both languages? From a clinical perspective, the question arises about whether an adolescent ELL student with LI may be more re liably identified through examining the interaction of various aspects of academic language proficiency in both L1 and L2, as compared to more traditional measures that evaluate specific linguistic features and/or assess only one language. In the education al realm, it will be important to address how adolescent ELLs might be taught
180 the necessary metalinguistic awareness to develop and access academic language knowledge to achieve literacy goals across both L1 and L2. Follow up studies The current findings consistent with previous research, suggest that micro level, text analyses of ELL writing may not be the most effective solutions to assess bilingual academic language proficiency. Instead, an integrated framework that recognizes language as an interacti ve system may prove more useful to shed light on the cross language aspects of the academic language proficiency of adolescent ELLs. Additionally, a mixed methods approach could increase opportunities to effectively identify adolescent ELLs with undiscover ed LI, such as the focal participant, Manuel, who, in grade 8, continued to struggle with literacy in both Spanish and English. An objective for future research, then, would be to apply integrated language measures to delve more deeply into the cross langu age aspects of ELL academic language proficiency, as well as to explore whether these integrated measures are more effective for this purpose than more traditional, micro level (or monolingual) measures such as NDW and MLT. One such integrated research app roach could extend the analysis of the present data set to explore the relationships between vocabulary depth and morphosyntactic awareness. For example, examination of the elaborated noun phrases used by the participants in both languages could provide no t only increased understanding of how more structurally complex forms in ways other than the embedding of dependent clauses. In addition to comparing outcomes ac ross languages, it would also be useful to contrast
181 the elaborated noun phrase findings with those from equivalent writing samples collected from monolingual English (and, ideally, also Spanish) speakers at the middle school level. Moving beyond the prese nt corpus of student writings, another study involving the crossover between vocabulary depth and morphosyntactic awareness could morphology. This might be explored by a apply semantic morphemes from L1 and L2 to change word meanings and/or create novel words that make sense in sentences. Performance could be compared across groups of ELL students with diverse language histories focusing on such variables as more or less time attending schools in the U.S. Alternately, and considering that bilingual/monolingual emerged as an identity variable for the focal participants in the present study, students could also be grouped based on t heir self identification as bilingual or monolingual (or, more or less proficient in L2). This type of semantic morpheme measure would address the question of cross language academic language proficiency in the area of morphosynta ctic awareness in relation to knowledge of derived words, which may be characteristic of more elaborate semantic networks. A second area of an integrated language paradigm warranting study involves the interactions between syntax and discourse and how these play out in bilingual a cademic language proficiency. Integrated measures in this area might employ larger units of text to examine the nature of syntax as it is situated in written discourse. An example of this type of measure is clause pac kaging (Berman & Nir Sagiv, 2005 ; Katze nberger, 2004). A
182 embedded unit of two or more clauses that are linked by syntactic criteria, but also taking into account thematic and & Nir Sagiv, 2005, p. 5). These authors used clause packages to ex plore differences between spoken and written texts and among different age groups of monolingual participants. A similar measure might be applied to the current data set to further compare and contrast discursive and sy n tactic patterns across L1 and L2 wri ting. Additionally outcomes could be compared with those of adolescent, monolingual speakers of English and Spanish to enhance our understanding of bilingual versus monolingual patterns in academic language proficiency. Finally, sentence combining, which has yet to be explored with ELLs at the middle school level, could also serve as a vehicle to achieve deeper insights into the cross language aspects of academic language proficiency at the syntactic discourse interface. One possibility is to employ an analytical approach used b y Scott and Nelson (2009) short scenarios consisting of simple one clause sentences into new stories that because principles of senten ce combining have been applied (e.g., deletion, insertion, subordination, coordination, etc.) languages as well as between groups distinguished by the variables of languag e experience or language identity. Outcomes might also be compared with those of syntactic proficiency in both languages and also may serve as a prospective approach to identify ELLs with LI.
183 Additionally, previous research has shown that sentence combining may be an effective instructional tool. In a meta analysis, Graham and Perin (2007) found a moderate effect size (0.50) for the impact of sentence combining as an in structional strategy to improve writing for English speaking adolescent students. Further, this instructional strategy could become more engaging to adolescent ELLs if it were applied in the context of culturally relevant, or autobiographical, texts that a cknowledged the Dueas, 2004). Instructional/intervention strategies Keeping in mind the sociocultural aspects of language and literacy learning, in addition to the research directions described above, t here is an additional need for instructional/intervention studies whose objective is to determine which strategies are most effective in promoting cross language academic language proficiency for adolescent ELLs. One possibility for a response to instructi on study involves the direct instruction of systematic, contrastive language analysis strategies in Spanish and English to an experimental group as compared with a control group receiving instruction/intervention in English only. Some such contrastive stra tegies include direct instruction in Spanish versus English inflectional and derivational morphemes, use of compound words and conjunctions, sentence word order, and use of prepositions, including in phrasal verbs (see Kohnert & Derr, 2004 for a summary of non overlapping features of Spanish and English). An exploratory investigation of this kind might be structured as a clinical intervention study involving an individual or small group of adolescent ELL students diagnosed with LI. These students would be evaluated before and after receiving the
184 contrastive analysis intervention in order to determine how this type of treatment impacted on their metalinguistic awareness in both languages. They could also be compared with a control group of ELL students with curriculum. At the moment, there remain many unanswered questions regarding best practices in adolescent ELL instruction. This is particularly the case when individual differences, including LI, are taken into account. F inal Thoughts This mixed methods study has explored expository and narrative writing in Spanish and English of ELLs in middle school. A quantitative examination of these re topics has provided a glimpse into the multiple factors that take part in ELL writing development. Namely, the significance of the role of the writing topic and prompt, the potential transfer of academic language proficiency across languages, and the in tegration of vocabulary, grammar, and text level skills stand out as key outcomes to glean from this study. Beyond these quantitative patterns, the qualitative profile analysis of 6 focal participants has highlighted additional factors that play a role in the orchestration of bilingual writing. Such elements include language choice, reaction to language prejudice, identity as bilingual or monolingual, and ethnic identity. While it is not possible, given the scope of this investigation, to delve into each o f these aspects in depth, it is critical to recognize the value of individual profile analysis for these types of students.
185 throughout the bilingual writing samples as well as in th e journal content and interview discourse of the ELL students who participated in this study. A common challenge to the educator or speech language pathologist working with ELL students is the variation in experiences and abilities of these students. Perha ps a deeper exploration into this diversity, integrated with big picture patterns of ELL language and literacy development, will offer some solutions to overcome these challenges.
186 References Alarcos Llorach, E. (1996). Gramtica de la lengua espaola. Madrid, Spain: Espasa Calpe. Anfara, Jr., V. A., Brown, K. M., & Mangione, T. L. (2002). Qualitative analysis on stage: Making the research process more public. Educational Researcher, 31 28 38. Alliance for Excellent Education (2007). Federal support for adolescent literacy: A solid investment. Issue Brief retrieved July 6, 2007 from http://www.all4ed.org/publications/FedAdLit.pdf Ariel, M. (2009). Discourse, grammar, discourse. Discourse Studies, 11, 5 36. Au, K. H. P., & Mason, J. M. (1981). Social organizational factors of learning to read: The balance of rights hypothesis. Reading Research Quarterly, 17 115 152. Bakeman, R., & Gottman, J. M. (1986). Observing interaction: An introduction to sequential analy sis. Cambridge, MA: Cambridge University Press. Ball, A. F. (2006). Teaching writing in culturally diverse classrooms. In C. A. MacArthur, S. Graham, & J. Fitzgerald (Eds.), Handbook of writing research (pp. 293 310) New York: Guilford Press. Bailey, A L., & Butler, F. A. (2007). A conceptual framework of academic English language for broad application to education. In A. L. Bailey (Ed.), The language demands of school: Putting academic English to the test (pp. 68 102). New Haven, CT: Yale University Press.
187 Beck, I. L., McKeown, M. G., and Kucan, L. (2008). Creating robust vocabulary: Frequently asked questions and extended examples New York: Guilford Press. Beck, S. W., & Jeffery, J. V. (2009). Genre and thinking in academic writing tasks. Journal of Literacy Research, 41 228 272. Beers, S. F., & Nagy, W. E. (2009). Syntactic complexity as a predictor of adolescent writing quality: Which measures? Which genre? Reading and Writing, 22 185 200. Bejarano, C. L. (2005). Qu onda? Urban youth culture and border identity Tucson: University of Arizona Press. Berdan, K., Boulton, I., Eidman Aadahl, E., Fleming, J., Gardner, L., Rogers, I., & Solomon, A. (2006). Writing for a change: Bo o sting literacy and learning through social action. San Francisco, CA: John Wiley & Sons. Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Hillsdale, NJ: Erlbaum. Berman, R. A. (2008). The psycholinguistics of developing text construction. Journal of Child Language, 35 735 771. Berman, R. A., & Nir Sagiv, B. (2007). Comparing narrative and expository text construction across adolescence: A developmental paradox. Discourse Processes, 43 79 120. Berman, R. A., & Nir Sagiv, B. (2005). Clause packaging in narratives: A cross linguistic developmental study. Unpublished manuscript.
188 Berman, R. A., & Verhoeven, L. (2002). Cross linguistic perspectives on the development of text production abilities: Speech and writing. Written Language and Literacy, 5 1 43. Bernard, H. R. (2 002). Research methods in anthropology: Qualitative and quantitative approaches, 3 rd Edition. Walnut Creek, CA: AltaMira Press. Berninger, V. W., & Hooper, S. (2003). A developmental neuropsychological perspective on writing disabilities in children and youth. Unpublished manuscript. Berninger, V. W, Vaughan, K., Abbott, R. D., Begay, K., Coleman, K. B., Curtin, G., Hawkins, J. M., & Grahan, S. (2002). Teaching spelling and composition alone and together: Implications for the simple view of writing. Jou rnal of Educational Psychology, 94 291 304. Bialystok, E. (2007). Acquisition of literacy in bilingual children: A framework for research. Language Learning, 57 45 77. Bialystok, E., Luk, G., & Kwan, E. (2005). Bilingualism, biliteracy, and learning to read: Interactions among languages and writing systems. Scientific Studies of Reading, 9 43 61. Biancarosa, G., & Snow, C. (2004). Reading next: A vision for action and research in middle and high school literacy (2 nd Ed.). Report by Alliance for Excel lent Education, Carnegie Corporation of New York.
189 Bloome, D., Power Carter, S., Morton Christian, B., Otto, S., & Shuart Faris, N. (2005). Discourse analysis and the study of classroom language and literacy events: A microethnographic perspective. Mahwa h, NJ: Lawrence Erlbaum. Boyd, F., & Brock, C. H. (2004). Constructing pedagogies of empowerment in multicultural and multilingual classrooms: Implications for theory and practice. In F. Boyd & C. H. Brock (Eds.). Multicultural and multilingual literacy a nd language: Contexts and practices (pp. 1 14) New York: Guilford Press. Brantlinger, E., Jimenez, R., Klingner, J., Fugach, M., & Richardson, V. (2005). Qualitative studies in special education. Exceptional Children, 71 195 207. Briggs, C. (1986). Learning how to ask: A sociolinguistic appraisal of the role of the interview in social science research. Cambridge: Cambridge University Press. Brinton, B., & Fujiki, M. (2004). Social and affective factors in children with language impairment: Implicat ions for literacy learning. In C. A. Stone, E. R. Silliman, B. J. Ehren, and K. Apel (Eds.), Handbook of language and literacy: development and disorders (pp. 130 153). New York: Guilford Press. Carvalho, A. M., da Silva, A. J. B. (Fall 2008). O papel do c onhecimento metalingstico nos padres de transferncia no desenvolvimento da interlingua e suas implicaes pedaggicas [The role of metalinguistic knowledge on patterns of transfer in the development of interlanguage and its instructional implications] Portuguese Language Journal, 3. Retrieved March 14, 2009, from http://www.latam.ufl.edu/portugueselanguagejournal/articles.html.
190 Christian, B., & Bloome, D. (2004). Learning to read is who you are. Reading and Writing Quarterly, 20 365 384. Cobo Lewis A. B., Pearson, P. Z., Eilers, R. E., & Umbel, V. C. (2002). Effects of bilingualism and bilingual education on oral and written Spanish skills: A multifactor study of standardized test outcomes. In D. K. Oller & R. E. Eilers (Eds.), Language and litera cy in bilingual children (pp. 98 117). Clevedon, UK: Multilingual Matters. Collins, J., & Blot, R. K. (2003). Literacy and literacies: Texts, power, and identity. Cambridge: Cambridge University Press. Creswell, J. W. & Plano Clark, V. L. (2007). Designing and conducting mixed methods research. Thousand Oaks, CA: Sage. Cronnell, B. (1985). Language influences in the English writing of third and sixth grade Mexican American students. Journal of Educational Research, 78, 168 173. Cummins, J. (20 00). Language, power, and pedagogy: Bilingual children in the crossfire Buffalo, NY: Multilingual Matters. Cummins, J. (1991). Interdependence of first and second language proficiency in bilingual children. In E. Bialystok (Ed.), Language processing in bilingual children (pp. 70 88). New York: Cambridge University Press. Cummins, J. (1984). Wanted: A theoretical framework for relating language proficiency to academic achievement among bilingual students. In C. Rivera (Ed.), Language proficiency and aca demic achievement Clevedon, England: Multilingual Matters.
191 Cummins, J. (1979). Linguistic interdependence and the educational development of bilingual children. Review of Educational Research, 49 222 251. Cummins, J., Bismilla, V., Chow, P., Cohen, S ., Giampapa, F., Leoni, L., Sandhu, P., & Sastri, P. (2005). Affirming identity in multilingual classrooms. Educational Leadership, 63 (1), 38 43. Cummins, J., & Schecter, S. R. (2003). Introduction: School based language policy in culturally diverse cont exts. In S. R. Schecter & J. Cummins (Eds.) Multilingual education in practice: Using diversity as a resource (pp. 1 16). Portsmouth, NH: Heinemann. Daisey & Jos Kampfner (2002). The power of story to expand possible selves for Latina middle school stude nts. Journal of Adolescent & Adult Literacy, 45 578 587. Danzak, R. L., & Silliman, E. R. (2005). Does my identity speak English? A pragmatic approach to the social world of an English language learner with language impairment. Seminars in Speech and Language, 26 189 200. DaSilva Iddings, A. C., & Katz, L. (2007). Integrating home and school identities of recent immigrant Hispanic English language learners through classroom practice. Journal of Language, Identity, and Education, 6 299 314. De Fina A. (2006). Group identity, narrative, and self representations. In A. De Fina, D. Schiffrin, & M. Bamberg (Eds.), Discourse and identity (pp. 351 375). Cambridge: Cambridge University Press. De La Pea, R. A. (1999). Gramtica terica y prctica de la lengua castellana. Mexico City: Universidad Nacional Autnoma de Mxico.
192 De Saussure, F. (1959). Course in general linguistics (3 rd Ed., W. Baskin, Trans.). New York: McGraw Hill (Original work published 1915). Dressl er, C., & Kamil, M. L. (2006). First and second language literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second language learners: Report of the National Literacy Panel on Language Minority Children and Youth (pp. 197 238). Mahwah, NJ : Lawrence Erlbaum. Durgunoglu, A. Y. (2002). Cross linguistic transfer in literacy development and implications for language learners. Annals of Dyslexia, 52, 189 204. Dworin, J. E. (2006). The family stories project: Using funds of knowledge for writin g. The Reading Teacher, 59 510 520. Edelsky, C. (1982). Writing in a bilingual program: The relation of L1 and L2 texts. Tesol Quarterly, 16 211 228. Escamilla, K. (2006). Semilingualism applied to the literacy behaviors of Spanish speaking emerging bilinguals: Bi illiteracy or emerging biliteracy? Teachers College Record, 108 2329 2353. Ferreiro, E., & Teberosky, A. (1982). Literacy before schooling Portsmouth, NH: Heinemann. Fey, M. E., Catts, H. C., Proctor Williams, K., Tomblin, J. B., & Zhang X. (2004). Oral and written story composition skills of children with language impairment. Journal of Speech, Language, and Hearing Research, 47, 1301 1318.
193 Fitzgerald, J. (2006). Multilingual writing in preschool through 12 th grade: The last 15 years. In C. A. MacArthur, S. Graham, & J. Fitzgerald (Eds.), Handbook of writing research (pp. 337 354) New York: Guilford Press. Flores Dueas, L. (2004). Reader response, culturally f amiliar literature, and reading comprehension: The case of four Lat ina(o) students. In F. Boyd & C. H. Brock (Eds.). Multicultural and multilingual literacy and language: Contexts and practices (pp. 180 206) New York: Guilford Press. Francis, D. J., Lesaux, N., & August, D. (2006). Language of instruction. In D. August & T. Shanahan (Eds.), Developing literacy in second language learners: Report of the National Literacy Panel on Language Minority Children and Youth (pp. 365 413). Mahwah, NJ: Lawrence Erlbaum. Francis, N. (2006). The development of secondary discourse ab ility and metalinguistic awareness in second language learners. International Journal of Applied Linguistics 16 37 60. Gee, J. P. (2004). Situated language and learning: A critique of traditional schooling. New York: Routledge. Gee, J. P. (1996). Social linguistics and literacies: Ideology in discourses. London: Routledge Falmer. Gentry, J. R. (1982). An analysis of developmental spelling in GNYS at WRK. The Reading Teacher, 36 192 200. Gentry, J. R. (2000). A retrospective on invented spellin g and a look forward. The Reading Teacher, 54 318 332.
194 Geva, E. (2006). Second language oral proficiency and second language literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second language learners: Report of the National Literacy P anel on Language Minority Children and Youth (pp. 123 140). Mahwah, NJ: Lawrence Erlbaum. Geva, E., & Genesse, F. (2006). First language oral proficiency and second language literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second langu age learners: Report of the National Literacy Panel on Language Minority Children and Youth (pp. 185 196). Mahwah, NJ: Lawrence Erlbaum. Gili Gaya, S. (1972). Curso superior de sintaxis espaola. Barcelona, Spain: Bibliograf. Goldenberg, C., Rueda, R. S., & August, D. (2006). Sociocultural influences on the literacy attainment of language minority children and youth. In D. August & T. Shanahan (Eds.), Developing literacy in second language learners: Report of the National Literacy Panel on Language Minori ty Children and Youth (pp. 269 318). Mahwah, NJ: Lawrence Erlbaum. Graham, S. & Perin, D. (2007). Writing next: Effective strategies to improve writing of adolescents in middle and high schools. Report by Alliance for Excellent Education, Carnegie Corpor ation of New York. Grosjean, F. (1998). Studying bilinguals: Methodological and conceptual issues. Bilingualism: Language and Cognition, 1 131 149. Guba, E. G., & Lincoln, Y. S. (1989). Fourth generation evaluation. Newbury Park, CA: Sage.
195 Gutirrez, K. D., & Orellana, M. F. (2006). Constructing genres of difference. Research in the Teaching of English, 40, 502 507. Gutirrez, K., Asato, J., Pacheco, M., Moll, L., Olson, K., Horno, E., Ruiz, R., Garcia, E., on English language learners. Reading Research Quarterly, 37 328 343. Gutirrez Clellen, V. F. (2002). Narratives in two languages: Assessing performance of bilingual children. Linguistics and Education, 13, 175 197. Gutirrez Clellen, V. F. (1999). Mediating literacy skills in Spanish speaking children with special needs. Language, Speech, and Hearing Services in Schools 30 285 292. Gutirrez Clellen, V. F ., & Hofstetter, R. (1994). Syntactic complexity in Spanish narratives: A developmental study. Journal of Speech and Hearing Research, 37 645 654. Harley, B., & King, M. L. (1989). Verb lexis in the written compositions of young L2 learners. Studies in Second Language Acquisition, 11 415 439. Hayes, J. R., & Flower, L. S. (1980). Identifying the organization of writing processes. In L. W. Gregg & E. R. Steinberg (Eds.), Cognitive processes in writing (pp. 3 30). Hillsdale, NJ: Erlbaum. Hewitt, L. E. Hammer, C. S., Yont, K. M., & Tomblin, J. B. (2005). Language sampling for kindergarten children with and without SLI: Mean length of utterance, IPSYN and NDW. Journal of Communication Disorders, 38 197 213.
196 Hunt, K. W. (1965). Grammatical structures w ritten at three grade levels. NCTE Research Report No. 3. Champaign, IL: National Council of Teachers of English. Inspiration (Version 8) [Computer software] (20 08). Beaverton, OR: Inspiration Software, Inc. Janzen, J. (2008). Teaching English language learners in the content areas. Review of Educational Research, 78 1010 1038. Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed methods research: A research paradigm whose time has come. Educational Researcher, 33 14 26. Johnson, B. & Turner, L. A. (2003). Data collection strategies in mixed methods research. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 297 320) Thousand Oaks, CA: Sage. Kameen, P. T. (1979). Syntactic skill an d ESL writing quality. learner in focus; selected papers from the thirteenth annual convention of Teachers of English to Speakers of Other Languages, 343 364. Katzenberger, I. (2004). The development of clause packaging in spoken and written texts. Journal of Pragmatics, 36 1921 1948. Koda, K. (2007). Reading and language learning: Crosslinguistic constraints on second language reading development. Language Learning, 57 1 44. Koelsch, N. (2006) Improving literacy outcomes for English language learners in high school: Considerations for states and districts in developing a coherent policy framework. Research brief, National High School Center, retrieved June 12, 2006
197 from http://www.betterhighschools.org/docs/NHSC_ImprovingLiteracy_010907.pdf Kohnert, K. (2008). Language disorders in bilingual children and adults San Diego: Plural Publishing. Kohnert, K. & Derr, A. (2004). Language intervention with bili ngual children. In B. A. Goldstein (Ed.), Bilingual language development and disorders in Spanish English speakers (pp. 311 338). Baltimore: Paul H. Brookes Publishing. Lafrance, A., & Gottardo, A. (2005). A longitudinal study of phonological processing skills and reading in bilingual children. Applied Psycholinguistics, 26 559 578. Lanauze, M., & Snow, C. (1989). The relation between first and second language writing skills: Evidence from Puerto Rican elementary school children in bilingual programs. Linguistics and Education, 1, 323 339. Language Assessment Scales Reading/Writing (LAS R/W) (1988 ). CTB/McGraw Hill. Lyons, J. (1977). Semantics, Volume II Cambridge, MA: CUP. Maxwell, J. A. (2005). Qualitative research design: An interactive approach (2nd. ed.). Newbury Park, CA: Sage. A. MacArthur, S. Graham, & J. Fitzgerald (Eds.). Handbook of writing research (pp. 115 130). New York, NY: Guilford Press. McHatt on, P. A., Shaunessy, E., Hughes, C., Brice, A., & Ratliff, M. A. (2007). You gotta represent! Ethnic identity development among Hispanic adolescents. Multicultural Perspectives, 9 12 20.
198 Meltzer, J., & Hamann, E. T. (2004). Meeting the literacy develop ment needs of adolescent English language learners through content area learning part one: Focus on motivation and engagement. Providence, RI: Brown University, The Education Alliance. Miles, M. B., & Huberman, A L. (1994). Qualitative data analysis: An expanded sourcebook (2 nd ed.). Thousand Oaks, CA: Sage Publications. Miller, J. F., Heilmann, J., Nockerts, A., Iglesias, A., Fabiano, L., & Francis, D. J. (2006). Oral language and reading in bilingual children. Learning Disabilities Research and Practice 21 30 43. Moll, L. C. (1986). Writing as communication: Creating strategic learning environments for students. Theory into Practice, 25 102 108. Moll, L. C., Amanti, C., Neff, D., & Gonzlez, N. (1992). Funds of knowledge for teachers: Using a qualitative approach to connect homes and classrooms. Theory into Practice, 31 132 141. Moll, L. C., Sez, R., & Dworin, J. (2001). Exploring biliteracy: Two student case examples of writing as a social practice. The Elementary School Journal, 101 4 35 449. Nelson, N. W., & Van Meter, A. M. (2007). Measuring written language ability in narrative samples. Reading and Writing Quarterly, 23 287 309. Nieto, S. (2002). Language, culture, and teaching: Critical perspectives for a new century. Mahwah, N J: Lawrence Erlbaum.
199 Norton Peirce, B. (1995). Social identity, investment, and language learning. TESOL Quarterly, 29 9 31. Onwuegbuzie, A. J., & Leech, N. L. (2007). Validity and qualitative research: An oxymoron? Quality & Quantity: International Journal of Methodology, 41 233 249. findings: The role of mixed methods research. The Qualitative Report, 9 770 792. Onwuegbuzie, A. J., & Teddlie, C. (2003). A frame work for analyzing data in mixed methods research. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 351 383) Thousand Oaks, CA: Sage. Ordez, C. L., Carlo, M. S., Snow, C. E., & McLaughlin, B. (2002) Depth and breadth of vocabulary in two languages: Which vocabulary skills transfer? Journal of Educational Psychology, 94 719 728. Paradis, J., & Crago, M. (2000). Tense and temporality: A comparison between children learning a second language and ch ildren with SLI. Journal of Speech, Language, and Hearing Research, 43 834 848. Paradis, J., Crago, M., Genesee, F., & Rice, M. (2003). French English bilingual children with SLI: How do they compare with their monolingual peers? Journal of Speech, Language, and Hearing Research, 46 113 127.
200 Perera, K. (1984). Oxford: Basil Blackwell. Prez, B. (2004). Becoming biliterate: A study of two way bilingual immersion education. Mahwah, NJ: Lawrence Erlbaum. Perkins, K. (1980). Using objective methods of attained writing proficiency to discriminate among holistic evaluations. TESOL Quarterly, 14 61 69. Prior, P. (2006). A sociocultural theory of writing. In C. A. MacArthur, S. Graham, & J. Fitzgerald (Eds.). Handbook of writing research (pp. 54 66). New York, NY: Guilford Press. Quellmalz, E. S. (1982). Designing writing assessments: Balancing fairness, utility, and cost (CSE Report No. 188). Center for the Study of Evaluation, Los Angel es, CA: University of California at Los Angeles, Graduate School of Education. Quellmalz, E. S. (1980). Problems in stabilizing the judgment process (CSE Report No. 136). Center for the Study of Evaluation, Los Angeles, CA: University of California at Lo s Angeles, Graduate School of Education. Quellmalz, E. S. (1979). Domain referenced specifications for writing proficiency (CSE Report No. 127). Center for the Study of Evaluation, Los Angeles, CA: University of California at Los Angeles, Graduate School of Education. and narrative writing skills. CSE Resource Paper No. 5. Center for the Study of Evaluation, UCLA Graduate School of Education. Los Angeles, CA.
201 Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive grammar of the English language. Essex, England: Longmann. Ravid, D. (2006). Semantic development in textual contexts during the school years: Noun scale analyses. Journal of Child La nguage, 22 791 821. Ravid, D., & Berman, R. A. (2006). Information density in the development of spoken and written narratives in English and Hebrew. Discourse Processes, 41, 117 149. Riley, P. (2006). Self expression and the negotiation of identity in a foreign language. International Journal of Applied Linguistics, 16 295 318. Rosenblatt L. M. (1978). The reader, the text, and the poem. Carbondale, Southern Illinois University Press. Rubin, R., & Carlan, V. G. (2005). literacy development. The Reading Teacher, 58 728 739. Rueda, R. S., August, D., & Goldenberg, C. (2006). The sociocultural context in which children acquire literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second language learners: Report of the National Literacy Panel on Language Minority Children and Youth (pp. 319 340). Mahwah, NJ: Lawrence Erlbaum. Ruiz, N. T. (1995). The social construction of ability and disability: I. Profile t ypes of Latino children identified as language learning disabled. Journal of Learning Disabilities, 28 476 490. Saddler, B, & Preschern, J. (2007). Improving sentence writing ability through sentence combining practice. Teaching Exceptional Children, 39 6 11.
202 Salahu Din, D., Persky, H., & Miller, J. (2008). (NCES 2008 468). National Center for Education Statistics, Inst itute of Education Sciences, U.S. Department of Education, Washington, D. C. Schriver, K. (1 990). Evaluating text quality: The continuum from text focused to reader focused methods. Technical Report No. 41, National Center for the Study of Writing. Educat ional Psychologist, 35 51 62. Scott, C. & Nelson, N. W. (2009). Sentence combining: Assessment and intervention perspectives. Perspectives on Language Learning and Education, ??? Sfard, A., & Prusak, A. (2005). Telling identities: In search of an analy tic tool for investigating learning as a culturally shaped activity. Educational Researcher, 34 (4), 14 22. Shanahan, T., & Beck, I. L. (2006). Effective literacy teaching for English language learners. In D. August & T. Shanahan (Eds.), Developing liter acy in second language learners: Report of the National Literacy Panel on Language Minority Children and Youth (pp. 415 488). Mahwah, NJ: Lawrence Erlbaum. Short, D. J., & Fitzsimmons, S. (2007). Double the work: Challenges and solutions to acquiring lan guage and academic literacy for adolescent English language learners A report to the Carnegie Corporation of New York Washington, DC: Alliance for Excellent Education.
203 Siegel, S. & Castellan Jr., N. J. (1988). Nonparametric statistics for the behavioral sciences, 2 nd Edition. Boston, MA: McGraw Hill. Silliman, E. R., & Mody, M. (2008). Individual differences in oral language and reading: Brain, behavior, and learning in language and reading disorders (349 386). New York: Guilford Snow, C. E., & Kim, Y. S. (2007). Large problem spaces: The challenge of vocabulary for English language learners. In R. K. Wagner, A. E. Muse, & K. R. Tannenbaum (Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 123 139). New York, NY: Guilford Press. Spradley, J. P. (1979). The ethnographic interview Fort Worth, TX: Holt, Rinehart & Winston. SPSS Statistics (Version 17.0) [Computer software] (2008). Chicago, IL: SPSS, Inc. Stake, R. E. (1995). The art of case study research Thousand Oaks, CA: Sage. Torrance, M., & Galbraith, D. (2006). The processing demands of writing. In C. A. MacArthur, S. Graham, & J. Fitzgerald (Eds.). Handbook of writ ing research (pp. 67 80). New York, NY: Guilford Press. University of Wisconsin (2008). Systematic Analysis of Language Transcripts [Computer software]. Madison, WI: SALT Software, LLC. Wagner, R. K., Francis, D. J., & Morris, R. D. (2005). Identifying E nglish language learners with learning disabilities: Key challenges and possible approaches. Learning Disabilities Research & Practice, 20 6 15.
204 Walqui, A. (2007, October). Apprenticing the academic uses of language: Middle school English language lear ners and their teachers. Paper presented at the inaugural meeting of the Center for Research on the Educational Achievement & Teaching of English Language Learners (CREATE), Oak Brook, IL. Wilkinson, L. C., & Silliman, E. R. (2008). Academic language prof iciency and literacy instruction in urban settings. In L. C. Wilkinson, L. M. Morrow, & V. Chou (Eds.), Improving literacy achievement in urban schools: Critical elements in teacher preparation (pp. 121 142) Newark, DE: International Reading Association. XSight (Version 2) [Computer software] (2006). Doncaster, Victoria, Australia: QSR International Pty Ltd. Yin, R. K. (2003). Case study research: Design and methods 3 rd Ed. Thousand Oaks, CA: Sage.
206 Appendix A Summary of Data from Informal ESL Teacher Interviews The foll owing table summarizes the observations obtained from informal phone interviews with middle and high school ESL teachers. The average phone conversation lasted 30 45 minutes. Main topics addressed included: a) stude nt demographics b) how they taught writing, c) how their students develop ed English writing proficiency, and d) what they recommend ed for the present study. Note: All names have been changed. Table 14 Summary of ESOL Teacher Interview Data Teacher, date of conversation Characteristics of students Writing instruction Advice on elicitation of writing samples Ms. Brady, Middle School, 7/17/07 Majority Spanish L1: N = ~38 40. Grade 6: ~25; Grade 7: ~13; Grade 8: ~13. Most students are from same town in Mexico. Journaling, daily bell work stresses writing, writing in L1 and L2, stresses transfer of mechanics, gives guidance. Uses literature to inspire. Writing topics: Hometow n stories? No most kids from same area, will not be interesting. Border crossing? Yes. She did autobiography project and students shared a lot.
207 Teacher, date of conversation Characteristics of students Writing instruction Advice on elicitation of writing samples Ms. Michaels, Middle school, 9/5/07 Speakers of various languages. Spanish L1: About 25 in grades 6 8, many new arrivals (last school year only had about 10). Students write independently; she conferences with proficient enough for Encourages use of thesaurus, focuses on building vocabulary. Uses Florida Writes outline format for narrative. Planning and prewriting are import ant. Biggest problem for ELL writing is lack of They will need SPECIFIC instructions of what we are looking for (e.g. description, interpretation, etc.). Only higher level students would be able to wr ite a narrative. Recommended not using computer to write students are more proficient by hand.
208 Teacher, date of conversation Characteristics of students Writing instruction Advice on elicitation of writing samples Ms. Thomas, High school, 9/12/07 Majority Spanish L1: N = 70 in grades 9 12. Origin: Mexico, South America, Central America, Caribean Structured, systematic, process oriented. Writing is so complex, many layers ta kes time to build this. Does not see evidence of transfer of higher level skills (more likely with phonetic/lexical items). Writing topics: Border crossing: No, too problematic. Her kids like to write about friends, families, here vs. there, being homes ick, going back to visit, differences in themselves from when they arrived until now, the day they left their home country. Abilities vary greatly, SES/educational opportunities in home country play a big role.
209 Appendix B Summary of Results of Participant Questionnaire Table 15 Personal and Educational Characteristics of Participants Participant Gender Grade level Age Place of birth Birthplace of mother Birthplace of father Grades attended in U.S. Grades attended outside U.S. Country of schooling 1 M 6 13 U.S. Mexico Mexico K 1, 5 6 2 5 Mexico 2 M 6 12 U.S. Honduras Mexico 5 6 1 4 Mexico 3 M 6 12 Mexico Mexico Mexico 2 6 1 Mexico 4 M 6 13 Mexico Mexico Mexico 1 6 K Mexico 5 M 6 12 U.S. Mexico El Salvador 5 6 K 5 Mexico 6 F 6 12 Mexico Mexico Mexico 2 6 K 2 Mexico 7 F 6 12 U.S. Mexico Mexico 1 6 K Mexico 8 M 6 11 U.S. PR 12 PR 5 6 K 5 PR 9 M 7 13 Mexico Mexico Mexico 5 7 1 5 Mexico 10 F 7 12 Mexico Mexico Mexico K 1, 4 7 2 3 Mexico 11 F 7 12 U.S. Mexico Mexico K 7 None N/A 12 F 7 13 U.S. PR PR K 2, 7 2 6 PR 12 PR = Puerto Rico
210 Participant Gender Grade level Age Place of birth Birthplace of mother Birthplace of father Grades attended in U.S. Grades attended outside U.S. Country of schooling 13 M 8 13 Mexico Mexico Mexico 4 8 K 4 Mexico 14 M 8 14 Mexico Mexico Mexico 4 8 1 3 Mexico 15 F 8 13 Mexico Mexico Mexico 6 8 1 5 Mexico 16 F 8 14 U.S. Mexico Mexico 2 8 K 1 Mexico 17 F 8 14 Mexico Mexico Mexico 4 8 K 3 Mexico 18 M 8 13 PR PR PR 7 8 K 6 PR 19 M 8 13 PR PR PR 7 8 K 6 PR 20 F 8 14 Dom. Rep Dominican Republic Dominican Republic 7 8 1 6 Dominican Republic
211 Appendix C Writing Prompts for Formal Samples and J ournals Formal samples Narrative 1 Special or funny family memory (Family) a. Remember something funny or special that happened in your family. b. Tell the story of what happened. i. Who was involved? ii. Where were you? iii. What happened? c. How did you feel at the time? d. Why is this a funny or special memory for you? Un recuerdo especial o chi stoso con la familia (Familia) a. Recuerda un evento chisoto o espec ial que pasaste con tu familia. b. Cuenta la historia del evento. i. Quin estaba? ii. Dnde estaban ustedes? c. Cmo te sentist e en este momento? d. Por qu este evento fue chistoso o especial?
212 Expository 1 (description) A person I admire (Family/friends) a. We all admire people for different reasons. b. Whom do you admire? (It can be someone in your family, a friend, professional, or celebrity). Describe this person with lots of details. i. What does this person do? ii. What makes him/her special? iii. Why do you admire him/her? c. If you could spend a day with this person, what would you do? Una persona que yo admiro (Familia/amigos) a. Todos admiramos a alguien por alguna razn. b. A qu in admiras? (Puede ser alguien en tu familia, un amigo, un profesional, o una persona famosa). Describe esta persona con muchos detalles. i. Qu hace esta persona? ii. Qu tiene de especial? iii. Por qu admiras a esta persona? c. Si pudieras pasar un da con esta p ersona, qu haran?
213 Narrative 2 First day of school in the U.S. (or First day of middle school) ( School) a. Think back and remember your first day of school in the US. b. Tell the story of what happened that day. i. What did you do? ii. Whom did you meet? c. How did you feel throughout the day? d. What did you learn that day about yourself and life at your new school? Primer da de clases en los EE.UU. (o Primer da de la escuela media) (La Escuela) a. Acurdate de tu primer d a de escuela aqu en los EEUU. b. Cuent a la h istoria de lo que pas ese da. i. Qu hiciste? ii. A quin conociste? c. C mo te sentiste durante ese da? d. Ese da, qu aprendiste sobre ti misma y sob re la vida en tu escuela nueva? Expository 2 (compare/contrast) Letter to a new student (School) a. Now that you have been here for a while, you can help a new student coming to this school from your home country.
214 b. Write a letter to this student explaining what it is like here. How is this school the same/different from school in your home country? i. What ar e the students and teachers like? ii. What kinds of things do you do with your friends? c. What advice would you give to this new student about starting school here? Carta a un estudiante nuevo (La escuela) a. Ya que llevas tiempo aqu en este pas, puedes ayudar a un estudiante nuevo que viene a esta escuela de tu pas natal. b. Escribe una carta a dicho estudiante nuevo, explicndole cmo es aqu. Cmo es esta escuela igual o diferente a la esc uela en tu pas natal? i. Cmo son lo s profesores y los estudiantes? ii. Qu haces con tus amigos? c. Cmo aconsejaras a este estudiante nuevo en cuan to a asistir a la escuela aqu? Journals (English versions) Journal 1 (Expository describe) Intro to journal a. Pretend your journal is a new friend you just met.
215 b. Write a letter to yo ur journal introducing yourself. Tell your journal about yourself. Give an overview of your family, school, friends, hobbies, and goals. We will expand on these topics in other journal entries. c. What are your expectations for the autobiography project? Jo urnal 2 (Narrative) Happy moments a. Remember a very happy moment in your life. b. Tell the story of this happy time. i. Who were you with? ii. Where were you? iii. What happened? c. How did you feel at the time? d. Why is this moment a special memory for you? Journal 3 ( Narrative) Sad moments a. Remember a very sad moment in your life. b. Tell the story of this sad time. i. Who were you with? ii. Where were you? iii. What happened? c. How did you feel at the time? d. What did you learn from this sad experience?
216 Journal 4 (Narrative) Problem or conflict a. Remember a time when you had a problem/conflict with a friend or family member. b. Tell the story about this problem or conflict. i. What happened? ii. What was the problem/conflict? iii. How did you resolve this conflict or solve the problem? c. H ow did you feel during this time? d. What did you learn from this experience? Journal 5 (Expository compare/contrast) The languages we speak a. Everyone in this class is bilingual we speak more than one language. b. Compare and contrast the language/s you an d your family speak. i. With whom do you speak each language? ii. When, under what circumstances do you speak each language? c. Have you ever experienced prejudice for speaking a minority language? How did it feel? Journal 6 (Expository describe) Sports /hobbies a. We all have hobbies and interests. b. Tell your journal about your favorite activity to do for fun?
217 i. When did you learn to do this activity? ii. How/why did you become interested in it? iii. When, where, and with whom do you practice this activity? c. Would you reco mmend this activity to someone else? Why/why not? Journal 7 (Expository cause/effect) Goals a. Everyone has goals and plans for the future. b. Explain your goals and why you have these goals. i. How do you envision yourself when you are an adult? ii. How will eac h of your goals help you achieve these plans for the future? c. What steps must you take to achieve your goals? Journal 8 (Expository explain) Tradition/family/culture a. We all practice family/cultural traditions. Think of a tradition in your family/cultur e. It can be something complex like celebrating a holiday or something simple like eating pizza every Friday night. b. Explain this tradition using lots of details. i. Why do you practice it? ii. What do you do (dress, eat, music, place)? iii. When? c. When you grow up will you continue this tradition with your own family?
218 J ournal 9 (N arrative ) Dream vacation a. Imagine you are able to go on a trip anywhere in the world, for free! Where would you go? b. Tell the story of your dream vacation, as if it really happened. i. Where did you go? ii. Who was with you? iii. What kinds of things did you do? c. How did you feel while you were on your dream vacation? Journal 10 (Expository explain) Three wishes a. We often wish to change our lives or the world. b. Imagine you have a magic wand and can have 3 wishes, whatever you want! Explain each wish and why you wish it. i. Will the wish help you, your family, or other people? ii. Will the wish improve the environment or the planet? iii. Why is this wish important to you? c. Now imagine you have one extra wish to give away. Who will you give it to and why?
219 Appendix D Topics that were addressed in writing prompts are underlined. Grade 6 School Culture Family Sports Country Hobbies Food Religion Fashion Music Movies Language Fun Holidays Grades 7 8 Shopping Friends Family
220 Shoes School Hobbies Sports Culture Goals Sad moments Language Where we are from Country Foods Future We are Latinos Tradition How we dress out Problems with family/friends Love life Happy moments Chores First bad hair day First award
221 Appendix E Participant Questionnaire Name: _______________________ Grade:_________________ Date of birth (month, day, year): ______________________ Place of birth (c ity/state, country): _______________________ Do you have sisters and/or brothers? List each sibling, their age, and country of birth below (for example: Francisco, 15, Mexico): ________________________________________________________________________ ____________________________________________________________ ____________ Circle the grades when you were in school in the United States : Kindergarten 1 st 2 nd 3 rd 4 th 5 th 6 th 7 th 8 th Circle the grades when you were in school in a different country : Kindergarten 1 st 2 nd 3 rd 4 th 5 th 6 th 7 th 8 th Where did you go to school outside the US? (country) _____________________ When and where did you start speaking Spanish? ________________________________________________________________ When and where did you start speaking English? ________________________________ ________________________________
222 Cuestionario La fecha de hoy _____________________ Nombre ___________________ ________ Grado ___________ Fecha de nacimiento (mes, da, ao) _________________ Lugar de nacimiento (ciudad, estado, pas) __________________________________ Lugar de nacimiento de tus paps (pas): Mam _____________ Pap ____________ Tienes hermanos? N ombra cada hermano/a, su edad, y su pas de nacimiento (por ________________________________________________________________________ ________________________________________________________________________ Circula los grado s cuando asistas a la escuela en los Estados Unidos : Kinder 1 2 3 4 5 6 7 8 Circula los grados cuando asistas a la escuela en otro pas : Kinder 1 2 3 4 5 6 7 8 Dnde asististe a la escuela fuera de los EEUU (pas)? ___________________________ Cundo y dnde empezaste a hablar el espaol? _________________________________ Cundo y dnde empezaste a hablar el ingls? __________________________________
223 Appendix F Interview Guide for Focal Participants I. Background information: 1. Participant name: 2. Age: 3. Grade: 4. Place of birth: II. Language history/use: 1. At what age and where did you begin to study/learn Spanish, English? 2. Do you or your family speak any other language(s) besides Spanish and English? 3. What age/grade were you in when you came to the US? 4. What language(s) do you speak to parents? Siblings? Grandparents, extended family? 5. What language(s) do you speak with friends in/outside of school? Phone? Email? Chat? 6. What language do you prefer for TV/radio/movies at home/with friends? (examples) 7. What lang uage do you prefer for reading for fun? (examples) III. Attitudes, beliefs, and feelings about language and literacy: 1. What do you most remember about school in your home country? Tell me about it. What did you like/not like? 2. Tell me about your experiences learning to read and write in your home country. What did you like/not like about it? What did you find difficult/easy?
224 3. Did you study English in your home country? Tell me about that experience. What did you like/not like about it? How was it different/si milar to learning English now? 4. How did you feel when you came to the US? How is it different from your home country? What was most difficult/easy to get used to? 5. Tell me about your experience learning English here. How does it feel to speak another langu age? 6. How did you learn to read and write in English? What was most difficult/easy? 7. bilingual? Why or why not? How does this make you feel? How does it make other people feel?
225 Ap pendix G Data Displays for Qualitative Analysis Figure 8 Data display for Diego.
226 Figure 9 Data display for Carolina.
227 Figure 10 Data display for Edgar.
228 Figure 11. Data display for Sara
229 Figure 12 Data display for Manuel
230 Figure 13. Data display for Juan.
231 Figure 14 Data display for cross case analysis.
About the Author Robin. L. Danzak received a Bachelor of Arts degree in Art History/Hispanic Language and Culture from New College of Florida in 1997. She completed her Master of Arts in Linguistics at the University of Concepcin, Chile, in 2001, where her research focused on literacy of deaf adol escents and adults. In Chile, Ms Danzak earned a national scho larship for graduate study and participated in a national grant for the arts to research life histories of rural craftswomen, resulting in a book publication. Upon her return to the U.S., Ms Danzak worked as a public school educator and as a doctoral s tudent at the University of South Florida, shifted her research focus to language and literacy o f English language learners. While at USF, Ms Danzak co authored an article and a book chapter, and participated in several national and international conferen ce presentations. She is currently an instructor at USF Sarasota Manatee, where she teaches online post baccalaureate courses in Language, Speech, and Hearing Sciences.
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 2200397Ka 4500
controlfield tag 001 002063949
007 cr mnu|||uuuuu
008 100318s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003032
Danzak, Robin L.
Exploring writing of English language learners in middle school :
b a mixed methods study
h [electronic resource] /
by Robin L. Danzak.
[Tampa, Fla] :
University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 231 pages.
Dissertation (Ph.D.)--University of South Florida, 2009.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
ABSTRACT: The study's purpose was to assess, through mixed methods, written linguistic features of 20 Spanish-speaking English language learners (ELLs) in middle school. Students came from Mexico, Puerto Rico, and the Dominican Republic. Participants wrote two expository and two narrative formal texts, each in Spanish and English, for a total of eight writing samples each. Additionally, students developed 10 journal entries in their language of choice, and 6 randomly selected, focal participants were interviewed for the qualitative analysis. The quantitative analysis involved scoring formal texts at the lexical, syntactic, and discourse levels. Scores were analyzed using Friedman's 2-way ANOVA by ranks, and resulting ranks were compared across genre-topic and language. A key outcome was that the text topic, rather than genre or language, impacted on rank differences at all levels, possibly due to student engagement or influence of the prompt structure.Performance at the three levels was essentially similar across both languages, revealing that participants were emerging writers in Spanish and English. Similar outcomes in Spanish and English also implied potential cross-language transfer of academic language proficiency. Results further highlighted the interaction of multiple linguistic levels in text composition. Finally, students appeared to apply a knowledge telling strategy to writing, resulting in unsophisticated vocabulary and structures. For the qualitative analysis, focal participants' journals and interview transcripts were analyzed with domain and taxonomic analyses to discern how their language learning experiences shaped their identities as bilinguals.Results showed that 1) Spanish was preferred for all focal participants; 2) students shared the experience of language discrimination; 3) bilingual and monolingual identities resulted in different attitudes toward language learning and varied writing performance; and 4) Mexican and Puerto Rican students had diverse language learning experiences, leading to differences in identities and writing outcomes. Overall, the quantitative and qualitative findings raise two questions: 1) which aspects of academic language proficiency are shared across both languages, and how might these be assessed with bilingual, integrated language measures? 2) How might integrated assessment in L1 and L2 aid in identifying adolescent ELLs with language impairment?
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Elaine R. Silliman, Ph.D.
Academic language proficiency
x Communication Sciences and Disorders
t USF Electronic Theses and Dissertations.