• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

The effects of bilingual education on language, achievement, and self-efficacy of Hispanic students

Dissertation
Author: Lisa A. Lockwood Hewitt
Abstract:
Much controversy exists surrounding the education of Hispanic English Language Learners (ELLs). This large and growing group presents significant challenges to educators. Foremost among these challenges is the question of whether bilingual or English-only education is most appropriate for enhancing ELLs' language proficiency and achievement. Despite decades of controversy and research in the field of bilingual education, the debate is ongoing. Additionally, Hispanic ELLs are profoundly affected by other cultural and educational factors. One potentially important factor with limited research involves the academic self-efficacy of ELLs. This study examined the relationship of bilingual and English-only education to Hispanic ELLs' language proficiency, academic achievement, and academic self-efficacy. Participants were eighth-grade Hispanic students from a large southwestern school district. Data were collected from school district records, and a self-efficacy questionnaire was administered to a subsample of students. Analyses included t-tests, ANOVA, ANCOVA, and regression procedures to measure relationships between ELL students who received bilingual education (the Bilingual group) or English-only education (the Mainstream ELL group), as well as a Comparison group of non-ELL Hispanic students. It was hypothesized that the Bilingual group would demonstrate advantages over the other two groups in English proficiency, academic achievement, and academic self-efficacy. Analyses revealed few significant group differences. The Bilingual group did not attain significantly higher English proficiency than the Mainstream ELL group by third grade. The Bilingual group did not demonstrate significantly higher achievement scores than the Mainstream ELL group, but their scores were significantly higher than the Comparison group. There were no significant differences between groups on academic self-efficacy. Regression analyses indicated that the length of time spent in bilingual education did not predict students' language proficiency, achievement, or academic self-efficacy. A final analysis indicated that academic self-efficacy and third grade English proficiency scores were significant predictors of eighth-grade achievement. Conclusions indicate modest benefits for ELL students attending bilingual education, but more advantages may have been evident had more years of bilingual education been provided to students. Clearly, increased attention to academic self-efficacy and English proficiency may be appropriate regardless of the type of educational placement. Further studies should examine other factors affecting the quality of education provided to ELL students.

viii TABLE OF CONTENTS

Page ABSTRACT.............................................................................................................. iii DEDICATION.......................................................................................................... v ACKNOWLEDGEMENTS...................................................................................... vi TABLE OF CONTENTS.......................................................................................... viii LIST OF TABLES.................................................................................................... x CHAPTER I INTRODUCTION................................................................................ 1 Significance of the Problem........................................................... 3 Purpose of Research....................................................................... 4

II REVIEW OF RELEVANT LITERATURE........................................ 7 The Education of Hispanic ELL Students...................................... 7 Self-Efficacy................................................................................... 26 Statement of the Problem............................................................... 35

III METHOD............................................................................................. 37

Participants..................................................................................... 37 Measures......................................................................................... 42 Procedure........................................................................................ 46

IV RESULTS............................................................................................. 49

Treatment of Missing Data............................................................. 51 Power Analysis............................................................................... 51 Initial Group Comparability........................................................... 52 Research Question 1....................................................................... 53 Research Question 2....................................................................... 55 Research Question 3....................................................................... 58 Research Question 4....................................................................... 58

ix CHAPTER Page

Additional Analysis........................................................................ 59

V DISCUSSION AND CONCLUSIONS................................................ 61 Discussion of Results..................................................................... 61 Limitations..................................................................................... 64 Directions for Future Research...................................................... 67

REFERENCES.......................................................................................................... 68 VITA......................................................................................................................... 79

x LIST OF TABLES Page Table 1 Summary of Descriptive Statistics of Measures............................... 50

Table 2 ANCOVA for Achievement Differences Between Overall Sample and Subsample, Gender as Covariate................................... 53

Table 3 ANCOVA for Effect of Bilingual Education on RPTE Score, OLPT as Covariate............................................................................ 54

Table 4 ANOVA Comparing Bilingual ELL, Mainstream ELL, and Comparison Groups on Eighth Grade Achievement......................... 56

Table 5 Tukey Test of Differences Between Groups on Academic Achievement...................................................................................... 57

Table 6 ANCOVA for Effect of Bilingual Education on Achievement, OLPT as Covariate............................................................................ 57

Table 7 Regression Analyses of Years of Bilingual Education Predicting RPTE, Achievement, and Self-Efficacy.......................... 59

Table 8 Hierarchical Regression with RPTE and Academic Self- Efficacy Scores Predicting Academic Achievement......................... 60

1 CHAPTER I INTRODUCTION

A significant and ever-growing number of students in the United States are of Hispanic origin (National Center for Education Statistics [NCES], 2005). These students and their families contribute a wealth of talents, potential, and cultural diversity. The term “Hispanic American” may refer to those with origins in any Central or South American country, in effect, anyone with ancestry from Spain (“Hispanic,” 2006). While Hispanic students constitute a valuable asset to the cultural and academic fabric of the U.S., this group also brings its share of challenges to the field of education. For example, the dropout rate among Hispanic American students is higher than among any other ethnic group, especially among immigrants (NCES, 2002). Immigrant and first- generation Hispanic children may begin school speaking Spanish only, with little or no working English knowledge. These individuals, as well as any other non-English- Speaking students, are referred to as English Language Learners (ELLs) by current research. Older literature, as well as many school districts, has referred to these students as Limited English Proficient (LEP) or English as a Second Language (ESL) students. For consistency, the ELL designation will be used throughout this study. To remedy the achievement discrepancy between ELL students and English-speakers, legislation now requires that ELLs be included in any annual testing used to measure the progress of ____________________ This dissertation follows the style of School Psychology Review.

2 schools and students (Porter, 2000a). The challenge facing educators seems vast: How can the educational system take into account extensive cultural and linguistic factors in order to best serve its population of Hispanic American students? The degree to which educators are able to effectively serve the ELL population, and indeed all Hispanic students, has significant implications for the academic health and future success of this group, as well as for the entire country. One of the major issues debated in the education of ELL Hispanic students (as well as other ELL students) is the degree to which these students should be taught in their native language. Throughout the history of the United States, bilingual education has received varying levels of support: at times, educators and lawmakers have allowed or supported ELL students to be taught in their native language in public schools, while at other times bilingual education has been discouraged or even outlawed (Baca & Cervates, 2004). Currently, bilingual education programs are widely used in many parts of the country, while some states have enacted laws requiring that ELL students be fully immersed in English-only education after a year of limited support (Rolstad, Mahoney, & Glass, 2005). Empirical research examining bilingual education shows a similar controversy. Some research supports the efficacy of bilingual education for English acquisition and achievement (Slavin & Cheung, 2005; Thomas & Collier, 2002; Willig, 1985), while other studies support English-only education as superior to bilingual methods (Porter, 2000b; Rossell & Baker, 1996). Aside from the type of educational program attended by Hispanic ELL students, many other factors contribute to this group’s academic success and must be considered.

3 One of these influential variables is self-efficacy. Self-efficacy is defined as the belief in one’s abilities to perform a certain task (Bandura, 1997). There are many different domains of self-efficacy, and many may strongly impact school success; however, one category very much of interest to educators is academic self-efficacy. Academic self- efficacy has been shown to impact student outcomes in a variety of subjects (Pajares & Valiante, 2006; Shell, Colvin, & Bruning, 1995; Zimmerman, Bandura, & Martinez- Pons, 1992). To date, little research is available on the academic self-efficacy of Hispanic ELL students. There is some research to indicate that children of non-English- speaking parents may develop self-efficacy skills when they act as “language brokers” to help their parents communicate with English-speakers (Buriel, Perez, DeMent, Chavez, & Moran, 1998; Tse, 1995). Significance of the Problem The number of school-age children who speak a language other than English doubled between 1972 and 2003, with the majority of this growth attributed to an increase in the number of Spanish-speaking immigrants (NCES, 2005). Meeting these students’ educational needs is imperative, and the task is made more urgent because of the many risk factors facing these students: their minority status and their lack of English knowledge not the least among them. Foremost among the educational concerns surrounding Spanish-speaking ELLs’ language proficiency and achievement is the question of which type of education is best: bilingual education or English-only immersion. The debate between proponents of each type of education continues to rage, with evidence supporting and condemning both sides of the issue (e.g. Ochoa, 2005;

4 Rossell & Baker, 1996). In another area of study, students’ academic-self-efficacy has shown to be a powerful factor in their success in academic tasks and courses. Little research is available investigating the academic self-efficacy as it relates to Hispanic ELL students, the interaction of students’ language proficiency and academic self- efficacy, or of the effects of different language learning methods on self-efficacy. Purpose of Research This study, then, will examine the relationship of bilingual or English-only education to the English language proficiency, academic achievement, and academic self-efficacy of Hispanic ELL students. As mentioned previously, the controversy over which type of education is best for ELL students is a vast, longstanding one, and the results of this study are unlikely to settle the question. They may, however, provide one more piece of evidence to support one type of education over another for fostering language proficiency, achievement, and/or self-efficacy. Also, this study will provide information about the potential relationship between educational programming for Hispanic ELLs and their academic-self-efficacy. If a relationship exists between the bilingual education or English immersion and students’ sense of self-efficacy about academic abilities, it could provide useful information for improving student outcomes. Such information may also be useful in untangling the complex bilingual-immersion debate by highlighting another area of functioning that these types of education can affect. Specific research questions to be addressed include:

5 Research Question 1 Are language proficiency scores higher for Hispanic ELL children who attended bilingual education classes or those who attended English-only classes? It is hypothesized that Hispanic ELL students who attended bilingual education will achieve higher language proficiency scores in English than those ELLs attending English-only classes by the time students are in the eighth grade, after controlling for English proficiency levels at school entry. Research Question 2 Are achievement scores higher for ELL children who attended bilingual education classes or those who attended English-only classes? How does the achievement of ELL children in both these programs compare to that of Hispanic non- ELL children? It is hypothesized that Hispanic ELL students who attended at least two years of bilingual education classes will attain higher achievement scores than those Hispanic ELL students who were placed in English-only classes, by the time they reach eighth grade. It is further hypothesized that Hispanic ELL students with bilingual education experience will reach parity on achievement variables by eighth grade, as compared with their Hispanic non-ELL peers. Research Question 3 How does the attendance of bilingual or English-only classes affect the academic self-efficacy of Hispanic ELL students? How does the academic self-efficacy of Hispanic ELL students compare to that of Hispanic non-ELL students? It is hypothesized that those Hispanic ELL students who attended bilingual education classes

6 will exhibit higher academic self-efficacy than those Hispanic ELL students who did not attend bilingual education classes. It is further hypothesized that, due to their bilingual abilities, the Hispanic ELL students will exhibit higher academic self-efficacy than their Hispanic non-ELL peers. Research Question 4 How does the length of attendance of bilingual education classes predict Hispanic ELL students’ language proficiency, academic achievement, and academic self-efficacy? It is hypothesized that the levels of achievement, language proficiency, and self-efficacy attained by Hispanic students will be directly predicted by the amount of time spent by students in bilingual education.

7 CHAPTER II REVIEW OF RELEVANT LITERATURE

The Education of Hispanic ELL Students Hispanic ELL Students Hispanics comprise the largest ethnic minority group in the country. The Hispanic population of the U.S. nearly doubled between 1990 and 2004 and currently constitutes about 14 percent of the national population. Of the foreign-born population of the U.S., over half (53%) are from Latin American countries, and about 10 million emigrated from Mexico alone. (“Hispanic heritage,” 2005). While the majority of these immigrants are Mexican nationals, many of these Hispanic newcomers come from other countries such as Cuba and the Dominican Republic. Additionally, many Spanish- speakers come from the U.S. territory of Puerto Rico and are not technically immigrants at all. Since many new immigrants are Mexicans who cross the border and settle into the southern United States, Hispanic immigrant populations are greatest in border states such as Texas, Arizona, and California. However, both Mexican immigrants and those from other countries and territories have settled in diverse parts of the country. For example, New York City is home to a large population of Puerto Rican origin, and the Cuban population in Florida, particularly in Miami, is very significant. Additionally, many northern and midwestern states have recently experienced increased populations of Hispanic immigrants, attracted by the availability of unskilled labor positions. For example, Delaware, Nebraska, Indiana, Minnesota, and Idaho have all experienced

8 increases of over 150% in their foreign-born populations since 1990 (Migration Policy Institute, 2006). In short, Spanish-speakers comprise a significant population in various parts of the country, especially in certain states. Therefore, the education of ELLs, particularly those from a Hispanic background, is of great interest to educators across the nation. Outcomes for Hispanic Youth In addition to being a large and growing group, Hispanic students consistently achieve below-average scores on language and achievement measures, failing to match “basic” achievement levels as defined by the National Assessment of Educational Progress (NAEP; Grigg, Daane, Jin, & Campbell, 2003). Also compelling are dropout rates for immigrant Hispanic students, as compared to other groups. More than forty percent (44.2) of immigrant Hispanics between the ages of 16 and 24 had dropped out of high school, according to data from 2000 1 (NCES, 2002). In contrast, Hispanic students born in the United States had dropout rates between 14 and 16 percent, depending on whether they were of the first or second generation to be born in this country (NCES). These statistics clearly illustrate that immigrant Hispanic students are far less likely to complete school than those born in the United States, even those comprising the first generation. This is almost certainly due to a variety of factors: first- and second- generation Hispanic children are different from their non-immigrant peers on many important variables. These include differences in socioeconomic status, parental education and family involvement in education (Crosnoe, 2006). One notable and

1 In data provided by the NCES, the term “Hispanic immigrant” is used to include immigrants from Spain or Spanish-speaking Latin American countries, as well as Puerto Ricans who moved to the U.S. mainland.

9 obvious difference between immigrant and first-generation Hispanics is their level of exposure to English, which in turn affects status as an ELL or an English-speaking student. As the immigrant/ELL group has a much higher dropout rate than the groups of Hispanic students who have had more opportunity to learn English, it is important to determine the most effective means of education for these students. Since Hispanic language-minority students comprise such a significant portion of the overall student population in the U.S., the success of this group is of vital importance not only in itself, but for the outcomes of children in the U.S. as a whole. The question is obvious: Can the educational systems of the country argue that they have left no child behind, if such a substantial group of students continues to experience such low educational outcomes? Historical Overview of Bilingual Education Directly related to Hispanic ELL students is the topic of bilingual education. Bilingual education on some level has existed in the United States for hundreds of years, ever since immigrants to the brand-new country had to figure out how to educate their non-English-speaking children (Baca & Cervantes, 2004). From the late 18 th century and throughout the 19 th century, the state and federal governments did not discourage bilingual education. The federal government did not address the matter, and the policy of the states varied from specific allowance of bilingual education, to simply not prohibiting it. Things changed during the early 20 th century, when World Wars I and II sparked sheer xenophobia among Americans. Bilingual education, along with other practices supporting foreign cultures, was prohibited. New immigrants were under

10 pressure to conform quickly to the culture of the United States, or risk being seen as a subversive or traitor (Baca & Cervantes). The practice of native language instruction in schools, which had enjoyed such a lengthy period of support, was in danger of dying out completely. In the 1960s, however, policy toward bilingual education began to change as the U.S. became keenly conscious of the importance of civil rights. The Bilingual Education Act, passed in 1968, granted funding to schools to encourage the implementation of native-language instruction in the classroom (Baca & Cervantes, 2004). In 1974, the Supreme Court ruled in the landmark Lau v. Nichols case that language should not be a barrier between a child and his or her public education, and that schools should take steps to overcome the obstacle of language in educating their students (Jacob & Hartshorne, 2003). This era of support for bilingual education lasted through the 1970s and 1980s, but then once again the tide began to change. As early as 1981, lobbyists began demanding that English be made the official language of the United States (Baca & Cervantes). In the educational arena, support for bilingual education has waned dramatically since the 1990s, in favor of English-only instruction (Porter, 2000b) In schools, native-language instruction is no longer the trend, and English-only education has become more popular (Slavin & Cheung, 2005). For example, California’s Proposition 227 prohibits bilingual education for ELL students without a special waiver. Instead, ELL students are given one year of structured English immersion before being placed in regular English classes (Valdés, 2004). Following the

11 trend, Arizona’s Proposition 203, and Massachusetts’ Question 2 are among other states’ policies limiting or prohibit the bilingual education of ELL students, in favor of English- only education (Rolstad et al, 2005). In federal policy, too, bilingual education has lost support in recent years. The passage of the No Child Left Behind act also brought with it the repealing of the 1968 Bilingual Education Act. It was replaced with the English Language Acquisition Act, a term reflective of the current emphasis on English learning, without support of native language instruction (Crawford, 2002). Theories Supporting Native Language Instruction It is apparent, then, that the policy favoring English-only or bilingual education has changed dramatically over the past few decades, from one extreme to the other, and now back again. But what is the basis for these educational decisions and policies? Regrettably, for such a political issue, decisions on which type of education to enact and support are often made on the basis of ideologies, without serious consideration of empirical evidence (Hakuta, Butler, & Witt, 2000). For example, the 1968 Bilingual Education Act came in an era of civil rights, when concerned leaders noticed that ELL students, taught only in English, were performing more poorly academically than their native English-speaking peers. The Bilingual Education Act sought to remedy the difference by allowing these students to be taught in their native language (Baker & Rossell, 1993). More recent changes toward English-only education also have a largely political basis (Valdés, 2004). However strong the role of politics has proven in this debate, researchers have not been silent. About a decade after the Bilingual Education Act, Cummins (1979)

12 began developing and publishing his theory of second language acquisition to explain why some ELL children, schooled in their native language while learning English, still lagged behind in their linguistic skills. His facilitation theory, developed in a number of papers in the late 1970s and early 1980s (e.g. Cummins, 1979, 1981) postulated that, to learn a second language (known as L2) well, a student must build a strong foundation in his or her first language (L1). Cummins’ (1979) theory explained that some students placed in bilingual education still may flounder if not given enough L1 instruction—that the failing was not in the bilingual education itself, but in insufficient time given for language development. Similar to the facilitation theory is the threshold hypothesis. This idea asserts that second language learners must reach a certain threshold of language ability in their L1 before they can achieve fluency in L2 (Cummins, 1979; Toukomaa & Skutnabb-Kangas, 1977). In a similar vein, the contrastive analysis framework theory suggests that, as much as L1 is similar to L2, instruction in L1 will also facilitate better knowledge of L2 (Fashola, Drum, Mayer, & Kang, 1996). In contrast to Cummins’ facilitation theory is the “time-on-task” theory: the idea that the more time English learners spend learning English, the better they will learn it (Baker, 1992, Baker & Rossell, 1993). This line of thought is the major argument for proponents of English-only education: that time spent learning a student’s L1 is wasted, since it is not dedicated specifically to the development of the L2. As mentioned previously, however, most theories of second language acquisition cite the support of the L2 by further development of the L1. Along this vein, De Houwer (2005) states that

13 there is as yet no empirical basis for the claim that, as a group, bilingual children develop their languages more slowly than monolingual children” (p. 41). Other theorists’ views on second language acquisition further bolster the case for bilingual education. For example, Krashen (2000), postulated that receiving comprehensible input in a second language is crucial to developing proficiency in that language. In other words, mere exposure to a language is insufficient; a learner must be presented with new language content in a way that he or she can interpret and understand the message. This tenet supports the use of bilingual education, since it is more likely that students will encounter slower, smaller doses of English in such programs, rather than being flooded with large amounts of English while their proficiency is still very limited. Also, building a strong native language base helps make new input in a second language more comprehensible (Krashen, 2000). Review of Bilingual Education Research In research as well as ideology and school policy, the superiority of bilingual or English-only education has been hotly contested for decades (Croft & Franco, 1983; Willig, 1985; Slavin & Cheung, 2005; Ochoa, 2005), and hundreds of individual studies of varying scope and integrity have examined these educational programs (Rossell & Baker, 1996). Some of the largest and most notable studies will be discussed here, as well as a number of the most notable literature reviews and meta-analyses on this topic. The Ramirez Report, 1991. In 1991 the findings were published from a federally-commissioned study examining the effectiveness of various forms of bilingual education. The study followed about 2000 elementary school ELL students who were

14 placed in either structur ed English immersion, with no native-language instruction, early-exit transitional bilingual education (TBE), with native language instruction phased out over two to three years, or late-exit TBE, in which students receive native language instruction through the sixth grade. Results of the study indicated that ELL students placed in structured English immersion did not achieve English proficiency faster than their peers in bilingual education; in contrast, by the sixth grade some of their achievement scores lagged behind their bilingually-instructed peers. On a related theme, the study’s results did not indicate that students in bilingual education took longer to learn English than their peers in structured English immersion. The data suggested that it took the students in the sample about six years to achieve English proficiency, regardless of the program attended (Ramirez, Yuen, Ramey, Pasta, & Billings, 1991, 1992). The release of the much-anticipated Ramirez report findings incited a flurry of responses from other researchers. Baker (1992), though usually considered an opponent of bilingual education, conceded that the results suggested that ELLs in the early grades learn better from bilingual education rather than English-only immersion. Some recommended that only conservative conclusions be made from the results of the study, since many variables were different between groups. For example, though students in late-exit programs performed better than those in early-exit programs, definitions of these programs varied across districts. Moreover, the late-exit programs appeared to garner more parental support, which may have affected overall outcomes (Cazden, 1992).

15 Reports by Thomas and Collier. In the past ten years, Thomas and Collier (1997, 2002) have published two large-scale longitudinal reports on the effectiveness of bilingual education. The first study analyzed over 700,000 student records across several school districts, representing students in bilingual and English as a Second Language (ESL) programs. This large sample size was achieved by aggregating students from a number of sequential years into a “mega-cohort”, for which extant data was examined through the twelfth grade. Findings from the report indicated that students who receive the most native-language instruction, such as in maintenance and dual- language programs, showed the greatest achievement gains, especially toward the end of their secondary education. Students receiving only ESL support made much smaller progress (Thomas & Collier, 1997). Despite this study’s large sample size and the level of anticipation it received in the research community, it demonstrated significant weaknesses. First, and most seriously, Thomas and Collier did not adequately outline the statistical procedures used in their analyses. While there is some discussion about the relative merits and pitfalls of the ANCOVA procedure and a short description of a complicated matched-groups design employed by the researchers to control for differences, actual statistical procedures are not described. Tables and graphs are presented to illustrate student averages and trajectories, but no specific statistical or procedural data is provided as the basis for these figures. Even participant characteristics are ignored. For example, nothing is reported on the size of the cohorts in the sample or their demographic characteristics. The sample is vaguely referred to as being primarily Spanish-speaking,

16 with no mor e specific information given. Second, students in the study were not compared to other comparable students in the area; rather, their achievement gains were compared to norms on national achievement measures. This is not an adequate comparison, since by definition ELL students score below the national average on language measures (Rossell, 1998). The second large-scale report by Thomas and Collier examined groups of students in various locations across the U.S., including student data utilized in the earlier study (2002). As in the previous study, samples were compiled across multiple years into large, artificial cohorts, in order to increase sample sizes (1997). Also similar to their earlier research, the 2002 report found positive achievement effects for students in long- term bilingual education programs, such as maintenance and 50-50 dual-language programs. Some improvements were made between the two reports. For example, in this study, Thomas and Collier reported their statistical procedures in far more detail than they had done previously. Some drawbacks are still apparent, however: No pretest measures were administered, and groups of ELL students were again compared to national norms instead of a local comparison group. Also, in the final summary of the outcomes, achievement results for various educational programs are outlined, but these results are not consistent across groups and cannot be compared. For example, the group of students in 50-50 one-way programs (all ELL students, taught half in English, half in Spanish) are described as reaching the 62 nd normal curve equivalent for achievement after four years of bilingual education. In contrast, the achievement for students in 50-

Full document contains 90 pages
Abstract: Much controversy exists surrounding the education of Hispanic English Language Learners (ELLs). This large and growing group presents significant challenges to educators. Foremost among these challenges is the question of whether bilingual or English-only education is most appropriate for enhancing ELLs' language proficiency and achievement. Despite decades of controversy and research in the field of bilingual education, the debate is ongoing. Additionally, Hispanic ELLs are profoundly affected by other cultural and educational factors. One potentially important factor with limited research involves the academic self-efficacy of ELLs. This study examined the relationship of bilingual and English-only education to Hispanic ELLs' language proficiency, academic achievement, and academic self-efficacy. Participants were eighth-grade Hispanic students from a large southwestern school district. Data were collected from school district records, and a self-efficacy questionnaire was administered to a subsample of students. Analyses included t-tests, ANOVA, ANCOVA, and regression procedures to measure relationships between ELL students who received bilingual education (the Bilingual group) or English-only education (the Mainstream ELL group), as well as a Comparison group of non-ELL Hispanic students. It was hypothesized that the Bilingual group would demonstrate advantages over the other two groups in English proficiency, academic achievement, and academic self-efficacy. Analyses revealed few significant group differences. The Bilingual group did not attain significantly higher English proficiency than the Mainstream ELL group by third grade. The Bilingual group did not demonstrate significantly higher achievement scores than the Mainstream ELL group, but their scores were significantly higher than the Comparison group. There were no significant differences between groups on academic self-efficacy. Regression analyses indicated that the length of time spent in bilingual education did not predict students' language proficiency, achievement, or academic self-efficacy. A final analysis indicated that academic self-efficacy and third grade English proficiency scores were significant predictors of eighth-grade achievement. Conclusions indicate modest benefits for ELL students attending bilingual education, but more advantages may have been evident had more years of bilingual education been provided to students. Clearly, increased attention to academic self-efficacy and English proficiency may be appropriate regardless of the type of educational placement. Further studies should examine other factors affecting the quality of education provided to ELL students.