English language proficiency assessments are used to report English Language Learner (ELL) proficiency and progress, and classifications from these assessments help determine what educational services students receive. While classification validity is routinely reported by test vendors, empirical evidence is rarely used to justify the use of a certain classification model or scheme. In this study, fifth grade ELL and native English speaker (non-ELL) performance data on a State's Standards-Based Achievement Assessment (SBAA) and the State's English Language Proficiency Assessment (ELPA) are used to evaluate "proficient" and "non-proficient" classifications on the ELPA. Findings indicate that the currently-used conjunctive model, compared to compensatory and mixed models, creates classifications that are the least congruent with other sources of proficiency evidence for ELL and non-ELL students. Findings from this study can directly inform how policy, practice and research communities define and verify ELPA classifications for use in high-stakes decision making.

Since working memory (WM) has important implications in academics, this study investigates the construct of Baddeley’s model of WM in English language learners (ELLs). This secondary analysis of data was taken from two Institute of Educational Science Grants (#R324A090092 & #R324A090002), which included samples of cognitive and academic performance of ELL and non-ELL students in central and southern California schools. Specifically, this study asked: (1) Does the theoretical framework of Baddeleys’ multicomponent model of working memory (WM) in English and Spanish fit the data for children who are English language learners (ELLs)?, (2) Are there differences in English working memory structures between ELL and non-ELL groups?, and (3) Does the structure of WM for ELL children remain stable over time in English and Spanish measures? To examine whether the construct of WM memory varies between ELL and non-ELL children and between English and Spanish measure of WM over time, various structural equation models were conducted. This study yielded three important findings: (1) the theoretical framework for WM in English and Spanish fit the data for children who are ELLs, (2) working memory structures measured in English are comparable in ELL and non-ELL groups, and (3) the structure of WM for ELL children does not remain stable over time in English and Spanish measures. Taken together, this research finds that Baddeley’s model of WM is an adequate conceptualization for ELLs and that the construct of WM can be measured within both English and Spanish language systems. In addition, there was evidence for the measurement invariance across ELL and non-ELL groups. These finding suggest that the construct of WM is being measured consistently across the two language systems and that results from subsequent studies can be generalized to ELLs whose primary language is Spanish.

There are two camps of thought on the early education of English Language Learners (ELLs) in the United States: i) They should be taught primarily in their native language with a gradual increase in the amount of English in their curriculum over the course of Kindergarten through 5th grade, or ii) They should be fully immersed in English immediately, with only ancillary support in their native language for particularly difficult academic concepts. When California voters passed Proposition 227 in 1998 they effectively agreed with the latter theory and banned bilingual education in public elementary schools for children who are not yet proficient in English.

This thesis investigates the apparent effectiveness of Two-Way Spanish Immersion in English language acquisition by analyzing aggregate results of samples taken from two distinct populations of Public Elementary Schools in California: schools that are exclusively Two-Way Spanish Immersion (\textgreater50\% of their curriculum is taught in Spanish), and schools that are exclusively Structured English Immersion and/or English Language Mainstream. This analysis uses a sampling design that controls for confounding factors such as income and parent education level.

This study offers a mixed-methods analyses of formal and informal screening tools in L1 and L2 to identify English Language Learners in who are “at risk” for language learning disabilities. It was conducted in Bangalore, India and the sample consisted of 104 participants in Grades 2-5 from low, middle and high-income private schools. Teachers currently use school-based performance scores in English to classify students as persistent low-achievers. The purpose of this study was to provide teachers with a screening tool in both L1 Kannada and L2 English to be able classify two sub-populations of low-achieving students: students who are delayed in the second language acquisition process and students who are at risk for an underlying language learning disability.

Two formal bilingual screening tools were adapted and rendered culturally relevant in both British English and Kannada, namely the Preschool Language Scale 5 Screening Test (Zimmerman, Steiner, & Pond, 2012) and the Clinical Evaluation of Language Fundamentals 5 Screening Test (Semel, Wiig, & Secord, 2013). Both tests were efficacious in assessing general language ability, and there was a statistically significant relationship between the test scores. The PLS 5 was used to compare language competencies across age, as the same test that was developed for 7-year olds was administered to all students in the population, whose ages ranged from 7-10 years. Quantitative analysis revealed a statistically significant difference between 7-8 year olds and 9-10 year olds in their English scores but not in their Kannada scores, suggesting that L2 English was maintained as an academic language while L1 Kannada was not. The CELF 5 Test was used to classify students as “bilingual” (if they passed both tests in L1 and L2), “dominant L1”(if they only passed the Kannada test), “dominant L2”(if they only passed the English test) and “at risk for a language learning disability” (if they did not pass either the L1 or L2 tests). When CELF 5 scores were compared to school-based assessment scores, more than half of the students who were classified as being “at risk” by their teachers turned out to be dominant in their L2 according to their CELF5 classification.

Four informal screening tools were used for the study: Narrative Assessment, Parent Questionnaire, Teacher Interview and Classroom Observation. Students’ narrative skills were assessed using the Narrative Scoring Scheme (Heilmann et al, 2010). A high degree of overlap was observed between the students’ NSS scores and their CELF5 scores. Students who were identified as being “dominant L1 or L2” according to their CELF5 scores, also got an overall “proficient” classification on the NSS and students who were considered “at risk” by the CELF 5, were classified as “minimal” or “emerging” in their narrative skills. Quantitative analysis revealed that the CELF5 English and Kannada scores significantly predicted students’ NSS scores.

The other informal tools, the parent questionnaire, teacher interview and classroom observation checklist were efficacious in pinpointing external factors such as parents’ educational attainment, parents’ income levels, pedagogical practices, and special education resources, that are important when interpreting students’ performance scores across low, middle and high-income schools. Parents’ educational attainment predicted income levels in the low-income school and reading frequency in the middle-high income schools respectively.

Qualitative analyses of the teacher interviews emphasized the differences in language of instruction between low-income and middle-high income schools; whereas teachers in the former school alternated between English and Kannada, teachers in the latter schools used English only. The teacher interviews were also useful in highlighting the special education support at each school site: (a) in the low-income school, teachers treated low-achieving students as one group and they received small group instruction that targeted rote-memorization of the content related to school exams; (b) in the middle-income school, teachers viewed special education as occurring outside the purview of their classrooms, as the school had a moderate-severe special education program on the school site, but no resources for students with mild-moderate disabilities; and (c) in the high-income school, teachers followed an inclusive special education model and had access to a special education department on the school site as well as a consultancy service for assessment and intervention of students with disabilities.

Finally, qualitative analyses of the classroom observation checklist stressed the pedagogical differences across the three schools, with low and middle income schools focusing more on students’ content knowledge and rote memorization skills and high-income schools focusing more on students’ presentation skills and conceptual knowledge.

The study has implications for theoretical and applied issues concerning assessment, differentiation of language learning difference versus disability in ELLs and models and approaches for intervention.

English Language Arts has historically been tied to the civic purposes of schools, and this qualitative study of a social design-based project (Gutiérrez & Vossoughi, 2010) examines the intersection of language and literacy learning and youth civic engagement, a problem space I call “Civics English.” In this dissertation, I describe and analyze the experimentation and inquiry process of a Professional Learning Community of English teachers in a diverse middle school as they integrated civic learning and action into their English teaching practices. The dissertation examines this teacher team’s development and shifts through various tensions and challenges that arise, analyzing through the lenses of Cultural Historical Activity Theory the ways their Professional Learning Community operated as an English teaching activity system attempting to integrate the cultural activity of civic engagement, leading to the teachers’ expansive professional learning (Engeström, 2001) about possibilities and challenges of Civics English.

The English teachers implemented various civic action projects, including producing and sharing multimodal civic advocacy essays online, composing and presenting children’s storybooks about civics issues, and organizing and conducting a Town Hall with local leaders about civic dimensions of allyship and youth sports. This study looks at how, contextualized by these civics activities, they adapt and innovate customary English Language Arts practices, such as reading novels, writing in authentic genres with blended text types, and developing literacy and discourse. As the teachers encounter various tensions that arise in their attempts at Civics English, I present evidence of how these tensions emerge from the contradictions of two intersecting cultural activity systems, and what adaptations and innovations the teachers develop to overcome these tensions.

Integrating civics causes shifts in the teachers’ practices of literary study, writing, and classroom discussion, as they orient students’ learning towards public audiences, collective action, and discursive models of political and professional discourse. I identify how reading literature creates an imaginative space for civic deliberation. And I demonstrate how the Town Hall civics project shifts various dimensions of literacy and language activity by recontextualizing them. The potentials and the constraints of these shifts are examined through studying the teachers’ work, students’ language and activity, and the civic event’s efficacy as an English teaching focal point.

As an initial step toward understanding which features of academic language make science-based expository text difficult for students with different English language proficiency (ELP) designations, this study investigated fifth-grade students' thoughts on text difficulty, their knowledge of the features of academic language, and the relationship between academic language and reading comprehension. Forty-five fifth-grade students participated in the study; 18 students were classified as English language learners (ELLs) and 27 students were fluent-English speakers. Participants read two science passages, answered comprehension questions, and engaged in a retrospective interview which probed their knowledge on the academic language features of vocabulary, grammar, and discourse. Qualitative analysis was used to code students' thoughts about the challenges to reading comprehension and to identify the challenges that were related to academic language. Quantitative analyses were conducted to examine whether students' knowledge of academic language features and reading comprehension differed by students' ELP designations, as well as to investigate the relationship between students' knowledge of academic language features and reading comprehension. Results for the qualitative analysis revealed that students found difficult vocabulary, reading abilities, and prior knowledge as the greatest challenges to comprehending the science passages. Results from the quantitative analyses indicated that ELL students' knowledge of academic vocabulary, grammar, discourse knowledge, and reading comprehension (as measured by multiple-choice questions) were significantly lower than the fluent-English speaking students. The results also indicated that vocabulary, not grammar or discourse features, was significantly related to students' comprehension scores. The results have implications for understanding the features of academic language that influence students' comprehension of expository texts in science.

This case study, drawing upon the ecological perspectives (Kramsch, 2002; van Lier, 2004) as a theoretical framework, described the learning experiences of two second generation and first grade Vietnamese English Language Learners navigating between home and school to develop oral and written L1 Vietnamese and L2 English competence for one school year. In the second school year, the focal students' oral and written language samples were collected without classroom observations or interviews. Focal student one had the advantage of learning language in three settings--home, public school, and Vietnamese Sunday School--while focal student two had access to the first two only. A variety of qualitative tools were developed to capture the language affordance, interaction and emergence of ELL learners: (a) observation fieldnotes (b) parent and teacher interviews, and (c) L1 and L2 oral and written reading survey and assessment. Metalinguistic awareness, through language play, language rehearsal and repeated reading emerged as an important mediator of language competence and as an interpretative framework to drive the analytic induction process (Erickson, 1986; Bogdan & Biklen, 2003) that I employed. Data interpretation focused on the L1 Vietnamese and L2 English oral and written modes.

The study findings suggest that when teachers, parents and students collaborate to generate and activate the L1 and L2 language affordances (van Lier, 2000) of Vietnamese ELLs through reflexive form of interaction, language learning and competence will emerge. Awareness of language form and function assisted focal students in developing L1 and L2 oral and written competence. Overall, reading assessment results indicate moderate-high growth by the end of the school year for focal student one. She mastered the early reading abilities, such as letter names and sounds, high frequency word and decodable words in L1 Vietnamese and L2 English. Listening and reading comprehension improved more for L2, compared to L1. Focal student two, who had access to only 2 sites (home and public school), showed low-moderate growth in listening and reading comprehension skills in L2 English. Her L2 early reading skills were high, compared to no improvement in L1. There was no growth in L1 listening and reading comprehension through oral retell and writing tasks. Because observations began at the start of first grade, after one year of L2 English exposure in kindergarten, both focal students used more L2 oral language to communicate with family members, peers and teachers. Focal student one

increased her codeswitching as her fluency improved in both L1 Vietnamese and L2 English at age six. Additional participation in Vietnamese Sunday School was beneficial for focal student one as evidenced by her overall improvement in L1 Vietnamese and L2 English oral and written language competence.

California policy defines English Language Learners (ELLs) as students who have not reached English proficiency and who speak another language at home. The state sets minimum requirements to determine students’ English proficiency, but individual districts have the freedom to be more rigorous. Therefore, a student can be considered ELL in one district and Reclassified Fluent English Proficient (RFEP) in another. My work examines how districts establish their language classification policies and the extent to which the classification itself determines students’ opportunities to learn.

My three-study dissertation addresses the following research questions: 1) How do school districts establish and implement their language classification policies for middle school students? 2) How does language classification affect eighth-grade English and math course placement? 3) How does language classification affect high school students’ English and math achievement and behavioral outcomes? I use data from two Southern California districts, including interviews, documentation and student-level data, to answer these questions. I focus on middle school because it is less frequently studied and it determines adolescents’ high school outcomes.

In study one, MUSD and GUSD administrators developed individual language classification to equitably determine their students’ English proficiency. However, in both districts, assessments served as gatekeepers that students must pass to be reclassified. Male, Hispanic, and low-income students were less likely to be reclassified, even if we consider only those who passed the district’s or state’s requirements. In study two, MUSD administrators stated language classification itself did not determine eighth-grade English and math placement. However, data show under what circumstances language classification influences students’ course placement. The year ELL started to receive specialized English courses, they were placed in lower math courses. In study three, in most instances middle school language classification itself does not affect achievement or behavioral outcomes in high school. At MUSD, ELL may be receiving appropriate English and math; therefore, there is no effect. Alternately, reclassification may not have impact since neither ELL nor RFEP may be receiving adequately rigorous content. The few times classification does affect outcomes, RFEP perform worse than ELL students.

Qualitative research methods guide data collection and analysis (Bogdan & Biklen, 2007; Dyson & Genishi, 2005; Heath & Street, 2008; Schram 2006) of this 18 month study researching the academic lives of 21 Spanish-English language learners classified as limited English proficient (LEP) in two urban, elementary school communities. In this dissertation, I argue language learners in elementary school begin to construct identities with language through the school community resources they are able to access as they participate in school-based activities. These language identities are related to students' engagement and motivation to participate in school, and also hold a relationship with students' language classification status. By triangulating the data of 21 focal students to include: 1) student interviews, 2) observations of students in academic and non-academic school activities, and 3) students' performance on academic tests, findings reveal how all language learners construct one of three language-learning identities - dual, separation, or distant - where a distant identity is associated with students who remain classified LEP or are reclassified as a long-term English language learner upon exiting elementary school.

ABSTRACT OF THE DISSERTATION

A Study of the Effects of Bilingual and Structured

English Immersion Programs on the Oral and Literacy Development

of Students Learning English as a Second Language

by

Danny Lee Wood

Doctor of Philosophy in Education

University of California, Los Angeles, 2014

Professor Concepción M. Valadez, Chair

This study had two purposes: (1) to compare two different English language learner (ELL) programs--transitional bilingual education (TBE) programs and transitional structured English immersion (SEI) programs--to identify which, if either, was superior in facilitating English language acquisition; and (2) to identify the relationship between English oral proficiency and English literacy proficiency. To answer these questions, this study analyzed English proficiency scores from the California English Language Development Test (CELDT) from a large school district in California from 2002 - 2007. The CELDT is an instrument that was aligned with the English Language Development (ELD) standards of California and first administered in 2001 to measure English language proficiency. The CELDT provided measurements in three English skill scales: oral (listening and speaking), reading, and writing. The CELDT also produced a composite overall English proficiency score from the combined performance scale scores.

The students in the study were primarily Hispanic (98.26%) native Spanish speakers (99%) from lower socioeconomic backgrounds (91.4%). Of the 2,731 students whose scores were analyzed, 1,196 (44%) were in their respective SEI or TBE programs the entire 6-year period of the study.

For the program comparisons, there were four groups of students: (a) students who were in transitional bilingual programs for the entire length of the study, (b) students who were in transitional bilingual programs for various lengths of time before transferring to SEI programs, (c) students who were in the SEI program for the entire length of the study, and (d) students who were in the SEI program for various lengths of time before transferring to other programs or schools.

The bilingual and SEI programs in this study were analyzed in four different ways: raw CELDT score means, CELDT proficiency levels, CELDT highest scorers, and the comparison of oral to literacy skills. When CELDT means were compared, the SEI groups performed slightly better throughout the study period over all. However, the bilingual students narrowed the scoring gap each year across all skill areas, and in the case of oral skills, one of the bilingual groups scored higher than both SEI groups in 2007.

When CELDT overall and oral proficiency levels were compared, all groups finished in the same level by grade four (2006) for overall proficiency, and were the same for oral proficiency by grade five (except for the higher bilingual score just mentioned). By 2007, while SEI students slightly outperformed bilingual students each year, the proficiency scores for reading and writing were close between both programs. When scores were calculated for the percentage of students that scored at the highest levels between the groups, the SEI scores were almost always the highest. However, bilingual students out performed SEI students in 2007 on CELDT and academic measures.