Teaching English-Language Learners: What Does the Research Say?

AFT Policy Brief Number 14


The argument over how U.S. schools should educate non-English-speaking students is as old as the nation itself. The fact that many immigrant children were being taught in their native languages — including French, German, Dutch, and Swedish — was alarming to some 18th-century commentators, who feared a threat to political cohesion and the social order. To some extent, the practice of teaching children in languages other than English — and the backlash against this practice — have persisted throughout the intervening years.

Issues in current debate were framed by a 1974 Supreme Court case, Lau v. Nichols. The lawyers for a young Chinese boy named Kenny Lau argued that it was unfair for him to be taught only in English, a language he didn't understand. The Supreme Court agreed, ruling that it was the same thing as denying him an education — that is, a violation of his civil rights. The court did not prescribe any specific remedy, however. Schools could take a variety of approaches, from teaching English to students as quickly as possible to providing academic instruction in their native language(s) while helping them make a gradual transition into English.*

Data from across the country demonstrated that the majority of non-English-speaking students (identified variously as English-language learners [ELL] or students with limited-English proficiency [LEP]) — performed below grade level and suffered from alarmingly high dropout rates. These students needed all the help they could get. Unfortunately, there was very little hard evidence to guide schools in figuring out the best way to provide that help. In the absence of research, most schools developed programs for English-language learners in accord with local circumstances (availability of qualified personnel, funding constraints, number of students and their needs, etc.) and educators' best instincts. Most existing programs can be clustered into five broad categories (Hakuta, 2000), of which English as a second language and transitional bilingual education are by far the most common (August & Hakuta, 1997):

English as a second language (ESL)
Specified periods of instruction are aimed at developing students' English-language skills, with a primary focus on communication, grammar, and vocabulary. Students may be from the same or different language backgrounds. Academic content is addressed through mainstream classes. In some cases, native-language support (from a multilingual paraprofessional, for example) is provided.
Sheltered instruction/structured immersion
Students (often from several different language backgrounds) are clustered by English proficiency level(s) and provided with instruction in English that is modified so that subject matter is more comprehensible to students with limited vocabularies. Typically, there is little or no native-language support.
Transitional/early-exit bilingual education
Students receive academic instruction in their native language while making a gradual transition to English-language instruction. The goal of the program is to make the shift to English as quickly as possible (approximately two to four years).
Maintenance/late-exit bilingual education
Students receive academic instruction in their native language while making a gradual transition to English-language instruction. The goal of the program is to develop academic proficiency in both English and the native language, with some native language instruction maintained for an extended period (perhaps seven or more years).
Two-way bilingual education/dual-language immersion
About half of the students are native English speakers; half are English-language learners who speak the same native tongue. Both groups of students receive instruction in both languages, with the goal of helping all students develop academic proficiency in both languages. This program is often offered as a magnet school option.

Beyond the philosophical debate, discussed below, several factors influence the type of program(s) that a given school or district is likely to adopt. These include school demographics (type, number, and concentration of ELL students), student characteristics (age, academic background, and English-proficiency level), and available resources (ability to find qualified personnel, availability of classroom space), etc. (See appendices.)

Data and Debate

As the number of U.S. students with limited-English proficiency has grown — from about 1.25 million in 1979 to approximately 4.1 million today — so too have the arguments over the most appropriate methods for educating students. Although the positions of a few extremists may be grounded in xenophobic or ethnic separatist impulses, most of the conflicting viewpoints seem to arise from genuine concern over students' wellbeing. At the risk of oversimplifying the debate, these arguments fall into three basic camps.

First, the proponents of bilingual education (some academic instruction in the native language) argue that ELL students are harmed when schools sacrifice content knowledge on the altar of the earliest possible acquisition of English. The fact that students are taught to read in more than one language is seen as an important benefit that may prove valuable in later life. It is argued further that bilingual education programs may foster stronger connections between the school and students' home cultures and communities, thereby reducing student alienation and related problems.

On the other side, the critics of bilingual education believe that this approach has worked to trap students in culturally and linguistically isolated settings, thus impeding their ability to enter the American mainstream. These critics point out that, while families can provide children with grounding in their native language and culture, many students are totally dependent on schools to equip them to succeed in the English speaking world — thus, they call for immediate immersion, arguing the more time spent on English, the better.

A third camp, located somewhere in between the other two, is likely to support either ESL or transitional bilingual education programs, depending on circumstances. Here, the use of the student's first language is seen primarily as an instructional tool, not necessarily as a major instructional goal. The earliest possible acquisition of English proficiency remains the primary focus, whatever the instructional method.

As the various approaches to the instruction of ELL students have been implemented in classrooms across the country, all sides of the debate have worked to find and analyze student performance data that bolster their case. To date, no camp has been able to gather enough definitive evidence to resolve the dispute and gain consensus on the best way to accelerate achievement (which, for most ELL students, has been improving at a rate that is alarmingly glacial). Given the state of the research, this is understandable.

The need for strong research into the educational attainment of English-language learners poses a serious challenge. Assessing Evaluation Studies: The Case of Bilingual Education Strategies, a 1992 report from the National Research Council (NRC), helps articulate one central dilemma: "In the absence of a well-defined set of program objectives, any research effort to assess success' of programs will encounter problems and difficulties from the start" (Meyer & Fienberg, 1992). For some, the primary objective of educational programs for ELL students is the swift development of English language proficiency. Thus, measures such as annual tests of students' English reading, writing, and speaking skills and the rate at which students are moved into mainstream classes are essential. For others, the primary objective of ELL programs is academic success across the curriculum. Thus, an immediate gain in English proficiency, in and of itself, may be less significant than other measures of long-term success, such as high school graduation and college attendance rates. As a consequence, many of the competing studies are actually weighing apples against oranges.

As Improving Schooling for Language-Minority Children: A Research Agenda, a 1997 report from the NRC, points out, a second problem lies in the nature of assessment. If a test of students' academic knowledge is the goal, then translating the regular statewide or district wide assessments into another language can create serious technical problems and raise questions about validity (August & Hakuta, 1997). The assessment of students' knowledge of English is also problematic. Different tests may focus on different aspects of language ability and may rely on different definitions of proficiency. Thus, school districts may actually be able to vary the number of ELL students they identify just through their choice of a proficiency test (Ascher, 1990). As a consequence, which students are classified as English-language learners (or as Limited-English proficient) may differ significantly from study to study.

A third problem is the fact that definitions of "bilingual," "ESL," and "immersion" vary hugely from state to state and district to district — at times, even from classroom to classroom. Thus, a program that one school regards as "bilingual" may involve a limited amount of native-language instruction, while another school across town might label a program with a substantial amount of first-language support as "ESL" or "immersion." Implementation of these programs also tends to fluctuate widely. As the NRC report points out, these "approaches do not exist in isolation, coexist even within schools, and are often combined in various ways, depending on the availability of staff and resources" (August & Hakuta, 1997). In other words, it's difficult to determine the outcome in a research comparison when there is no real clarity about what is being compared.

It should come as no surprise that both of the expert panels assembled by the National Research Council in the 1990s to review the research on English-language learners found that most existing studies had serious flaws. Nevertheless, the two groups concluded that — all else being equal — some native-language instruction is better than none. How much native-language instruction is desirable, its duration, and the context of the program in which it is presented remain matters of serious dispute.

Arguably, the NRC's most important contribution to this field is the observation of its 1997 report that, whatever the label, each approach to helping ELL students can be implemented well or poorly. There is evidence to demonstrate the positive effects of quality programs characterized as "bilingual" and "immersion" and "ESL," just as there is ample evidence to show the negative effects of bad programs that are similarly labeled. Further, successful programs that are nominally very different may actually have more points of similarity than of difference.

The NRC's 1998 summary report provided this synopsis of the research literature on the achievement effects of educational programs for English-language learners:

  • The major national-level program evaluations suffer from design limitations; lack of documentation of study objectives, conceptual details, and procedures followed; poorly articulated goals; lack of fit between goals and research design; and elaborate statistical designs to overcome shortcomings in research designs.
  • In general, more has been learned from reviews of smaller-scale evaluations, although these, too, have suffered from methodological limitations.
  • It is difficult to synthesize the program evaluations of bilingual education because of the extreme politicization of the process…
  • The beneficial effects of native-language instruction are clearly evident in programs that are labeled "bilingual education," but they are also clear in some programs labeled "[ESL or] immersion"…
  • There is little value in conducting evaluations to determine which type of program is best. The key issue is not finding a program that works for all children and all localities, but rather finding a set of program components that works for the children in the community of interest, given that community's goals, demographics, and resources (August & Hakuta, 1998).

Appendix A: English-language Learners: Facts & Figures

  • In 1979, approximately 1.25 million students in U.S. schools were identified as English-language learners. By 1995, that number had almost doubled to 2.44 million students (National Center for Education Statistics, 1997a). In 2000, there were an estimated 4.1 million ELL students in U.S. schools (Macías, et al., 2000).
  • By far, the largest proportion of ELL students are native speakers of Spanish (73 percent), followed by speakers of Vietnamese (3.9 percent), Hmong (1.8 percent), Cantonese (1.7 percent), Cambodian (1.6 percent), Korean (1.6 percent), Laotian (1.3 percent), Navajo (1.3 percent), Tagalog (1.3 percent), Russian, French Creole, Arabic, Portuguese, Japanese, Armenian, Chinese, Mandarin, Farsi, Hindi, and Polish (unspecified). (Fleischman & Hopstock, 1993; August & Hakuta, 1998).
  • A disturbingly large percentage of ELL students receive low grades, score below their classmates on standardized reading and mathematics tests, and drop out of school. For example, the U.S. Department of Education's Prospects study reported that third-grade ELL students had a mean percentile score of 24.8 in reading and 35.2 in math on the Comprehensive Tests of Basic Skills (CTBS) test, as compared to a mean percentile score of 56.4 and 56.8, respectively, for all third-graders in public school (Moss & Puma, 1995; August & Hakuta, 1997).
  • Some but not all of this achievement gap can be explained by the fact that a disproportionate number of ELL students tend to be from disadvantaged socioeconomic backgrounds. For example, 77 percent of ELL students are eligible for free or reduced-price lunches, compared with 38 percent overall in the same schools. A large percentage of ELL students also attend schools with a high proportion (75 percent to 100 percent) of students in poverty, which is in itself a risk factor for low achievement. About 43 percent of ELL first-graders and 51 percent of ELL thirdgraders attend such schools, compared with about 13 percent of the overall student population (Moss & Puma, 1995; August & Hakuta, 1997).
  • Schools use various methods to identify ELL students, chiefly a home language survey, language assessment tool, previous student record, and teacher referral (National Center for Education Statistics, 1997b).
  • According to a National Center for Education Statistics survey, about 76 percent of ELL students receive ESL instruction, compared to approximately 40 percent of such students who receive bilingual education services. Bilingual programs are generally implemented in schools with the highest concentrations of ELL students (National Center for Education Statistics, 1997b).
  • Roughly 30 percent of all public schools with ELL students provide both ESL and bilingual education programs. About 71 percent of all ELL students attend these schools. Roughly 13 percent of all public schools provide neither ESL nor bilingual education programs. About 3 percent of all ELL students (more than 59,000) attend these schools (National Center for Education Statistics, 1997b).
  • While 42 percent of all public school teachers report that there is at least one ELL student in their classes, only 7 percent of these teachers have classes in which ELL students represent more than 50 percent of students (National Center for Education Statistics, 1997b).
  • Only 2.5 percent of teachers with ELL students have a degree in ESL or bilingual education. Only 30 percent of teachers with ELL students in their classes have received any training in teaching English-language learners. Teachers in schools with the highest concentrations of ELL students are the most likely to have received such training (National Center for Education Statistics, 1997b).

Appendix B: Choosing a Program Model**

There are several variables that influence a district's decision to develop a specific type of program for the instruction of English-language learners (such as transitional bilingual, maintenance bilingual, two-way bilingual, structured/sheltered immersion, or ESL program). These include:


Districts and schools may find themselves with many different types of students who are English-language learners. Some districts have a large, relatively stable population of ELL students who share the same language and cultural background. Other districts may have to respond to a sharp upsurge in ELL students from a particular language background, such as the students of Cuban, Haitian, Mexican, Russian, Somali, and Vietnamese heritage that a number of districts have had to accommodate in recent years. Still other districts may have a small number of students from a great number of different language backgrounds. For example, a few districts report having ELL students who speak more than 100 distinct languages scattered across the school system.

Student Characteristics

The circumstances of individual students can also influence the choice of program. Some students will enter school with no English at all. Others will know enough English to make themselves understood on the playground, but will still be at a significant disadvantage when forced to compete in English with classmates who are fully proficient in the language. Students' academic backgrounds and whether they are in primary or secondary school must also be considered. For example, some ELL students enter U.S. schools from countries where they have already learned to read and write in their first language, and where they have achieved a level of academic mastery that is on par with (or more advanced than) their U.S. peers. Others come from impoverished backgrounds in countries that offer little or no formal schooling.

Available Resources

The resources that schools have to work with and the ability to implement particular program choices will vary a great deal from district to district. Schools nationwide are reporting shortages of certified bilingual, ESL, and foreign language teachers, as well as a shortage of teachers who are both certified in a specific academic content field and fully proficient in more than one language. But other districts have a large, stable community group with a particular language background that makes it easier to recruit qualified bilingual teachers and paraprofessionals. Still other districts will have to scramble to find a native language speaker to work with a few students on a volunteer basis. In addition, districts with declining enrollments may find it easier to locate space for special magnet schools and ESL classrooms, while districts that are experiencing severe overcrowding may have limited options.

*In 1975, the predecessor of the U.S. Department of Education's Office for Civil Rights (OCR) released a detailed report, Task-Force Findings Specifying Remedies Available for Eliminating Past Educational Practices Ruled Unlawful under Lau v. Nichols, that went well beyond the Lau decision, recommending that elementary schools provide English-language learners with some form of bilingual education.

**Adapted from a September 1993 ERIC Digest, "ESL and Bilingual Program Models" (Rennie, 1993).


American Federation of Teachers (1990). Children with Limited English Proficiency, resolution adopted at the American Federation of Teachers Convention, July 1990, Boston, Mass.

Ascher, C. (1990). Assessing Bilingual Students for Placement and Instruction, ERIC Clearinghouse on Urban Education, Digest No. 65. New York, N.Y.: U.S. Department of Education.

August, D. & Hakuta, K., eds. (1997). Improving Schooling for Language-Minority Children: A Research Agenda. Committee on Developing a Research Agenda on the Education of Limited- English-Proficient and Bilingual Students; Board on Children, Youth, and Families; National Research Council. Washington, D.C.: National Academy Press.

August D. & Hakuta, K., eds. (1998). Educating Language-Minority Children. Committee on Developing a Research Agenda on the Education of Limited-English-Proficient and Bilingual Students; Board on Children, Youth, and Families; National Research Council. Washington, D.C.: National Academy Press.

Fleischman, H. & Hopstock, P. (1993). Descriptive Study of Services to Limited English Proficient Students, Volume 1. Summary of Findings and Conclusions. Arlington, Va.: Development Associates.

Greenberg, E., et al. (2001). English Literacy and Language Minorities in the United States. Washington, D.C.: National Center for Education Statistics (www.nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2001464).

Hakuta, K. (Jan. 27, 2000). Improving Schooling for Language Minority Students: Moving from What We Know, presentation to a Pew Network meeting.

Macías, R.F., et al. (2000). Summary Report of the Survey of the States’ Limited English Proficient Students and Available Educational Programs and Services, 1997-98. Washington, D.C.: National Clearinghouse for Bilingual Education (www.ncbe.gwu.edu/ncbepubs/seareports/97-98/).

Meyer, M. & Fienberg, E., eds. (1992). Assessing Evaluation Studies: The Case of Bilingual Education Strategies. Panel to Review Evaluation Studies of Bilingual Education, Committee on National Statistics, National Research Council. Washington, D.C.: National Academy Press.

Moss, M. & Puma, M. (1995). Prospects: The Congressionally Mandated Study of Educational Growth and Opportunity. Cambridge, Mass.: Abt Associates.

National Center for Education Statistics (1997a). The Condition of Education 1997. Washington, D.C.: U.S. Department of Education.

National Center for Education Statistics (1997b). 1993-94 Schools and Staffing Survey. A Profile of Policies and Practices for Limited English Proficient Students: Screening Methods, Program Support, and Teacher Training. Washington, D.C.: U.S. Department of Education.

Rennie, J. (September 1993). ESL and Bilingual Program Models, ERIC Clearinghouse on Languages and Linguistics. Washington, D.C.: U.S. Department of Education.

American Federation of Teachers. (2002). Teaching English-Language Learners: What Does the Research Say?. AFT Educational Issues Policy Brief Number 14.


For any reprint requests, please contact the author or publisher listed.

More by this author

Donate to Colorin Colorado

Add new comment

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.