The Effects of a Comprehensive Language Arts Transition Program on the Literacy Development of English Learners

William M. Saunders & Claude Goldenberg
California State University, Long Beach

1999

ABSTRACT
This study examines the effects over the course of one school year of a language arts transition program on second and fifth grade English language learners. The program was designed and successfully evaluated at a previous site with Spanish-speaking transitional bilingual students and is currently being implemented and tested at a new site as part of both transitional bilingual and all-English programming. Teachers had been implementing the program for two years and were successfully using most, although not all, program components. The study includes 70 matched pairs of "program" and "nonprogam" English learners (matched on prior achievement), including 49 from transitional bilingual programming (16 in 2nd and 33 in 5th grade) and 21 from all-English programming (7 in 2nd, 14 in 5th). Literacy outcomes include standardized tests and a performance assessment conducted at the end of the school year. Transitional bilingual 2nd graders were tested in Spanish; all other students were tested in English. Results indicate that (1) overall, the program produces significantly higher levels of literacy attainment for English learners; (2) program effects are noticeably stronger on students in transitional bilingual programming in contrast to students in all-English programming; and (3) program effects are stronger on language measures than they are on reading. Results indicate program effects are being successfully replicated for transitional bilingual students. However, continued work is needed to improve reading effects and to effectively adapt the program for students in all-English programming.

The Effects of a Comprehensive Language Arts Transition Program on the Literacy Development of English Learners
The number of limited-English speakers in U.S. schools has risen over the past two decades and continues to grow. At a time when the size of the general school population has remained essentially stable, the number of limited English proficient students (three-fourths of whom are Spanish speakers) grew by 85% nationwide between 1985 and 1992 (National Clearinghouse for Bilingual Education, 1995). More than 2 million students--4% of the U.S. student population--speak a language other than English in their homes and are not fluent in English (Fleischman & Hopstock, 1993; National Clearinghouse for Bilingual Education, 1995).
Perhaps 50% of these students are in some form of "transitional bilingual education," or TBE (August & Hakuta, 1997). These students receive academic instruction in their native language during the early years of schooling then "transition" into mainstream English once they are proficient enough to participate fully in all-English instruction and activities. The rest of these students are in all-English education. They are, essentially, expected to learn English while acquiring academic competencies in language arts, mathematics, and other curriculum areas.
Recent evidence suggests programs that maintain and actively promote continued use of the primary language, rather than having students transition to all-English instruction in elementary school, produce superior academic outcomes in English (Thomas & Collier, 1997). Yet these programs are exceedingly rare (August & Hakuta, 1997). The fact remains that for the large majority of English learners, even if their home language is used academically at all, it is for a relatively brief amount of time and generally for no more than a few years during elementary school. Once children have reached a certain level of literacy in their home language and oral fluency and comprehension in English, they are transitioned into all-English instruction. For those students in all-English programs from the beginning of their schooling careers, they are immediately thrust into a situation where they must learn alongside or be compared to students who have heard and spoken English since infancy. In either case, English learners—and their teachers—-face a daunting challenge.

Latino students, who comprise by far the largest group of English learners in U.S. schools, tend to do very poorly academically. They have lower levels of achievement and higher dropout rates than their white non-Latino counterparts (Valencia, 1991). Despite some progress over the past 15 to 20 years, 85% of Hispanic fourth and eighth graders still read at a "basic" level or below. Over half score even below "basic," meaning they cannot demonstrate understanding of a text written at their grade level (Mullis, Campbell, & Farstrup, 1993).

Achievement gaps between Latinos—whether they are U.S.- or foreign-born (Kao & Tienda, 1995)—and whites in all academic areas appear early and persist throughout their schooling careers. For example, 17-year-old Latinos read only slightly better than 13-year-old white students (U.S. Department of Education, 1995). The overall picture for students from Spanish-speaking backgrounds is troubling indeed. It is critical, therefore, that we develop and identify effective programs to improve academic achievement for this population of students.
Unfortunately, educators have little research upon which to base policy and practice when instructing these children in English (Goldenberg, 1996). Teachers tend to describe themselves as overwhelmingly uncertain about the appropriate methods to use to instruct these students in English (Gersten & Woodward, 1994). Even in schools and districts recognized for exemplary bilingual programs, transition—that is, when students move into all-English instruction—is often a conundrum (Berman, et al., 1992; Gersten, 1996).

As part of a previous project (Saunders, 1998; Saunders & Lennon, 1996; Saunders, O'Brien, Lennon, & McLean, 1998), members of our research team collaborated with educators from a school district in Southern California to develop and evaluate an effective language arts program for Spanish-speaking children. In general, our efforts proved successful: The program produced significantly higher levels of Spanish literacy achievement at grades 3 and 4 and English literacy achievement at grade 5 (as gauged by both standardized and performance-based assessments), significantly higher numbers of students who formally demonstrated fluent English proficiency by grade 5 (and were reclassified from "limited" to "fluent" English proficient), and more positive attitudes towards bilingualism (specifically students' attitudes towards Spanish literacy) for significantly larger numbers of students (Saunders, 1998).

In our current project, we are studying the implementation and effects of the program at a new school site. Using a strategy Slavin (1984) has called "component building," we are systematically testing through controlled studies the individual and combined effects of several of the 12 program components. In a previous study (Saunders & Goldenberg, 1997) we found that teachers consider all the program components important, although some more important than others. Our assumption is that teachers of transition students need precise and systematic information about the relative contribution to children's achievement of specific program components in order to make informed choices when faced with the inevitable constraints of time and resources. (See Saunders & Goldenberg, 1999, for the first in our series of component building studies.)

This language arts transition program was originally conceived to facilitate students' transition from Spanish to English instruction. However, the 1998 passage of Proposition 227 in California has effectively ended Spanish-language instruction in many schools throughout the state. We strongly object to attempts to end bilingual education. Nonetheless, we are now faced with a situation where large numbers of English learners, formerly in primary language programs until they transitioned into English instruction, are now learning academic skills in English. This suggests the question of whether the language arts transition program might also be helpful to English learners who have been learning academic skills in English. Although conceived in a context where students had received academic literacy instruction in their first language and were transitioning to academic instruction in English, we expected that the program could function as an effective "language arts" program, regardless of whether English learners were transitioning from Spanish into English instruction or if they were acquiring English academic skills in an all-English context. Some of the students with whom we are working have been in a transitional bilingual program where they learned initial literacy skills in Spanish; others have been in all-English programs. We are thus investigating the utility and effectiveness of this program as a means of helping English learners develop English language-arts knowledge and skills, regardless of their initial language of instruction.
Focus of This Study
This study examines the effects of the language arts transition program on English language learners participating in transitional bilingual and all-English programming. Teachers represented in this study had been implementing the program for two years and were successfully using most, though not all, program components. The focus of the study is threefold. We are (1) testing the combined effects of the cluster of components teachers successfully implemented, (2) evaluating the extent to which overall program effects are being replicated among TBE students, and (3) assessing, for the first time, program effects on students in all-English programming.
THE LANGUAGE ARTS/TRANSITION PROGRAM
Three-Year Design for Transition
Within the context of transitional bilingual programming, our design for transition optimally spans grades 3-5. Grade 3 is explicitly considered a Pre-Transition year, grade 4 is Transition I, grade 5 is Transition II (see Table 1). The concept of a Pre-Transition component is designed to emphasize the fundamental role of Spanish reading and writing and oral English development that precedes transition. The thrust of this phase is intensive Spanish reading and writing instruction and extensive oral English development. The goal is to have all students performing at grade level in Spanish reading and writing, and at the speech emergence level in oral English development by the end of 3rd grade, which in this particular district would qualify students to begin transitional language arts.

Table 1: The Design for Transition within Transitional Bilingual Programming.
Phases Optimal
Grade Goal Measurable
Outcome
K-2nd o Initial Reading & Writing Proficiency
(Spanish)
o Early Production II (oral English) Existing norm-or
criterion-referenced measures
Pre-Transition 3rd o Grade appropriate Reading & Writing
Achievement (Spanish)
o Speech Emergence (oral English) Pass CARE
(district transition instrument)
Transition
I 4th o Initial Reading and Writing Proficiency
(English)
o Academic Oral Language Proficiency (English)
o Grade appropriate Reading & Writing Achievement (Spanish) Existing norm-or
criterion-referenced measures
Transition
II 5th o Grade appropriate Reading & Writing
Achievement (English)
Redesignation:
LEP to FEP The concept of Transition I and II--grades 4 and 5--is designed to make explicit the need for a concrete transition program of serious substance and duration. By the end of Transition I students should be within a year of grade level in English reading and writing and have sufficient oral English language proficiency to participate actively in academically-oriented discussions. Finally, students should continue to demonstrate grade level Spanish reading and writing proficiency.
By the end of Transition II, students should be decoding and comprehending grade level material in English, both in terms of literature and in the content areas. The goal is reclassification: Grade level or close to grade level English skills that allow student to be reclassified as fluent rather than limited English proficient.
The Language Arts Program
The language arts progam used in conjunction with the 3-year design for transitional bilingual programming and, in our more recent work, within all-English programming includes 12 different components organized into three different strands: Studying Literature, Skill Building, and Other Supporting Components (see Table 2; see also Appendix for short descriptions of each component). Table 2: Language Arts Program Components.
Studying Literature Skill Building Other Supporting Components
Literature Units
Comprehension Strategies
Pleasure Reading
(experience-text-relationship)
Literature Logs
Assigned Independent Reading
Teacher Read-Alouds
Instructional Conversations
Dictation
Interactive Journals
Culminating Writing Projects
(writing-as-a-process)
Written Conventions Lessons

ELD through Literature
Studying literature. Across all phases of the program, from Pre-Transition to Transition II--from Spanish to English language arts and from grades 2 or 3 through 5--students study literature. We assumed that students would benefit from more extensive and intensive opportunities to work with text, to study interesting stories under the tutelage of a teacher. Based upon research conducted as part of the Kamehameha Elementary Education Program in Hawaii (Au, 1979, 1992; Tharp & Gallimore, 1988) and Spanish-speaking Latino communities in southern California (Saunders & Goldenberg, in press; Saunders, Patthey-Chavez, & Goldenberg, 1997; Goldenberg, 1992/93), we adapted the Experience-Text-Relationship (ETR) approach as our framework for studying literature. Through ongoing discussions (instructional conversations), writing activities (literature logs and culminating writing projects), and reading, the teacher helps students study the story in relationship to their own experiences and a central theme.
With respect to literacy development, we assume that through this recurrent process of individual and social discourse--of reading, writing, and discussing-- studying literature helps students learn to comprehend text, make connections between the text and their own lives, and develop more fully formed concepts about the themes addressed in the units (Tharp & Gallimore, 1988). In terms of second language acquisition (Cummins, 1989; Krashen, 1987), we assume literature units help provide substantial comprehensible input--language that includes slightly more sophisticated structures or vocabulary than the learner can produce on his/her own, but is understandable within the total context in which it is used. The literature unit becomes a meaningful social context in which words, phrases, language structures, and concepts are used, acquired, and learned (see Saunders et al., 1998, for a more detailed explanation of the ETR approach).
Skill building components. As we found throughout our research and development work, studying literature needs to be complemented by additional skill-building components. Students need direct instruction in specific reading comprehension strategies (predicting, summarizing, questioning), and they need daily opportunities to read texts geared to their reading level--assigned independent reading. Students need similar study and practice experiences for written language. As part of the weekly dictation program, students study a short but carefully targeted passage from the literature selection, and also receive regular lessons on written language conventions.
English Language Development through Literature (developed by project consultant Dolores Beltrán and project advisor, Gisela O'Brien) is a daily, 40 minute oral English program used in the Pre-Transition phases of the program. Instruction is delivered to students in small, homogeneous groups based on students' proficiency level. Lessons and independent activities are all drawn from a particular literature selection (typically one with predictable patterns, language structures, and target vocabulary for various domains). The focus of lessons and the teacher's talk are geared specifically to students' production level.
Other supporting components. Teacher read-alouds and pleasure reading are both designed to expose students to good literature and support their independent reading behaviors. At all grades, teachers read to students for approximately 20 minutes at least 3 times per week. In addition, a portion of time each day is devoted to pleasure reading. Students choose their own books and stories, keep records of their reading, and for those books they find most interesting, they complete short assignments (summaries, synopsis, oral presentations, drawings, etc.). In addition, many Transition I teachers use interactive journals during the first half of the year when students are making their first attempts at English writing. The immediate written response from the teacher provides both emotional support for the students and a highly contextualized and therefore comprehensible English text for students to read.
Theoretical Premises
Four theoretical premises under gird the program, all of which are assumed to promote 1st and 2nd language acquisition and achievement:
o Challenge: consistently challenge students academically--challenge them to think, learn, and engage intellectually;
o Continuity: achieve continuity in curriculum and instruction as students move from primary to middle to upper grades, and from Spanish to English language arts;
o Connections: build upon and make explicit connections between students' existing knowledge, skills, and experiences and the academic curriculum to be learned (including language, literacy, and content);
o Comprehensiveness: address both meaning and skills, both higher level thinking and appropriate drill and practice, and provide complementary portions of student- and teacher-centeredness.
These premises are grounded in the research literature, specifically studies that have tried to identified the characteristics of more and less successful programs for English learners (Berman et al., 1992; García, 1992; Gersten & Jiménez, 1993; Ramírez, 1992).
METHODS
Design & Procedures
Study design. We used matched samples of students to evaluate the effects of the language arts/transition program on English learners' literacy development over the course of one school year. Students who participated in the language arts/transition program ("program" students) were enrolled during the 1997-98 school year in the classrooms of teachers implementing the language arts/transition program at the project school. Matched comparison students ("nonprogram" students) were randomly selected from schools in the same community and district and and matched with program students based on standardized test scores (see below for matching procedures) . All students were English learners, who were enrolled in one of two district language-programming options: Transitional Bilingual Education (TBE) students receive academic instruction in their native language during the early years of school and then transition into an all-English program; English Language Development (ELDP) students are in all-English education.
Procedures. The study was conducted retrospectively. All student data used for this study were drawn from a larger database compiled under the direction of our research team at the end of the 1997-98 school year. The larger database (1311 students) comprises performance assessment measures, standardized test scores, demographic data and programming information on randomly selected 2nd and 5th graders from 22 district elementary schools, as well as 2nd, 3rd, 4th and 5th graders enrolled in the classrooms of eight program teachers at the project school. Students participated in the performance assessment during May or June, 1998. All test, demographic and programming data were compiled from computer print-outs generated through the district's official student information system. Samples for this study were drawn by matching students from other schools to the students of program teachers.
Establishing the composition of the "program" sample. In drawing matched samples, we first established the composition of our program sample. We included only 2nd and 5th program students because those were the two grades randomly sampled from all other schools (i.e., available in the larger database). This provided a pool of 5 program classrooms: three 2nd grade and two 5th grade. One 2nd grade classroom was excluded because the teacher was completing her first year of program implementation. The other four teachers were completing their second year of implementation and thus provided a better sample for the purposes of the study. One of the 2nd grade classrooms, a Spanish-bilingual class, included nineteen TBE students. The other 2nd grade classroom, an all-English class, included 11 English language learners (ELLs) in the all-English program and 7 native-English speakers (not included in this report). Both 5th grade classrooms (n = 29 and 26) were all-English classes with a mixture of ELLs: most had transitioned or were in the process of transitioning from the bilingual program (TBE; n = 40), some had received ELDP (all-English) programming (n = 15). Seven 2nd grade and eight 5th grade students were dropped from the study because of missing test scores. These four classrooms provided a total sample of 70 ELLs from "program" classrooms: 23 second graders (7 ELDPs and 16 TBEs); and 47 fifth graders (14 ELDPs and 33 TBEs).
Matching "program" and "nonprogram" samples. Program and nonprogram students were matched on 3 variables: programming option (TBE or ELDP), home language, and standardized reading and language test scores from the previous year, spring 1997 (end of 1st grade for 2nd graders and end of 4th for 5th graders). Only students with these data and with spring 1998 test scores available were included in the subject pool. All program and nonprogram TBE students are native Spanish-speakers. Two 2nd grade and one 5th grade ELDP (all-English) matched pairs are native Armenian-speakers; all other 2nd grade pairs are native Spanish-speakers.
Test language: Pre (for matching) and Post (for analyzing outcomes). All ELDP students were tested the previous year in English and matched based on measures of English reading and language. All TBE students were tested in Spanish the previous year and matched on the basis of Spanish reading and language measures. Second grade and fifth grade ELDP students were assessed at the end of the year in English, and second grade TBE students were assessed at the end of the year in Spanish (i.e., for these students Pre and Post measures are the same language). Fifth grade TBE students were assessed the previous year in Spanish but assessed in English at the end of the year (i.e., for these students the Pre-measures are in Spanish and the Post measures are in English). Table 3 shows the composition of the matched samples and the language of Pre and Post measures.

Table 3: Composition of the Matched Samples and Test Language for Pre and Post Measures.
Grade Treatments ELL Programming Options ALL
ELDP TBE
n Test Language n Test Language n Test Language
Pre Post Pre Post Pre Post
2nd Program
Nonprogram 7
7 Eng Eng 16
16 Span Span 23
23 E & S E & S
5th Program
Nonprogram 14
14 Eng Eng 33
33 Span Eng 47
47 E & S EngClarifying Treatment Conditions: "Program" vs. "Nonprogram"
Program components: Implementation status. Teachers were in their second year of program implementation during the year data were collected for this study. At the end of the year, teachers rated the consistency and effectiveness with which they thought they were implementing all program components. Project advisors also independently rated each teacher, component by component, for implementation consistency and effectiveness. These ratings provide an indication of the more and less prominent components comprising the "program" treatment. With respect to consistency, teachers rated on a 3-point scale the extent to which they were implementing or using each component an ongoing basis according to the designs of the program, from not using the component to using it consistently ("built-in" to the teacher's langauge arts program). For the 4 teachers represented in this study, teacher and advisor ratings are highly consistent (87% exact agreement; 13% within 1 point). Components are grouped in Table 4 according to teachers' consistency ratings, averaged across the four teachers.
With respect to effectiveness, teachers rated on a 4-point scale the level of effectiveness they achieve when they use the component: Not effective; Somewhat effective; Effective; Very Effective, with "effectiveness" operationalized as the degree to which the teacher feels the component is benefitting her students consistently and substantially. The 4 teachers in this study tended to rate their effectiveness one point lower than the advisors on about half the components (48% exact agreement; 45% within 1 point, 7% with a 2 point difference). In Table 4, components are rank ordered within each consistency level according to teachers' averaged effectiveness rating; advisors averaged ratings are shown, as well.

Table 4: Teacher and Advisor Ratings of Consistency and Effectiveness of Program Implementation.
Implementation Consistency
Scale = 1-3 Language Arts Program Components Implementation Effectiveness
Scale = 1-4
Teachers Ratings
of Themselves Teachers' Ratings of Themselves Advisors'
Ratings of Teachers
Using Consistently Dictation 3.25 3.50
(mean = 3.00) Literature Units 3.00 3.50
Literature Logs 2.75 3.50
Assigned Independent Reading 2.50 3.25
Instructional Conversation 2.00 3.00
Consistently/Intermittently Teacher Read-Alouds 3.75 4.00
(mean = 2.50) Conventions Lessons 3.25 3.50
Pleasure Reading 3.25 3.25
Using Intermittently Interactive Journals 2.50 2.75
(mean = 2.00) Comprehension Strategies 2.00 2.25
Culminating Writing Projects 1.75 3.25
Mean of means 2.72 3.25
Note: The Oral English Language Development component is not shown here as it applies to only the 2nd grade teachers (averaged consistency rating was 3; averaged teacher and advisor effectiveness ratings were both 3).
Study results should be interpreted with these implementation data in mind. First, according to the advisors' ratings, teachers are proficient enough to implement effectively or better (i.e., 3.00 or higher) all but 2 components (comprehension strategies and interactive journals). That teachers tend to rate their effectiveness lower than advisors likely reflects teachers' conscientiousness and the challenging nature of some of the components (e.g., Instructional Conversation, Culminating Writing Projects). Second, despite the generally strong effectiveness ratings, there is noteworthy variation among the components in terms of implementation consistency, corroborated by uniform teacher and advisor ratings. Dictation, Literature Units, Literature Logs, Assigned Independent Reading, and Instructional Conversation are used consistently across the four classrooms. Teacher Read-Alouds, Conventions Lessons, and Pleasure Reading are used consistently in some classrooms and intermittently in others. And finally, Interactive Journals, Comprehension Strategies and Culminating Writing Projects are not implemented consistently in any of the classrooms, only intermittently.
Components with lower consistency ratings likely reflect (1) the long-term nature of total program implementation (i.e., it takes time to learn to make all the components fit within a finite amount of instructional time), and (2) the continuing need for development and refinement of some components (e.g., Comprehension Strategies and Culminating Writing Projects).
Program teachers and their professional development opportunities. Study results also should be interpreted based on an accurate description of the program teachers and their unique professional development opportunities. The 4 teachers represented in this study are all veterans with from 8 to 12 years experience. Three of the four have either a bilingual or language specialist credential, and all four have no less than 6 years experience working primarily with ELLs. Like all other program teachers, these four volunteered and made a long-term commitment to work with our research team and the project advisors on implementing and refining the program. As part of that commitment they participate in a "teacher workgroup" lead by the two project advisors (Dr. Kathy Hasenstab and Gisela O'Brien). Across the school year, the group meets twice a month for two-hours afterschool and 3-4 times a year for a full day. In addition, project advisors regularly spend time in the classroom demonstrating for and collaborating with program teachers. Teachers are modestly compensated for their afterschool hours and provided with a substitute for the all-day meetings. Without question, these four individuals are among the strongest teachers implementing the program, and they are generally recognized among their colleagues and by project staff as highly conscientious, hardworking, effective and caring teachers.
The "nonprogram" program. The "nonprogram" condition in this study does not represent a known instructional entity. Based on best-possible matches, the nonprogram sample includes students from 18 different schools located in the same region of the same district. Students were selected randomly to be part of the larger database and entered this study randomly, as well, based solely on matching criteria--programming option, home language and prior achievement. The majority of nonprogram students (66%, 46 of 70) come from 10 schools with similar demographics and achievement levels as the project anchor school; the rest come from 8 schools that generally tend to have slightly higher socioeconomic and achievement levels and lower numbers of ELLs. As a consequence, the nonprogram sample provides a very closely matched norm group, of sorts, drawn randomly and likely representative of achievement levels for this specific ELL population from around the district. All district schools are mandated to follow the same guidelines for TBE and ELDP programming. Any description of the language arts program students received, however, would be pure speculation on our part. It is possible that some of the components that comprise our language arts program are being used elsewhere, although our team has not been involved in any effort to implement the program at any of the schools represented in this study, other than the project school.
Measures
Standardized tests. Norm-referenced, standardized tests are administered at all schools in accord with mandated district procedures. The English language instrument is Stanford 9 and the Spanish language instrument is APRENDA, both of which include subtests of reading and language (and both are published by Harcourt Brace Jovanovich, Inc.). The district maintains strict guidelines regarding the testing of ELLs: ELDP students are administered standardized tests in English beginning after 12-18 months of English literacy instruction. The same criterion holds for Spanish-speaking TBE students; however, prior to the 12-18 months of English instruction, students are tested in Spanish. These guidelines governed testing policies through the spring of 1997, when students in this study were "pre" tested. In the spring of 1998, when students were "post" tested, the State mandated English testing for all ELLs. In accord with district policies, ELLs normally tested in Spanish also took Spanish tests. For all standardized measures, statistical analyses were conducted on and results are reported in Normal Curve Equivalents.
Performance assessment content and administration. Project-developed Spanish and English performance assessments (Goldenberg & Saunders, 1996) measure: 1) Narrative comprehension--students read a portion of a grade level appropriate story, summarize what they read, and then write an ending for the story (scored for comprehension of story); 2) Informational comprehension--students read a grade level appropriate informational text, describe what they learned, and then respond to an inference question (scored for comprehension of text information); 3) Writing communication--students choose a favorite story they have read and summarize it (scored for the clarity with which students convey the content of the story); 4) Writing conventions--students' summaries of their favorite stories are also scored for conventions (spelling, punctuation, capitalization, usage); 5) Dictation--students take dictation for a grade level appropriate text (scored for spelling, punctuation, and capitalization).
The assessment is typically administered in a small group setting (n = 6-10) on three consecutive days (60-90 minutes each day). Students are given as long as they need to complete each task. The small group setting allows for close proctoring to insure that students understand all tasks Program students were administered the assessment in their own classrooms by their teacher. Project teaching assistants administered the assessments to all nonprogram students on a pull-out basis in small groups. All administrations followed common guidelines described in a proctor's manual.
Performance assessment scoring and reliablity. The work of students in this study was scored as part of a large scale performance assessment (including assessments for all 1311 students in the larger database) during a formal, 2-day scoring session involving 55 teachers from local schools and directed by the first author. At each scoring session, "readers" were trained on each task to be scored. Scoring leaders (researchers, project advisors) reviewed the scoring rubric and anchor papers in order to prepare readers to make consistent and accurate judgments. Each paper was read and scored blind (papers were coded with identification numbers). Papers from all 22 schools were randomly assembled into stacks so that each reader read and scored papers from a range of schools.
At least 20% of the papers for each task at each grade level were scored by two independent readers to gauge interreader reliability: the first and last 10% of papers scored for the task. Exact agreement is when two independent readers give a paper the same score. Adjacent agreement is when two readers "agree" within one score point (i.e., same score or scores that differ by one point). With a five-point scale, exact agreement levels above 70% and adjacent agreement levels above 90% are considered strong levels of reliability. Agreement levels averaged across all tasks and grade levels are: 81% exact and 99% adjacent.
Performance assessment data. Scoring rubrics were developed based on national and state models (National Assessment of Educational Progress, California Learning Assessment System). Based on the 5-point scale, a score of 4 indicates students are meeting challenging grade level standards; a score of 3 means students are approaching those standards and have demonstrated at least basic competence for that grade level. Results to be reported here are based on scores combined across tasks: We averaged students scores across all five tasks to produce a single composite score. The performance assessment was conducted only at the end of the year, that is, at Post (end of 2nd and end of 5th). Second grade TBE students took the assessment in Spanish. All other students took the assessment in English.
Data Analysis
All measures were analyzed using 2-way analysis of variance (treatment x programming option; i.e., "program" and "nonprogram" by ELDP and TBE). Standardized reading and language scores (NCEs) from the previous year ("pre") were analyzed to check the comparability of the matched samples. Standardized reading and language scores and performance assessment composite scores from the end of the year ("post") were analyzed to test for treatment main effects and also potential interactions (differential treatment effects, "program" or "nonprogram" on ELDP and TBE students). Data for 2nd and 5th grade samples were analyzed separately. Significance level is .05.
RESULTS
2nd Grade Pre-Measures: Comparability Check
There were no significant differences between treatment groups on the pre-measures. Results of two-way ANOVAs (treatment group x ELL programming option) on standardized reading and language scores were the same: no significant main effect for treatment group and no interaction (p-values > .72). As shown in Table 5 means are virtually identical for program and nonprogram TBE students, as well as for program and nonprogram ELDP students. However, there was a significant main effect for programming option (reading: 1,42 df, F = 18.54, p < .01; language: 1,42 df, F 4.41, p < .05). The noticeable and signficiant difference between TBE and ELDP students, regardless of treatment group, is not surprising: TBE students were tested in their home language, Spanish, while ELDP students were tested in their second language, English.

Table 5: Results on Pre-Measures for the 2nd Grade Matched Samples (End of Grade 1).
Group ELDP
n = 7 & 7
Language of Test is English TBE
n = 16 & 16
Language of Test is Spanish ALL
n = 23 & 23
mean sd mean sd mean sd
Standardized Program 37.81 12.16 66.79 23.53 57.97 24.57
Reading Non 38.29 11.52 64.00 20.83 56.17 21.87
Standardized Program 37.63 10.26 48.39 20.03 45.12 18.11
Language Non 38.59 8.36 51.46 19.79 47.54 17.96
Notes: Standardized measures are Normal Curve Equivalents. 2nd Grade Post-Measures: Testing for Treatment/Program Effects
Program students significantly outperformed nonprogram students on all post-measures. Results of the 2-way ANOVAs were the same for standardized reading and language, and also the performance assessments: significant main effects for treatment group (all Fs > 6.70, all p-values < .05) and ELL programming option (all F's > 6.10, all p-values < .05), but no significant interaction (all Fs < 2.10, all p-values > .15). Means and effect sizes are reported in Table 6. Overall, program students scored no less .90 standard deviations higher (the effect size) than nonprogram students on each measure.

Table 6: Results on Post-Measures for the 2nd Grade Matched Samples (End of Grade 2).
Group ELDP
n = 7 & 7
Language of Test is English TBE
n = 16 & 16
Language of Test is Spanish ALL
n = 23 & 23
mean sd ES mean sd ES mean sd ES
Standardized Program 42.36 15.79 +.34 75.41 17.09 +1.36 65.35 22.56 +.93
Reading Non 36.16 17.92 54.05 15.67 48.60 18.05 *
Standardized Program 52.51 16.37 +.42 72.19 18.26 +1.37 66.20 19.65 +.94
Language Non 41.66 25.82 51.21 15.34 48.30 19.04 **
Performance Program 2.51 .99 +1.13 3.62 .50 +2.30 3.29 .84 +1.63
Assessment Non 1.89 .55 2.56 .46 2.36 .57 **
Notes: Standardized measures are Normal Curve Equivalents. Performance assessments are scored on a 5-point scale, where 3 is approaching and 4 is meeting challenging grade level standards.
* p < .05; ** p < .01.
Because there were no significant interactions, we did not perform post hoc tests on program and nonprogram differences within the TBE and ELDP groups. However, we did calculate the effect sizes for those differences, which reveal a much stronger program effect for TBE students than ELDP students. Program TBE students scored at least 1.3 standard deviations higher than nonprogram TBEs on standardized measures and more than 2 standard deviations higher on the performance assessment. In contrast, program ELDP students scored .34 and .42 standard deviations higher than nonprogram students on reading and writing and 1.20 standard deviations higher on the performance assessment. These are modest to strong effect sizes, although they are strikingly smaller than those produced among TBE students. (Note: We will discuss the substantial differences between standardized and performance assessment effect sizes after reviewing the 5th grade results.)
5th Grade Pre-Measures: Comparability Check
The fifth grade samples are comparable on both pre-measures (see Table 7). As in grade 2, the two 2-way ANOVAs produced no significant main effects for treatment group and no interactions (all p-values > .79), but there was a significant main effect for language-programming option (reading: 1,42 df, F = 98.83, p < .01; language: 1,42 df, F 48.70, p < .01). Although not directly germane to this study's focus (i.e., end-of-year outcomes), the very low levels of English achievement among ELDP students is striking. While TBE students are performing at or close to the national norm, albeit in Spanish (50th NCE), ELDP students, after four years of all-English instruction, are performing well below (21st NCE).
Table 7: Results on Pre-Measures for the 5th Grade Matched Samples (End of Grade 4).
Group ELDP
n = 14 & 14
Language of Test is English TBE
n = 33 & 33
Language of Test is Spanish ALL
n = 47 & 47
mean sd mean sd mean sd
Standardized Program 21.20 8.87 50.68 14.53 41.90 18.83
Reading Non 21.69 9.19 50.80 14.24 42.13 18.60
Standardized Program 25.25 8.49 48.79 17.57 41.78 18.80
Language Non 25.34 9.02 49.62 16.68 42.39 18.50
Notes: Standardized measures are Normal Curve Equivalents.
5th Grade Post-Measures: Testing for Treatment/Program Effects
Results of the post comparisons for 5th graders, all of which were conducted in English, are shown in Table 8. Program students performed significantly higher than nonprogram students on standardized measures of language achievement and the performance assessment, but not on standardized measures of reading. First, with regard to reading, there was no main effect for treatment group (1, 90 df, F = 2.62, p = .1088), no main effect for programming option (p > .40), and no significant interaction (F = 2.31, p = .1317). Inspection of means and effect sizes (here used for descriptive purposes only) shows a noteworthy difference between results within the ELDP and TBE samples. Among ELDP (all-English) students, program and nonprogram means are virtually identical and noticeably low: 25.31 and 25.04, respectively (ES = +.03). In contrast there is a difference of + 8.9 NCEs between TBE program and nonprogram means, although again both means are fairly low: 31.98 and 23.08 (ES = +.68). (Note: With no significant interaction, we ruled out a post hoc test on this difference.)Table 8: Results on Post-Measures for the 5th Grade Matched Samples (End of Grade 5).
Group ELDP
n = 14 & 14
Language of Test is English TBE
n = 33 & 33
Language of Test is English ALL
n = 47 & 47
mean sd ES mean sd ES mean sd ES
Standardized Program 25.31 14.98 +.03 31.98 11.91 +.68 30.00 13.10 +.52
Reading Non 25.04 9.85 23.08 13.11 23.66 12.16
Standardized Program 35.37 15.43 +.72 45.58 16.25 +.81 42.54 16.53 +.78
Language Non 27.04 11.59 30.74 18.30 29.64 16.55 **
Performance Program 2.77 .55 + 2.09 2.90 .52 + 2.51 2.86 .52 +2.40
Assessment Non 2.06 .34 2.17 .29 2.14 .30 **
Notes: Standardized measures are Normal Curve Equivalents. Performance assessments are scored on a 5-point scale, where 3 is approaching and 4 is meeting challenging grade level standards.
* p < .05; ** p < .01.
In contrast to reading, the ANOVA on language measures produced a significant treatment main effect (F = 9.90; p < .01) but no significant main effect for programming option (F = 3.57, p > .06) and no significant interaction (p > .35). On average, program students scored .78 standard deviations higher than nonprogram students and effect sizes within both the ELDP and TBE samples are of approximately the same magnitude: +.72 for ELDPs and +.81 for TBEs. Moreover, means for both ELDP and TBE program students are noticeably higher in English language (35.37 and 45.58, respectively) than in English reading (25.31 and 31.98, respectively).
Performance assessment results are similar to those for language achievement: a significant treatment effect (F = 55.83; p < .01), no significant main effect for programming (F = 1.59, p > .21) and no significant interaction (p > .92). The overall effect size is +2.40 and effect sizes within the ELDP and TBE groups are both substantial: +2.09 and +2.51, respectively.
Synthesis of Results and Study Limitations
The language arts transition program is associated with overall higher levels of literacy attainment for English learners. However, the following qualifications apply:
1) Program effects are noticeably stronger on TBE students than ELDP students. As indicated by the effect sizes, TBE program students consistently outperformed their nonprogram matched pairs to a greater extent than ELDP program students, both at grade 2 in Spanish and at grade 5 in English. In fact, at grade 5, TBE program students outperformed ELDP program students on all English literacy measures, although we do not know if differences between TBE and ELDP program students are significant, nor do we have any basis for gauging the comparability of the two groups.
2) Program effects on reading and language vary by grade level. At grade 2, program effects on reading and language are of a similar magnitude (respectively, +.92 and +94 overall, with similar consistency though of different magnitudes for TBE, +1.36 and +1.37, and ELDP students, +.34 and +.43). At grade 5, however, when all assessments are conducted in English, program effects are clearly stronger and more robust in language than in reading (for language and reading respectively, +.78 and +.53 overall; +.81 and +.68 for TBEs and +.72 and +.03 for ELDPs).
3) Program effects appear most dramatically on performance-based measures, but this should be interpreted cautiously. TBE and ELDP program students at both 2nd and 5th grade scored no less than a full standard deviation (in some cases 2) higher than their nonprogram matched pairs on the performance assessment. Given the corroborating results from most of the standardized test comparisons, there is no doubt a real difference indicated in the performance assessments results. However, program students were administered the assessment in their own classrooms by their regular teacher; nonprogram students were administered the assessment on a pull-out basis by a project teaching assistance. Despite a common proctor's guide, the perfomance assessment can be proctor-sensitive, and it is possible that the different administration conditions influenced scores on the performance assessment. Nonetheless, we include these data since they are informative about authentic literacy tasks in which program students were gaining proficiency.
We explicitly encourage program teachers to help students learn to do these tasks (summarizing familiar stories, reading and summarizing stories and articles on demand, writing story endings, and answering inferential questions), and this also confounds the performance assessment comparisons. While we are pleased to see that program students seem to be performing with increasing success on these meaningful reading-writing tasks, our experience suggests these kinds of tasks are not widely used and taught, especially in the lower grades (where, not surprisingly, means for all nonprogram samples were low, standard deviations were small, and effect sizes were uniformly very large). If students are not receiving related instruction, the performance assessment likely underestimates underlying reading and language skills.
4) Results are based on a best-case comparison. We are studying the effects of the language arts transition program as it has been implemented by skilled and dedicated teachers working in a school context that supports high levels of implementation of instructional innovations. We are not studying the effects of the program independent of these conditions. Teachers in this study meet regularly to learn about, discuss, try out, and receive feedback on their implementation of the program’s different components. By their own account, teachers were participating in what they considered to be the strongest professional development opportunity they have ever experienced. In addition, and consistent with our prior and ongoing research into creating more effective school contexts for English learners, we have assisted project school administration and staff in improving identification, assessment, class organization and programming oversight for limited-English-speaking students (Goldenberg, Korostoff, & Saunders, 1998). This assistance and the improvement we have witnessed at the project school are likely to contribute to the increased achievement among English learners we have documented here.
5) The program has not been tested against an equivalent group of students using an established coherent program (e.g., Success for All, Project M.O.R.E.). While the nonprogram sample provides a carefully matched and meaningful reference group for comparing outcomes, nonprogram students as a group represent a range of language arts pedagogies, varying in emphasis and quality. So while we can say the the language arts transition program seems to help produce better results for English learners’ achievement, we cannot answer with any certainty, "Better than what?" Even our use of matched pairs to evaluate the program’s effects can be seen as problematic, since matching by individual scores at pretest can only appear to create equivalence (Campbell & Stanley, 1963). Ideally, this program (as with so many other programs that claim efficacy) should be subject to an experimental evaluation. But the education of English learners is an area of study notable for its very difficult issues of research design and methodology (August & Hakuta, 1997), even more so than other areas within education. Nonetheless, there is precedent for using a matched-pairs design to evaluate the effectiveness of an educational intervention (e.g., Slavin, Madden, Dolan, & Wasik, 1996). Readers should be aware, however, of its limitations.
DISCUSSION
We began the study knowing, on the basis of previous evaluations, that the language arts transition program we are investigating produces superior levels of achievement when compared to the school district’s standard approach to transition (Saunders, 1998). However, previous evaluation studies were (1) conducted with different teachers at different school sites where the program was initially developed and (2) focused solely on implementation within Transitional Bilingual Education programming. We did not know prior to this study the extent of program effects at the new project site where we have been attempting to implement and replicate effects in a context where some students had been in a transitional bilingual education program and others in an all-English environment.
We found that program effects previously observed were replicated for English learners in the bilingual program. It was for these students that the programs served as a true "transition" program, helping them make a more successful academic transition from Spanish to English instruction. However, the program was noticeably less successful for English learners in all-English (ELDP) programming. In general our sample was too small to permit detailed analysis, but it was still clear that relative to the students who had received earlier instruction in Spanish, the performance of program students in all-English since the beginning of their school careers was extremely low. Their very poor performance might be attributable to their language of instruction and the absence of primary language learning opportunities. The most direct interpretation is that as currently conceived and implemented, the language arts transition program may not be very effective in the absence of solid grounding in primary language and literacy development. The transition program was developed and conceived with students in mind who are proficient in native language literacy skills. Students in all English programs lack basic proficiency in their primary or their second language, therefore are less able to take advantage of the language arts transition program. They might need curriculum and instruction tailored to much lower levels of language and literacy skills. Indeed, these data provide indirect evidence to bolster the claims of those who argue that language and literacy instruction in the primary language provides a more substantive foundation for subsequent language and literacy development in the second language than does direct immersion into a second language.
Continued implementation and program refinement in both bilingual and all-English programming contexts should assign a high priority to reading achievement. The weaker reading effects we obtained might be explained by the implementation ratings, which clearly point to the possibility that comprehension strategies are simply not being consistently taught. We will soon be working toward systematizing, implementing, and testing that component of the language arts program. Another possibility that might explain the relatively poor showing in reading is that our language arts transition program has paid insufficient attention to promoting fluent word recognition in English by English learners. We are currently exploring the inclusion of components designed to address explicitly English orthography, phonics, and decoding and automaticity in word recognition and oral reading, all of which are known to make important contributions to reading development in any alphabetically-based language (Snow, Burns, & Griffin, 1998).
APPENDICES
Table A1: Descriptions of each component in the language arts model.
Literature Units (experience-text-relationship approach): On average, students engage in 4 literature units across the year. Titles are chosen to fit the students' grade level and language proficiency (in particular across Transition I & II). The literature unit is propelled by an ongoing process of reading, writing (literature logs), and discussion (instructional conversations). Discussions are conducted in small groups of 6-10 students and managed through a specifically designed rotation system. The instructional framework for the literature units is called Experience-Text-Relationship (Mason and Au, 1986): Throughout the course of the unit, the teacher tries to help students understand the relationship between their own experiences, the content of the literary selection, and one or more major themes that apply to the selection (e.g., friendship, sacrifice, perseverance, commitment, justice, cultural identity). In addition to those three critical elements (experience, text, theme), the teacher enriches the unit with lessons, activities, supplementary readings that build background knowledge necessary for developing a deeper understanding of the selection and theme(s). Typically, units culminate with a writing project (see Culminating Writing Projects) through which students elaborate on some aspect of the literature unit.
Literature Logs: Teachers divide the literary selection into "chunks" (manageable portions of reading) and assign a literature log entry for each chunk. Students complete the log entry at an independent center, and typically small group discussions begin with some or all students sharing their logs. Literature log prompts might ask students to (a) write about a personal experience (related to the story), (b) elaborate on something that has happened in the story (e.g., assume the role of the character), or (c) analyze/interpret some aspect of the story or theme. In preparing a literature unit, teachers develop specific log prompts for each chunk, but often times, prompts emerge naturally from small group discussions.
Instructional Conversations (small group discussions): Throughout the course of the literature unit, teacher and students meet in small groups to discuss the story, log entries, related personal experiences and the theme(s) for the unit. The amount of time allotted to the discussion segment and frequency vary from teacher to teacher, but on average students spend at least 45 minutes a week engaged in discussion. The discussion provides the teacher with the opportunity to (a) hear students articulate their understanding of the story, theme(s) and related personal experiences, and (b) in the process of facilitating the discussion, challenge but also help students to enrich and deepen their understandings. Facilitated by the teacher, the small group discussions, also referred to as Instructional Conversations (Goldenberg, 1992/93), allow students to hear, appreciate, and build on each others' experiences, knowledge, and understandings.
Culminating Writing Projects (writing-as-a-process approach): On average, students complete 4 major writing projects across the year, taking the pieces through the entire process of writing: prewriting, drafting, sharing, receiving feedback, revising, editing, and preparing a final, polished piece of work (Calkins, 1986, 1991; Graves, 1983, 1991). Typically these projects are directly related to the literature units which conclude with a culminating writing (e.g., fully developing a literature log, or a writing assignment tailored to the themes and content of the literature study). The key to this process is revision. Three things seem to promote meaningful revision: (a) helping students learn to share and receive/provide feedback, (b) discussing examples (student or published) of the kind of writing students are working on, highlighting for students things they might incorporate in their own pieces when they revise, and (c) one-on-one conferences with the teacher.
Comprehension Strategies: Students are taught specific strategies to use while they are reading in order to monitor their own comprehension (McNeil, 1984; Palinscar & Brown, 1985). The two essential strategies involve students pausing intermittently during reading to summarize what they've read and formulate and answer test-like questions about the reading material. Strategies are introduced during two week training modules provided at the beginning and middle of the year. Students practice the strategies in pairs at the assigned independent reading center.
Table A1 continued: Descriptions for each component in the language arts model.
Assigned Independent Reading: Students are regularly assigned reading selections from available materials (basals, Literature titles, and any other sources) to read independently. Optimally, selections are related to the themes and topics being discussed in the Literature Units. Students complete various accompanying assignments to promote comprehension and hold the students accountable for what they read (summaries, comprehension questions, graphic organizers, paired and group activities). Readings and assignments are completed in-class as part of an independent center and/or for homework.
Dictation: The most extensive dictation program (Seeds University Elementary School, UCLA, 1992) includes: Students engage in dictation exercises weekly, taking a cold dictation of a grade level appropriate passage (at the beginning of the week), studying the features of that particular passage and practicing the dictation (throughout the week), and then completing a final dictation (at the end of the week). But as we've found, even a less extensive dictation program (2 times per week) is beneficial. Two elements are critical for successful dictation: 1) explanations from the teacher about language and punctuation items featured in the dictation passage, and 2) opportunities for the students to proofread and check their dictation against the actual passage.
Written Conventions Lessons: Students receive directed lessons about the conventions of written language (punctuation, capitalization, grammar, word usage). Lessons include a presentation from the teacher, opportunities for guided and independent practice, and then application to writings the students are working on (e.g., literature logs, writing projects, even Dictation passages). The key is connecting what is studied in the lessons to the actual writing students are doing.
Oral English Language Development...through Literature: Used in grades K-3, the ELD program is based on a natural language approach and children's literature (Beltran & O'Brien, 1993). Literature provides a meaningful, motivational, and enjoyable context for learning and practicing specifically targeted English oral language skills. It also exposes children to English print well in advance of formal transition to English reading. On average, students receive 30 minutes of ELD per day. Lessons are conducted in small groups organized by English language production level. Organizing groups by production level allows the teacher to focus more successfully on students' specific needs.
Pleasure Reading: A portion of the language arts time is set aside for students to select and read things on their own for pleasure and interest. Students keep and review with the teacher a record of their ongoing readings (reading inventory), and often times complete assignments related to their readings: preparing summaries and synopses, oral presentations for book sharing time, drawings, etc. Three things help promote pleasure reading: 1) teachers introduce students to numerous selections (trips to library, a full classroom library, lending read-aloud selections, making recommendations); 2) teachers explicitly teach students how to choose and try-out books (reading the cover synopsis, reading a portion of the book, reading various books from the same author); and 3) students have a chance to share and discuss with each other and teacher what they are reading.
Teacher Read-Alouds: At least 3 times per week, teachers read to students for approximately 20 minutes. Read-alouds (Trelease, 1985) serve various purposes: promote pleasure reading; expose students to the language of expert writers and the fluency of an expert reader; engage students in reading material they may not yet be able to read themselves; and increase students' familiarity with different genres of writing.
Interactive Journals: Used primarily in grades K-2 and at the beginning of transition, interactive journals provide students with regular, non-threatening opportunities to write about topics of their own choice and participate in a written dialogue with the teacher (Flores, et. al., 1991). Teacher response occurs as often as possible and provides students with examples of conventional writing. Interactive journals help K and 1 students break the written language code; and later in grade 1 and 2, they help students develop initial writing fluency. Transition teachers use interactive journals during the first semester of transition when students are making their first attempts at English writing. The immediate response from the teacher provides both emotional support for students and a highly contextualized and therefore comprehensible English text for students to read.
REFERENCES
Au, K. H. (1992). Constructing the theme of a story. Language Arts, 69, 106-111.
Au, K. H. (1979). Using the experience-text-relationship method with minority children. Reading Teacher, 32, 677-679.
August, D. & Hakuta, K. (Eds.) (1997). Improving schooling for language-minority children: A research agenda. Washington, DC: National Academy Press.
Berman, P., Chambers, J., Gandara, P., McLaughlin, B., Minicucci, C., Nelson, B., Olson, L., & Parrish, T. (1992). Meeting the challenge of language diversity: An evaluation of programs for pupils with limited English proficiency (Executive summary, Volume I). Berkeley, CA: BW Associates.
Beltran, D., & O'Brien, G. (1993). English language development through literature. Workshop and documentation presented at the annual conference of the California Association for Bilingual Education, Anaheim, CA.
Calkins, L. M. (1991). Living between the lines. Portsmouth, NH: Heinemann.
Calkins, L. M. (1986). The art of teaching writing. Portsmouth, NH: Heinemann.
Campbell, D. & Stanley, J. (1963). Experimental and quasi-experimental designs for research in teaching. In N. Gage (Ed.), Handbook of research on teaching. Chicago: Rand-McNally.
Cummins, J. (1989). Empowering minority students. Sacramento, CA: California Association for Bilingual Education.
Feldman, D., Gagnon, J., Hoffman, R., & Simpson, J. (1987). Statview II. Berkeley, CA: Abacus Concepts.
Fleischman, H., & Hopstock, P. (1993). Descriptive study of services to limited English proficient students, Vol. 1: Summary of findings and conclusions. Arlington, VA: Development Associates. Available: Development Associates, 1730 N. Lynn St., Arlington, VA 22209, (703) 276–0677.
Flores, B., Garcia, E., Gonzalez, S., Hidalgo, G., Kaczmarek, K, Romero, T. (1991). Bilingual holistic instructional strategies. Unpublished manuscript.
Gándara, P., & Merino, B. (1993). Measuring the outcomes of LEP programs: Test scores, exit rates, and other mythological data. Educational Evaluation and Policy Analysis, 15, 320-338.
García, E. (1992). Effective instruction for language minority students: The teacher. Journal of Education, 173(2), 130-141.
Gersten, R. (1996). Literacy instruction for language minority students: The transition years. Elementary School Journal, 96(3), 227-244.
Gersten, R., & Woodward, J. (1994). The language minority student and special education: Issues, themes, and paradoxes. Exceptional Children, 60(4), 310-322.
Gersten, R., & Jiménez, R. (1993, April). Language minority students in transition, Symposium presented at the annual conference of the American Educational Research Association, Atlanta, Georgia.
Goldenberg, C. (1996). Commentary: The education of language-minority students: Where are we, and where do we need to go? The Elementary School Journal, 96(3), 353-361.
Goldenberg, C., Korostoff, M., & Saunders, W. (1998). Using the School Change Model to implement a successful language arts program for English learners. (Year 2 Deliverable, Project 1.5: "Upscaling for Transition: Schoolwide Factors to Support Latino Students' Transition from Spanish to English Instruction"). Santa Cruz, CA: Center for Research on Education, Diversity, and Excellence, University of California, Santa Cruz.
Goldenberg, C., & Saunders. W. (1996). Using standards and assessments to promote excellence and equity among diverse students. In R. Blum & J. Arter (Eds.), Who's learning what? A resource guide for student performance assessment in the context of school restructuring (VII-7). Alexandria, VA: Association for Supervision and Curriculum Development.
Goldenberg, C. (1992/93). Instructional conversations: Promoting comprehension through discussion. The Reading Teacher, 46, 316-326.
Graves, D. H. (1991). Build a literate classroom. Portsmouth, NH: Heinemann.
Graves, D. H. (1983). Writing: Teachers and children at work. Portsmouth, NH: Heinemann.
Greenfield, E. (1993). Louella's song. In J. Pikulski (Sr. Author), Dinosauring (4th grade reader), 430-436. Boston: Houghton Mifflin.
Los Angeles Unified School District. (1988). Master plan for the education of limited-English-proficient students. Los Angeles, CA: Office of Bilingual-ESL Instruction, LAUSD.
Los Angeles Unified School District. (1988 ed. & 1996 ed.). Criteria for the addition of reading in English: Description and testing materials. Los Angeles, CA: Office of Bilingual-ESL Instruction, LAUSD.
Krashen, S. (1987). Principles and practice in second language acquisition. New York: Prentice-Hall.
Mason, J. M., & Au, K. H. (1986). Reading instruction for today. Glenview, ILL: Scott, Foresman and Company.
McNeil, J. D. (1984). Reading comprehension: New directions for classroom practice. Glenview, ILL: Scott, Foresman and Company.
National Clearinghouse for Bilingual Education. (1995, January 20). Ask NCBE: How many school-aged limited English proficient students are there in the U.S.? Washington, DC: Author. Available: NCBE, 1118 22nd St. NW, Washington, DC, or on the internet at <<gopher://gopher.ncbe.gwu.edu/>>
Palinscar, A., & Brown, A. (1985). Reciprocal teaching: A means to a meaningful end. In J. Osborn, P. Wilson, & R. C. Anderson (Eds.), Reading education: Foundations for a literate America (299-310). Lexington, MA: D.C. Heath & Co.
Ramírez, J. D. (1992). Longitudinal study of structured English immersion strategy, early-exit and late-exit transitional bilingual education programs for language minority children (Executive summary). Bilingual Research Journal, 16 (1 & 2), 1-62.
Saunders, W. & Goldenberg, C. (1999). The effects of instructional conversations and literature logs on limited and fluent English proficient students' story comprehension and thematic understanding. Elementary School Journal, 99, 277-301.
Saunders, W. & Goldenberg, C. (in press). The effects of instructional conversation on transition students' concepts of friendship and story comprehension. In R. Horowitz (Ed.), The evolution of talk about text: Knowing the world through classroom discourse. International Reading Association.
Saunders, W. (1998, April). Improving literacy achievement for English learners in transitional bilingual programs. Paper presented presented at the annual conference of the American Educational Research Association, San Diego, CA.
Saunders, W., O'Brien, G., Lennon, D., & McLean, J. (1998). Making the transition to English literacy successful: Effective strategies for studying literature with transition students. In R. Gersten & R. Jiménez (Eds.), Promoting learning for culturally and linguistically diverser students: Classroom applications from contemporary research. Belmont, CA: Wadsworth.
Saunders, W. & Goldenberg, C. (1997). Identifying salient elements of a successful transition program. (Year 1 Deliverable, Project 1.5: "Upscaling for Transition: Schoolwide Factors to Support Latino Students' Transition from Spanish to English Instruction"). Santa Cruz, CA: Center for Research on Education, Diversity, and Excellence, University of California, Santa Cruz.
Saunders, W., Patthey-Chavez, G., & Goldenberg, C. (1997). Reflections on the relationship between language, curriculum content and instruction. Language, Culture, and Curriculum, 10(1), 30-51.
Saunders, W. & Lennon, D. (1996). Developing academic literacy in English through a program of integrated language arts: Year 5 evaluation (Project TOO3AOO185). Washington, D.C.: Office of Bilingual Education and Minority Affairs, U.S. Department of Education.
Seeds University Elementary School (1992). Dictation program: An approach to written language development. Los Angeles, CA: Seeds University Elementary School, University of California, Los Angeles.
Slavin, R., Madden, N., Dolan, L., & Wasik, B. (1996). Every child, every school: Success for All. Corwin Press.
Slavin, R. (1984). Component building: A strategy for research-based instructional improvement. The Elementary School Journal, 84, 255-269.
Tharp, R., & Gallimore, R. (1988). Rousing minds to life. Cambridge, England: Cambridge University Press.
Thomas, W. & Collier, V. (1997). School effectiveness for language minority students. Washington, DC: National Clearinghouse for Bilingual Education. Available on the internet at http://www.ncbe.gwu.edu/ncbepubs/resource/effectiveness/ thomas-collier97.pdf
Trelease, J. (1985). The read-aloud handbook. New York: Penguin.
TABLESTable 1: The Design for Transition within Transitional Bilingual Programming.
Phases Optimal
Grade Goal Measurable
Outcome
K-2nd o Initial Reading & Writing Proficiency
(Spanish)
o Early Production II (oral English) Existing norm-or
criterion-referenced measures
Pre-Transition 3rd o Grade appropriate Reading & Writing
Achievement (Spanish)
o Speech Emergence (oral English) Pass CARE
(district transition instrument)
Transition
I 4th o Initial Reading and Writing Proficiency
(English)
o Academic Oral Language Proficiency (English)
o Grade appropriate Reading & Writing Achievement (Spanish) Existing norm-or
criterion-referenced measures
Transition
II 5th o Grade appropriate Reading & Writing
Achievement (English)
Redesignation:
LEP to FEP

Table 4: Teacher and Advisor Ratings of Consistency and Effectiveness of Program Implementation.
Implementation Consistency
Scale = 1-3 Language Arts Program Components Implementation Effectiveness
Scale = 1-4
Teachers Ratings
of Themselves Teachers' Ratings of Themselves Advisors'
Ratings of Teachers
Using Consistently Dictation 3.25 3.50
(mean = 3.00) Literature Units 3.00 3.50
Literature Logs 2.75 3.50
Assigned Independent Reading 2.50 3.25
Instructional Conversation 2.00 3.00
Consistently/Intermittently Teacher Read-Alouds 3.75 4.00
(mean = 2.50) Conventions Lessons 3.25 3.50
Pleasure Reading 3.25 3.25
Using Intermittently Interactive Journals 2.50 2.75
(mean = 2.00) Comprehension Strategies 2.00 2.25
Culminating Writing Projects 1.75 3.25
Mean of means 2.72 3.25
Note: The Oral English Language Development component is not shown here as it applies to only the 2nd grade teachers (averaged consistency rating was 3; averaged teacher and advisor effectiveness ratings were both 3).
Table 5: Results on Pre-Measures for the 2nd Grade Matched Samples (End of Grade 1).
Group ELDP
n = 7 & 7
Language of Test is English TBE
n = 16 & 16
Language of Test is Spanish ALL
n = 23 & 23
mean sd mean sd mean sd
Standardized Program 37.81 12.16 66.79 23.53 57.97 24.57
Reading Non 38.29 11.52 64.00 20.83 56.17 21.87
Standardized Program 37.63 10.26 48.39 20.03 45.12 18.11
Language Non 38.59 8.36 51.46 19.79 47.54 17.96
Notes: Standardized measures are Normal Curve Equivalents.

Table 6: Results on Post-Measures for the 2nd Grade Matched Samples (End of Grade 2).
Group ELDP
n = 7 & 7
Language of Test is English TBE
n = 16 & 16
Language of Test is Spanish ALL
n = 23 & 23
mean sd ES mean sd ES mean sd ES
Standardized Program 42.36 15.79 +.34 75.41 17.09 +1.36 65.35 22.56 +.93
Reading Non 36.16 17.92 54.05 15.67 48.60 18.05 *
Standardized Program 52.51 16.37 +.42 72.19 18.26 +1.37 66.20 19.65 +.94
Language Non 41.66 25.82 51.21 15.34 48.30 19.04 **
Performance Program 2.51 .99 +1.13 3.62 .50 +2.30 3.29 .84 +1.63
Assessment Non 1.89 .55 2.56 .46 2.36 .57 **
Notes: Standardized measures are Normal Curve Equivalents. Performance assessments are scored on a 5-point scale, where 3 is approaching and 4 is meeting challenging grade level standards.
* p < .05; ** p < .01.
Table 7: Results on Pre-Measures for the 5th Grade Matched Samples (End of Grade 4).
Group ELDP
n = 14 & 14
Language of Test is English TBE
n = 33 & 33
Language of Test is Spanish ALL
n = 47 & 47
mean sd mean sd mean sd
Standardized Program 21.20 8.87 50.68 14.53 41.90 18.83
Reading Non 21.69 9.19 50.80 14.24 42.13 18.60
Standardized Program 25.25 8.49 48.79 17.57 41.78 18.80
Language Non 25.34 9.02 49.62 16.68 42.39 18.50
Notes: Standardized measures are Normal Curve Equivalents.

Table 8: Results on Post-Measures for the 5th Grade Matched Samples (End of Grade 5).
Group ELDP
n = 14 & 14
Language of Test is English TBE
n = 33 & 33
Language of Test is English ALL
n = 47 & 47
mean sd ES mean sd ES mean sd ES
Standardized Program 25.31 14.98 +.03 31.98 11.91 +.68 30.00 13.10 +.52
Reading Non 25.04 9.85 23.08 13.11 23.66 12.16
Standardized Program 35.37 15.43 +.72 45.58 16.25 +.81 42.54 16.53 +.78
Language Non 27.04 11.59 30.74 18.30 29.64 16.55 **
Performance Program 2.77 .55 + 2.09 2.90 .52 + 2.51 2.86 .52 +2.40
Assessment Non 2.06 .34 2.17 .29 2.14 .30 **
Notes: Standardized measures are Normal Curve Equivalents. Performance assessments are scored on a 5-point scale, where 3 is approaching and 4 is meeting challenging grade level standards.
* p < .05; ** p < .01.

©2002 Center for Research on Education, Diversity & Excellence. All rights reserved.