This paper was written under funding from the Office of Educational Research and Improvement, U.S. Department of Education (Grant Nos. OERI-R-117-R90002 and R-117D-40005), and the Pew Charitable Trusts. However, any opinions expressed do not necessarily represent the positions or policies of our funders.

We would like to thank Katherine Conner, Allie Mulvihill, MaryLouise de Nicola, Renee Yampolsky, and Norris Eldridge, of the Philadelphia Public Schools, Marcie Dianda of the Southwest Regional Lab, Jerry Frye of the Modesto City School District, Carley Ochoa of the Riverside Unified School District, and Robert Petza, Rachel Nicholas, Larry Dolan, and Gretta Gordy of Johns Hopkins University for their help with this project.

     The education of English language learners is at a crossroads. For many years, researchers, educators, and policy makers have debated questions of the appropriate language instruction for students who enter elementary school speaking languages other than English. Research on this topic has generally found that students taught to read their home language and then transitioned to English ultimately become better readers in English than do students taught to read only in English (Garcia, 1991; Willig, 1985; Wong-Fillmore & Valdez, 1986). More recently, however, attention has shifted to another question. Given that students are taught to read their home language, how can we ensure that they succeed in that language? (See, for example, Garcia, 1994). There is no reason to expect that children failing to read well in Spanish, for example, will later become good readers and successful students in English. On the contrary, research consistently supports the common-sense expectation that the better students in Spanish bilingual programs read Spanish, the better their English reading will be (Garcia, 1991; Hakuta & Garcia, 1989). Clearly, the quality of instruction in home-language reading is a key factor in the ultimate school success of English language learners, and must be a focus of research on the education of these children.

     Even if all educators and policy makers accepted the evidence favoring bilingual over English-only instruction, there would still be large numbers of English language learners being taught to read in English. This is true because of practical difficulties of providing instruction in languages other than English or Spanish; teachers fully proficient in Southeast Asian languages, Arabic, and other languages are in short supply, as are materials to teach in these languages. Speakers of languages other than English or Spanish are among the fastest-growing groups in our nation's schools (GAO, 1994). Further, many Spanish-dominant students are taught to read in English, either because of shortages of bilingual teachers, insufficient numbers of Spanish-dominant students in one school, parental desires to have their children taught in English, and other factors. For these reasons, a large percentage of English language learners will always be taught in English only, with instruction in English as a second language (ESL). As with bilingual programs, the quality of reading instruction, ESL instruction, and the integration of the two are essential in determining the success of English language learners being taught in English only.

     The renewed focus since the late 1980's on the quality of bilingual and ESL programs has led to numerous observational and descriptive studies of effective education for English language learners (see, for example, Fleishman & Hopstock, 1993; Leighton et al., 1993; Garcia, 1987; Tikunoff et al., 1991). However, few studies have directly compared outcomes of innovative bilingual or ESL programs to traditional programs (see Ramirez, 1986).

     There is remarkably little research evaluating programs designed to increase the Spanish reading performance of students in bilingual programs. Hertz-Lazarowitz, Ivory, & Calderon (1993) evaluated a bilingual adaptation of Cooperative Integrated Reading and Composition (BCIRC) in El Paso elementary schools starting in second grade. This program, based on a successful program originally developed in English for English proficient students (Stevens, Madden, Slavin, & Farnish, 1987; Stevens & Slavin, 1995), involves having students work in small cooperative groups. Students read to each other, work together to identify characters, settings, problems, and problem solutions in narratives, summarize stories to each other, and work together on writing, reading comprehension, and vocabulary activities. Students in BCIRC classes scored significantly better than control students on the Spanish Texas Assessment of Academic Skills (TAAS) at the end of second grade, and as they transitioned to English in third and fourth grades they performed significantly better than control students on standardized reading tests given in English.

     While it is important to improve the outcomes of bilingual and English-only reading instruction for English language learners at all grade levels, there is a particular need to see that students are successful in beginning to read in first grade. Many students fail to read adequately in this critical year and then never become good readers (see for example, Juel, 1988). These students are therefore at risk for being retained in grade or assigned to special education or long-term remedial services, all of which are key predictors of ultimate dropout (Lloyd, 1978). Latino students, with one of the highest dropout rates of all ethnic groups (Bureau of Census, 1992; Durán, 1983), are particularly at risk if they do not read well.

     Of course, the problems of early reading failure are by no means unique to limited English proficient students or to Latino students more generally. Every school has first graders who fail to read adequately and are therefore placed at great risk, and schools serving many disadvantaged children, regardless of ethnicity, have rates of reading failure above the norm (Mullis et al., 1991).

     If all students are to achieve their potential in school, all must begin with success in first grade reading. One program that has achieved a great deal of success in meeting this goal is called Success for All, a comprehensive model for restructuring elementary schools that focuses on prevention and early, intensive intervention. The program's philosophy is that learning problems must first be prevented by providing students with high-quality instruction from prekindergarten or kindergarten onward, improving school-family links, and assessing student progress on a regular basis. When problems appear despite effective preventive measures, interventions must be applied immediately and intensively to solve them before they become serious. In particular, one-to-one tutoring is provided to first graders who are failing to read well. The English version of Success for All has been evaluated in comparison to matched control schools in seven school districts throughout the U.S. and found to be consistently effective on measures of reading, reductions in retention and special education placements, and other outcomes (Slavin et al., 1994).

     The first application of Success for All to English language learners began in Philadelphia's Francis Scott Key School, which serves a high-poverty neighborhood in which more than 60% of students enter the schools speaking Cambodian or other Southeast Asian languages. An adaptation of Success for All was designed to meet the needs of these children. This adaptation focused on integrating the work of ESL teachers and reading teachers, so that ESL teachers taught a reading class and then helped limited English proficient students with the specific language and reading skills needed to succeed in the school's (English) reading program. In addition, a cross-age tutoring program enabled fifth graders, now fully bilingual in English and Cambodian, to help kindergartners succeed in the English program. The performance of students at Francis Scott Key has been compared to that of students in a matched comparison school each year, and the results have consistently favored Success for All (for Asian as well as non-Asian students (Slavin & Yampolsky, 1991). The present paper reports the reading performance of the English language learners at Key and its comparison school as of spring, 1994, the end of the sixth year of program implementation.

     In 1992, a Spanish adaptation of the Success for All reading program called Lee Conmigo ("Read With Me") was developed for use in Spanish bilingual programs. During the 1992-1993 school year the entire Success for All program (including Lee Conmigo for LEP students) was implemented in one Philadelphia school serving a predominately Latino (mostly Puerto Rican) student body. The first year results showed the Spanish bilingual students to be performing substantially better than controls on individually administered tests of Spanish (Slavin & Madden, 1994). This paper reports the results for the second graders who completed their second year in Lee Conmigo.

     A third evaluation of Success for All with English language learners was carried out by Marcella Dianda (1995) at the Southwest Regional Laboratory in Southern California. This study, reported in the following paper in this symposium, involved three schools. Fremont Elementary in Riverside, California, and Orville Wright Elementary in Modesto, are schools with substantial Spanish bilingual programs. The third, El Vista Elementary, also in Modesto, served a highly diverse student body speaking 17 languages using an ESL approach. Students in all three schools were compared to matched students in matched schools. In each case, students were assessed in the language of instruction (English or Spanish).

The present paper summarizes the results from all five of the schools implementing Success for All with English language learners.

Success for All: Program Description

     Success for All is a comprehensive reform program for elementary schools, especially those serving many students placed at risk. It restructures Title I staff and resources, plus any other available resources (such as special education or state compensatory education), to focus on prevention, early intervention, and long-term professional development, instead of remediation. Specific elements of the program, and adaptations for the needs of English language learners, are described in the following sections.

Reading Tutors

     One of the most important elements of the Success for All model is the use of tutors to support students' success in reading. One-to-one tutoring is the most effective form of instruction known (see Wasik & Slavin, 1993a). The tutors at Fairhill, Fremont, and Wright, the schools using Lee Conmigo, were Spanish bilingual teachers. At Key and El Vista, tutors were certified teachers paid for by Chapter I funds, plus ESL teachers from the schools' staffs. Tutors worked one-to-one with students who were having difficulties keeping up with their reading groups. Students were taken from their homeroom classes by the tutors for 20-minute sessions during times other than reading or math periods. In general, tutors supported students' success in the regular reading curriculum, rather than teaching different objectives. For example, if the regular reading teacher was working on stories with long vowels or was teaching comprehension monitoring strategies, so did the tutor. However, tutors identified learning deficits and use different strategies to teach the same skills.

     During daily 90-minute reading periods, tutors served as additional reading teachers to reduce class size for reading. Information on students' specific deficits and needs passed between reading teachers and tutors on brief forms, and reading teachers and tutors were given regular times to meet to coordinate their approaches with individual children.

     Initial decisions about reading group placement and need for tutoring were made based on informal reading inventories given to each child by the tutors. After this, reading group placements and tutoring assignments were made based on eight-week assessments, which included teacher judgments as well as more formal assessments. First graders received first priority for tutoring, on the assumption that the primary function of the tutors is to help all students be successful in reading the first time, before they become remedial readers.

Reading Program

     Students in grades 1-3 were regrouped for reading. That is, students were assigned to heterogeneous, age-grouped classes with class sizes of about 25 most of the day, but during a regular 90-minute reading period they were regrouped according to reading performance levels into reading classes of about 15 students all at the same level. For example, a 2-1 (second grade, first semester) reading class might contain first, second, and third grade students all reading at the same level. At the bilingual schools this regrouping was done separately for Spanish-dominant and English-dominant students; at Key and El Vista, all students were regrouped according to reading level, regardless of language background. Regrouping allows teachers to teach the whole reading class without having to break the class into reading groups. It is a form of the Joplin Plan, which has been found to increase reading achievement in the elementary grades (Slavin, 1987).

     The reading program emphasizes development of basic language skills and sound and letter recognition skills in kindergarten, and uses an approach based on sound blending and phonics starting in first grade. The K-1 reading program used in the bilingual program at Fairhill, Fremont, and Wright, Lee Conmigo, was built around the Macmillan Campanitas de Oro basals. English-dominant students in all schools experienced the same instructional methods, but in first grade used a series of "shared stories" and other materials designed for Success for All. This program emphasizes oral reading to partners as well as to the teacher, instruction in story structure and specific comprehension skills, and integration of reading and writing. It provides a rapidly paced, engaging set of routines that involve students in group response games that develop auditory discrimination skills, letter name and letter sound recognition and sound blending strategies based on the sounds and words used in the books. When they reach the primer reading level, students use a form of Cooperative Integrated Reading and Composition (CIRC) with Spanish or English novels and basals. CIRC uses cooperative learning activities built around story structure, prediction, summarization, vocabulary building, decoding practice, writing, and direct instruction in reading comprehension skills. Research on CIRC has found it to significantly increase students' reading comprehension and language skills in English (Stevens, Madden, Slavin, & Farnish, 1987) and in Spanish (Hertz-Lazarowitz et al., 1993).

Eight-Week Reading Assessments

Every eight weeks, reading teachers assessed student progress through the reading program. The results of the assessments were used to determine who is to receive tutoring, to suggest other adaptations in students' programs, and to identify students who need other types of assistance, such as family interventions or vision/hearing screening.

English as a Second Language

All schools had instruction in English as a second language (ESL). At Key and El Vista, ESL teachers taught regular reading classes during a common regrouped reading period. After this period, they tutored individual students or worked with groups of limited English proficient students. The emphasis of the ESL program in Success for All was on giving students assistance that is directly tied to success in the English curriculum. For example, ESL teachers used the same reading materials used in the classroom reading program. At Fairhill, Fremont, and Wright, ESL instruction was also closely connected to instruction in subjects in which students were being taught in English.

Kindergarten

All schools provided a full-day kindergarten for all eligible students. The kindergarten program provided a balanced and developmentally appropriate learning experience for young children. The curriculum emphasizes the development and use of language. It provides a balance of academic readiness and non-academic music, art, and movement activities. Readiness activities include use of integrated thematic units, and a program called Story Telling and Retelling (STaR) in which students retell stories read by the teachers.

Family Support Team

A Family Support Team in each provided parenting education and worked preventively to involve parents in support of their children's success in school. Also, family support staff provided assistance when there were indications that students were not working up to their full potential because of problems at home. For example, families of students who are not receiving adequate sleep or nutrition, need glasses, are not attending school regularly, or are exhibiting serious behavior problems receive family support assistance. Links with appropriate community service agencies were made to provide as much focused service as possible for parents and children.

Program Facilitator

A program facilitator worked at each school full time to oversee (with the principal) the operation of the Success for All model. Facilitators helped plan the Success for All program, helped the principal with scheduling, and visited classes and tutoring sessions frequently to help teachers and tutors with individual problems. They helped teachers and tutors deal with any behavior problems or other special problems, and coordinated the activities of the classroom teachers, tutors, Family Support Team, ESL teachers, and others.

Teachers and Teacher Training

The teachers and tutors were regular classroom teachers, bilingual teachers, or ESL teachers. They received detailed teacher's manuals supplemented by two days of inservice at the beginning of the school year and several inservice sessions throughout the year on such topics as classroom management, instructional pace, and implementation of the curriculum.

METHODS AND RESULTS

Francis Scott Key (ESL)

     Beginning in September 1988, researchers from The Johns Hopkins University began working with the staff at Philadelphia's Francis Scott Key Elementary School to implement Success for All in grades K-3. Sixty-two percent of its students were from Asian backgrounds, primarily Cambodian. Nearly all of these students enter the school in kindergarten with little or no English. The remainder of the school was divided between African American and white students. The school is located in an extremely impoverished neighborhood in South Philadelphia. Ninety-six percent of the students were from low-income families and qualified for free lunch.

     Because of the unavailability of Cambodian-speaking teachers, Francis Scott Key used an ESL approach to its LEP students. The only adult in the school who spoke Cambodian was a bilingual counseling assistant.

Evaluation Design

     The program at Francis Scott Key was evaluated in comparison to a similar Philadelphia elementary school. Table 1 compares the two schools on several variables. As the Table shows, the two schools were very similar in overall achievement level and other variables. Thirty-three percent of the comparison school's students were Asian (mostly Cambodian), the highest proportion in the city after Key. The percentage of students receiving free lunch was very high in both schools, though higher at Key (96%) than at the comparison school (84%). A few differences are worthy of note, however. The comparison school was larger than Key, with 1,128 students overall to Key's 622, and the non-Asian students at the comparison school were almost all African American, while 21% of Key's students were white.

     The data reported here are for all students in grades 3-5 in Spring, 1994. With the exception of transfers, all students had been in the program since kindergarten.

Measures

     At Francis Scott Key and its comparison school, all students in grades 3-5 were individually administered the Woodcock Language Proficiency Battery (Woodcock, 1984). The Woodcock scales, Word Identification, Word Attack, and Passage Comprehension, were given to students in grades 3-5. The Word Identification scale was used to assess recognition of common sight words, the Word Attack scale assessed phonetic synthesis skills, and the Passage Comprehension scale assessed students' abilities to read and comprehend meaningful text.

     Analyses of variance (ANOVA) were conducted on each outcome separately. Outcomes were characterized in terms of effect sizes, which are the difference between experimental and control means divided by the control group's standard deviation. Grade equivalents were not used in any analyses, but are presented as convenient indicators of students' absolute performance levels.

Results: Asian Students

     The results for Asian students are summarized in Tables 2-4. Success for All Asian students at all three grade levels performed far better than control students. Differences between Success for All and control students were statistically significant on every measure at every grade level (p<.001). Median grade equivalents and effect sizes were computed across the three Woodcock scales. On average, Success for All Asian students exceeded control in reading grade equivalents by almost three years in third grade (Median ES = +1.76), more than 2 years in fourth grade (Median ES = +1.46), and about three years in fifth grade (Median ES = +1.44). Success for All Asian students were reading more than a full year above grade level in grade 3 and more than a half-year above in fourth and fifth grade, while similar control students were reading more than a year below grade level at all three grade levels.

Results: Non-Asian Students

     Outcomes of Success for All for non-Asian students, summarized in Tables 2-4 and Figure 1, were also very positive in grades 3-5. Experimental-control differences were statistically significant (p<.05 or better) on every measure at every grade level. Effect sizes were somewhat smaller than for Asian students, but were still quite substantial, averaging +1.00 in grade 3, +0.96 in grade 4, and +0.78 in grade 5. Effect sizes were particularly large for the Passage Comprehension measure at all three levels. Success for All students averaged almost two years above grade level in third grade, more than a year above grade level in fourth grade, and about eight months above grade level in fifth grade; at all grade levels, Success for All averaged about 2.5 years higher than control students.

Fairhill (Bilingual)

     The bilingual version of Success for All, Lee Conmigo, was first implemented at Fairhill Elementary School, a school in inner-city Philadelphia. Fairhill serves a student body of 694 students of whom 78% are Hispanic (primarily from Puerto Rico) and 22% are African-American. A matched comparison school was also selected. Table 5 shows data on the two schools. From the table it is clear that the two schools were very similar in total enrollment, percent Hispanic and African-American, and historical achievement levels (from district records). The schools were also similar in the percent of students receiving instruction in Spanish. In both schools about half of all students were in the bilingual program in first grade. Nearly all students in both schools qualified for free lunches. Both schools were Chapter 1 schoolwide projects, which means that both had high (and roughly equivalent) allocations of Chapter 1 funds that they could use flexibly to meet student needs.

Measures

     All students defined by district criteria as LEP at Fairhill and its control school were pretested at the beginning of first grade on the Spanish Peabody Picture Vocabulary Test (PPVT). Each following May, these students were tested by native language speakers on three scales of the Spanish Woodcock (Bateria Woodcock de Proficiencia en el Idioma): Letter/Word Identification (Identificacion de Letras y Palabras), Word Attack (Analisis de Palabras), and Passage Comprehension (Comprension de Textos).

Results

     A check for pretest differences on the Spanish PPVT found that there were differences in favor of the experimental group (p< .03). PPVT scores were therefore used as covariates in all analyses of covariance (ANCOVA). These analyses showed that Success for All students scored substantially higher than controls on every measure (p< .01 or better). Table 6 shows the adjusted means, standard deviations, grade equivalents, and effect sizes. Figure 2 summarizes mean grade equivalents and effect sizes. Control second graders scored far below grade level on the all three scales. In contrast, Fairhill students averaged near grade level on all measures. Effect sizes on all measures were substantial. Fairhill students exceeded control by 1.8 standard deviations on Letter-Word Identification, 2.2 on Word Attack, and 1.3 on Passage Comprehension.

Fremont (Bilingual), Wright (Bilingual), and El Vista (ESL)

     Data from first graders in the three California Success for All schools were analyzed together by Dianda (1995), pooling data across schools in four categories: English-dominant students, Spanish-dominant students taught in Spanish (Lee Conmigo in Success for All schools), Spanish-dominant students taught in English ("sheltered students"), and speakers of languages other than English or Spanish taught in English. The pooled results are summarized in Figure 3 (from Dianda, 1995).

     As is clear in Figure 3, all categories of Success for All students scored substantially better than control students. The differences were greatest, however, for Spanish-dominant students taught in bilingual classes (ES = +1.03) and those taught in sheltered English programs (ES = +1.02). The bilingual students scored at grade level, and more than six months ahead of controls. The sheltered students scored about two months below grade level, but were still four months ahead of their controls. Both English-speaking students and speakers of languages other than English or Spanish scored above grade level and about two months ahead of their controls.

Discussion

     The effects of Success for All on the achievement of English language learners are substantially positive. Across three schools implementing Lee Conmigo, the Spanish curriculum used in bilingual Success for All schools, the average effect size for first graders on Spanish assessments was +0.88; for second graders (at Philadelphia's Fairhill Elementary) the average effect size was +1.77 For students in sheltered English instruction, effect sizes for all comparisons were very positive, especially for Cambodian students in Philadelphia and Mexican-American students in California.

     While the performance of English language learners in Success for All was quite positive, what is more striking across all five school evaluations is how poorly the control groups performed. In all three schools evaluating Lee Conmigo, control first graders consistently scored near the floor of the Spanish Woodcock Scales. Across all three control groups, first graders averaged a grade equivalent of 1.1, indicating little skill in reading. Looking at individual Woodcock scales, it is apparent that while control students did have some word attack and letter-word identification skills, their passage comprehension scores were near zero. Similarly, second graders at Fairhill's control school scored below grade level on all scales, but especially on passage comprehension.

     The extraordinarily low performance of these control students is not typical of control students in other Success for All evaluations. Across fifty cohorts of first graders studied over the years, English-only first grade control groups (primarily African-American students in high-poverty schools) have averaged a grade equivalent of 1.6 (Slavin, Madden, Dolan, and Wasik, 1995), well below grade level, but far ahead of the Spanish control groups' average of 1.1. In contrast, Spanish bilingual first graders' scores in Success for All schools averaged 1.9, close to the English-only average of 2.1 for all Success for All first graders (Slavin et al., 1995).

     In the longitudinal evaluation of Success for All at Philadelphia's Key School, the situation was somewhat different from that of the bilingual evaluations. There, the Cambodian control students did score well below grade level, but the more striking observation was the extraordinarily high performance of the Cambodian third, fourth, and fifth graders in the Success for All school. These students scored substantially above grade level, averaging 2.5 years above controls at all three grade levels.

     The low performance of English language learners in the control groups could be an artifact of the test, such as a difference between the norming of the English and Spanish Woodcock scales. However, the much higher performance of the bilingual Success for All students makes this explanation unlikely. If the results obtained here are valid and generalizable to other high-poverty schools serving English language learners, they raise some disturbing questions. Perhaps the shortage of bilingual teachers is leading schools to hire teachers with inadequate skills in teaching or in the language of instruction. Perhaps teacher preparation programs for bilingual teachers are inadequate. Perhaps bilingual teachers have adopted reading strategies that are particularly ineffective, especially in teaching comprehension skills. The very positive results for Success for All bilingual programs indicate that existing bilingual teachers are certainly capable of doing an outstanding job of teaching first graders to read, but there is clearly a need for better professional development and better instructional models for these teachers. The findings of this research suggest many areas in need of further investigation. First, they point to a need for more in depth qualitative investigations of instructional practices in traditional bilingual first grades, as well as in bilingual Success for All classes. Such an investigation and ethnography of Success for All and control schools is currently under way in Houston. In addition, it would be important to investigate the effects of the separate components of Success for All in bilingual and ESL classes and to relate these components to student outcomes. This is also a component of the ongoing Houston study, which is contrasting bilingual schools using the Lee Conmigo reading curriculum alone, schools using Lee Conmigo plus tutoring for first graders, schools using all components of Success for All, and traditional bilingual control schools. This study includes degree-of-implementation observations so that it will be possible to relate use of particular program elements and the quality of implementation to program outcomes. There is a need to continue to follow students in the schools studied so far, in particular to assess reading performance in English as bilingual students make the transition to English-only reading instruction. Continued assessments are under way in all five schools (and their controls) described in this paper.

Conclusion

     The research summarized in this paper supports two principal conclusions. First, the performance of English language learners in high-poverty schools is very poor, whether they are taught in English or in their home language (and assessed in the language of instruction). Second, this need not be the case. In every evaluation, English language learners in Success for all schools have scored substantially better than their control conterparts, and in all but one case (Spanish-speaking students in a sheltered English program in Modesto), these students scored at or above grade level on individually administered tests. More research is needed to better understand how Success for All affects daily practices in schools serving English language learners and to understand how these practices differ from those typical of traditional bilingual and ESL programs.

     More research is also needed to determine the effects of Lee Conmigo over a longer time period and in a larger number of schools. However, this study shows the impact of a structured approach to beginning reading in Spanish that emphasizes teaching phonics in the context of meaningful text, cooperative learning, story telling, and tutoring. A similar approach integrating ESL and classroom instruction for English language learners taught in English was also found to be effective.

     Both bilingual and ESL instruction are realities for hundreds of thousands of U.S. students. It is time to move beyond the debate about the relative benefits of each and to begin to investigate instructional strategies able to ensure the success of students in reading, whatever the language of instruction. The research summarized here provides a step in this direction.

References

Bureau of the Census, (1992). Current population survey. Washington, DC: U.S. Department of Commerce.

Calderón, M. (1994, April). Cooperative learning as a powerful staff development tool for school renewal. Paper presented at the annual meetings of the American Educational Research Association.

Dianda, M. (1995, April). Effects of Success for All on the reading achievement of first graders in California bilingual programs. Paper presented at the annual meeting of the American Educational Research Association, San Francisco.

Durán, R.P. (1983). Hispanics' education and background. New York: College Entrance Examination Board.

Fleischman, H.L., & Hopstock, P.J. (1993). Descriptive study of services to limited English proficient students. Arlington, VA: Development Associates.

Garcia, E. (1994, April). The impact of linguistic and cultural diversity on America's schools: A need for new policy. Paper presented at the annual meeting of the American Educational Research Association, New Orleans.

Garcia, E.E. (1991). Bilingualism, second language acquisition, and the education of Chicano language minority students. In R.R. Valencia (Ed.), Chicano school failure and success: Research and policy agendas for the 1990's. New York: Falmer.

General Accounting Office (1994). Limited English proficiency: A growing and costly educational challenge facing many school districts. Washington, DC: U.S. General Accounting Office.

Hakuta, K., & Garcia, E. (1989). Bilingualism and education. American Psychologist, 44, 374-379.

Hertz-Lazarowitz, R. Ivory, G. & Calderón, M. (1993). The bilingual cooperative integrated reading and composition (BCIRC) project in the Ysleta Independent School District: Standardized test outcomes. Baltimore, MD: Johns Hopkins University, Center for Research on Effective Schooling for Disadvantaged Students.

Juel, C. (1988). Learning to read and write: A longitudinal study of 54 children from first through fourth grades. Journal of Educational Psychology, 80, 437-447.

Leighton, M.S., Hightower, A.M, Wrigley, P., Pechman, E.M. & McCollum, H. (1930) Improving education for language minority students: Promising practices in preschool development. Washington, DC: Policy Studies Associates.

Lloyd, D. N. (1978). Prediction of school failure from third-grade data. Educational and Psychological Measurement, 38, 1193-1200.

Mullis, I., Dossey, J., Foertsch, M., Jones, L., & Gentile, C. (1991). Trends in academic progress. Washington, DC: National Center for Education Statistics, U.S. Department of Education.

Ramirez, J.D. (1986). Comparing structural English immersion and bilingual education: First year results of a national study. American Journal of Education, 95, 122-148.

Slavin, R.E. (1987). Ability grouping and student achievement in elementary schools: A best- evidence synthesis. Review of Educational Research, 57, 347-350.

Slavin, R.E. & Madden, N.A., (1994, April). Lee Conmigo: Effects of Success for All in bilingual first grades. Paper presented at the annual meeting of the American Educational Reserach Association, New Orleans.

Slavin, R.E., Madden, N.A., Dolan, L.J., & Wasik, B.A. (1995, April). Success for All: A multi-site replicated experiment. Paper presented at the annual meeting of the American Educational Research Association, San Francisco.

Slavin, R.E., Madden, N.A., Dolan, L.J., Wasik, B.A., Ross, S., & Smith, L. (1994). "Whenever and wherever we choose...". The replication of Success for All. Phi Delta Kappan., 75 (8), 639-647.

Slavin, R.E., Madden, N.A., Karweit, N.L., Dolan, L., & Wasik, B.A. (1992). Success for All: A relentless approach to prevention and early intervention in elementary schools. Arlington, VA: Educational Research Service.

Slavin, R.E., & Yampolsky, R. (1991). Effects of Success for All on students with limited English proficiency: A three-year evaluation. Baltimore, MD: Johns Hopkins University, Center for Research on Effective Schooling for Disadvantaged Students.

Stevens, R.J., Madden, N.A., Slavin, R.E., & Farnish, A.M. (1987). Cooperative Integrated Reading and Composition: Two field experiments. Reading Research Quarterly, 22, 433-454.

Stevens, R.J., & Slavin, R.E. (1995). The effects of Cooperative Integrated Reading and Composition (CIRC) on academically handicapped and non-handicapped students achievement, atitudes, and metacognition in reading and writing. Elementary School Journal, 95 (3), 241-261.

Wasik, B.A., & Slavin, R.E. (1993). Preventing early reading failure with one-to-one tutoring: A review of five programs. Reading Research Quarterly, 28, 178-200.

Willig, A.C. (1985). A meta-analysis of selected studies on the effectiveness of bilingual education. Review of Educational Research, 55, 269-317.

Wong-Fillmore, L., & Valadez, C. (1986). Teaching bilingual learners. In M.C. Wittrock (Ed.), Handbook of Research on Teaching (3rd Ed.). New York: Macmillan.

Woodcock, R. W. (1984). Woodcock Language Proficiency Battery. Allen, TX: DLM.