Testimony of Kenji Hakuta

(with the special assistance from Michele Bousquet Gutierrez)

to the United States Commission on Civil Rights

The Education of Language Minority Students

April 13, 2001

This testimony is organized around the three prongs of Castañeda v. Pickard (1981), which has been widely influential in policy and civil rights enforcement in defining the appropriateness of programs serving English Language Learners (ELLs). The standards are (1) that the program is based on sound educational theory, (2) that it is adequately implemented, and (3) that it is evaluated for its effectiveness in meeting the needs of the students.

Standard 1: Is the program based on an educational theory recognized as sound by at least some experts in the field?

Research findings on the effectiveness of bilingual education are far from definitive. However, the best research (i.e., research that investigates programs that are representative of their "bilingual" label and that are properly implemented) suggests that bilingual education is a successful model and does produce measurably better outcomes in academic achievement. This finding has been supported by two independent committees of the National Academy of Sciences, one of which I chaired. In fact, it is often the studies with the weaker research design that have reported negative findings on the success of bilingual education. Several meta-analyses that combine findings across independent studies have consistently supported bilingual education over English-only alternatives (Greene, 1998 and Willig, 1985).

In addition to the evaluation findings just noted, a bilingual education model is a necessary alternative to an English-only model based on educational theories found in the literature about transfer of knowledge, contexualized language, and affective and cognitive variables. For example, cognitive theory supports the notion that there is one common underlying proficiency for both a student’s languages and that skills learned in a student’s first language can be transferred to and accessible in the student’s second language. By teaching a child in a language she or he already understands, access to school subjects is accelerated and enhanced, and there is an immediate context for the content being taught. We further can assume, based on motivation theory, that when language minority students’ native language is valued and utilized in the curriculum, they are more likely to have increased self-esteem, less anxiety, and greater self-efficacy of their performance in educational settings. Finally, while there has been debate on the effectiveness of bilingualism to improve a learner’s cognitive skills, the most conservative conclusion would be that even if additive bilingualism does not contribute to increased metalinguistic awareness and cognitive flexibility, then it certainly does not detract from a person’s innate capabilities in these areas.

Research also suggests that success in non-bilingual education settings most likely occurs when children come from homes in which there is a high degree of education, where second language only programs are taught by teachers who have the skills to communicate with the child in his or her first language, and where there is a school-wide climate that supports high degrees of learning. However, absent these conditions—as is usually the case, the research strongly suggests that children are best taught at least to some degree in their first language until they have acquired proficiency in English. Because of this, native language instruction is one method of teaching that educators must be free to use in constructing effective programs for ELLs. Depriving educators of this method summarily removes one of the many useful tools that can be used to improve learning outcomes for these students.

There is no "formula" that can be used to determine how long it takes for ELL students to learn English. A myriad of factors contribute to ELL students’ rate of English acquisition, including whether they are proficient in their first language upon their arrival to the U.S. and at what age they enter U.S. schools. In a recent paper I co-authored with Yuko Goto Butler and Daria Witt, I outlined a study that showed the expected time it would take for these students to learn English is between 4 to 7 years, and that this time period varies considerably by the socioeconomic status of the students as well as the aforementioned variables. The study further showed that this estimate would hold both in districts that have bilingual education as well as those that use English-only and have intensive English as a Second Language instruction.

Basic research has corroborated my findings, showing that ELL students are learning English at a rapid and natural rate of development ("at the speed limit"), regardless of relative amounts of exposure to English vs. native language in school. It appears then that a simplistic, "time on task" theory fails to predict English language development, just as it fails to predict patterns in the acquisition of a first language. It also appears unlikely that any further intensifying of instruction in English would cause the rate of acquisition to become much faster.

Summary for Standard 1: In order to be compliant with Standard 1 of Castañeda, programs for ELLs must include at least some amount of native language support, and ELL students should receive some form of special instruction and accommodations for a period of at least 4 to 7 years.


Standard 2: Are the school district’s programs and practices, including resources and personnel, reasonably calculated to implement this theory effectively?

States need to think more broadly about ELL instruction, rather than just in terms of program type with respect to the language of instruction. What is critical is finding and communicating a set of program components that work for the children in a given community of interest, within the context of the goals, demographics, and resources of that community. This set of components will (and should) vary depending on factors that differ not only across but within immigrant groups, such as students’ first language, SES, previous academic experience, community and parental socio-linguistic climate, learner styles, and goals for proficiency (additive v. subtractive). Teacher availability and qualifications also play a vital role (National Research Council, 1997). A set of generally agreed upon "best practices" that can and should be found across program types to encourage the success of language minority students would include the following: some use of native language and culture in the instruction of language minority students, a balanced curriculum that incorporates both basic and higher-order skills, explicit skills instruction, opportunities for student-directed activities, use of instructional strategies that enhance understanding, opportunities for practice, systematic student assessment, staff development, and home and parent involvement. This important literature has been reviewed and supported by the National Research Council (1997).

Currently in typical programs for ELL students—regardless of program type—high levels of academic learning are not promoted. It stands to reason that if ELL students are going to have a chance at an equal education, they need to be included in standards-based reform movements in the various content areas and in academic measures of competitiveness. Clear academic standards for ELL students must be in place, confirming the need to set the same expectations for this population as for mainstream students. Emphasis should be placed on the integration of theory, standards, instruction, and assessment.

When setting these standards for ELL students, the focus should be on what ultimately matters most—their long-term performance and success. It is unreasonable to expect ELLs to perform comparably to their native English-speaking peers in their initial years of schooling (hence the need for standards specific to ELLs) and holding them to this expectation too early in their educational careers can be detrimental to their academic progress, not to mention their self-esteem. The problem enters when students are not pushed to go beyond this stage over time, are presumed to be at an elementary level, or are misdiagnosed as having educational disabilities by teachers unfamiliar with the needs of ELLs.

Currently, there is a dearth of appropriate materials for ELLs. Generally speaking what is needed are ELL textbooks and other instructional materials that are aligned to states’ ELD and content standards and that incorporate both basic and higher-order skills. Furthermore, it is questionable whether the limited ELL material that does exist is distributed equitably. Approximately 75% of all ELL students’ families settle in high poverty, urban areas where schools are "disproportionately poor, non-white, under-funded, and underachieving" (Crawford, 1997). The fact that lesser resources are available in many of the classrooms where ELLs can be found may be responsible for the fact that most ELL teachers do not use materials designed for ELLs. In fact in a nationwide sample, 64% of ELL teachers used the same reading materials for their English proficient students and their ELLs, despite the great disparities between these two groups (Moss & Puma, 1995).

The success of ELLs does not hinge solely on the resources and environment of the classroom but also on those of their school in general. Therefore, criteria for effective school environments for this population are important to identify as well. The following are recommended school attributes: a supportive school-wide climate, school leadership, a customized learning environment, articulation and coordination between and among schools in the district, school-wide coherence, rigorous standards for teaching and learning, assessment and accountability, continuous evaluation, and research of program effectiveness.

The academic success of all students rests in large measure in the hands of their instructors, and research suggests that this success is largely dependent upon teachers’ professional preparation and certification (Darling-Hammond, 2000; The Report of the Proposition 227 Task Force, 1999). Approximately 42% of U.S. teachers have at least one ELL in their classrooms (Moss & Puma, 1995). Proper training is particularly critical for those who work with ELLs, as members of the ELL population are not only vastly different from the mainstream student population, but there is considerable variability even within the population of ELL students. This heterogeneous group of students comes to U.S. classrooms with varying degrees of proficiencies in both their native language and English, differing amounts of academic content knowledge, and from varying socioeconomic and political circumstances, all of which affect learning readiness. Instructors of these students need additional teaching skills and theoretical knowledge beyond that which is taught to mainstream teachers in order to effectively instruct this population.

Despite the availability of programs like California’s Crosscultural Language and Academic Development (CLAD) & Bilingual CLAD (BCLAD) certificate programs that train teachers in 1) second language acquisition, 2) bilingual, ELD, and content instruction theories and methodology, 3) culture and diversity, and 4) (for B-CLAD only) proficiency in a target language, knowledge of the target culture, and methodology for instruction in the target language, many instructors of ELLs have not completed this training. In California alone, for example, approximately 101,000 public school teachers teach ELLs using English Language Development (ELD), Specially Designed Academic Instruction in English (SDAIE), or some combination of the two, yet only about half are certified to do so by the California Commission on Teacher Credentialing. Additionally, more than 16,000 teachers report that they use ELLs’ native language to offer content instruction bilingually, but less than 11,000 have CCTC authorization to do this (California Department of Education, Educational Demographics Unit, Language Census, 1999, Number of staff members providing primary language instruction to English learner students in California public schools, by language of instruction, 1998-99). Note that these numbers already reveal that a large portion of ELL teachers—in the state responsible for the education of the most ELLs in this country—are not trained to work with this population. Further magnifying the issue, however, is that these data do not account for teacher quality and availability in the other 49 states and the data only include California’s teachers who utilize ELD, SDAIE, and/or bilingual education approaches when teaching—not all of those teachers throughout the state who potentially have ELLs in their classes that could benefit from these types of instruction—potentially enlarging the unqualified ELL teaching force even more.

Faced with a shortage of teachers, particularly in urban schools, schools often make use of aides. While aides and paraprofessionals are valued and necessary employees, they are not appropriate as the primary (or only) instructors for ELLs (Proposition 227 Task Force, 1999). The danger of over-reliance on classroom aides for the education of ELL students can be seen at the national level, where 58% of ELL students in high-poverty schools have aides delivering their English/reading instruction, and less than 21% of nationwide ELL students have aides who possess any education beyond high school (Moss & Puma, 1995).

Beyond this, there is a larger problem of overall teacher qualification. In California alone, almost 15% of all public school teachers are working without a teaching credential. Ten percent of these un-credentialed classroom teachers do not hold a bachelor’s degree and are teaching on waivers (California Department of Education, Educational Demographics Unit, Statewide classroom teacher credential and experience report for the year 1999-00). ELLs in particular are greatly impacted by this. In California, schools with 40% or more ELLs have 6 times the percentage of teachers who are not fully credentialed than do schools that have 7.49% or fewer EL students (Rumberger & Gandara, 2000). In other words, the schools that ELL students are most likely to attend are the very ones in which qualified teachers are least likely to be employed.

There is also a mismatch between the racial and ethnic breakdowns of the nation’s teachers and students. In California for example, the teaching force is relatively homogeneous when viewed against the diversity of the state’s students. While minorities comprise only 24% of the teacher workforce (California Department of Education, Educational Demographics Unit, Number of teachers in California public schools by ethnic designation and gender, 1999-00,), 63% of the state’s students are minorities (California Department of Education, Educational Demographics Unit, Statewide enrollment in California public schools by ethnic group, 1999-00).

While it is critical that the ELL teaching force become more diverse and trained in the methodological and affective accommodations ELL students may need, having this group in place is just the beginning of what needs to be done to serve this population. Professional development is critical. Specifically, there is a need for continued, career-long professional development. Unfortunately, too often pre-and in-service programs provide superficial, cursory looks at ELL-specific issues but do not allow time for substantial teacher reflection (Walqui, 1997). Even completion of a program such as California’s CLAD or BCLAD certificate is insufficient training on its own when one is dedicating a career to ELL instruction. The resources and support necessary to affect long-term change are absent. Teachers need to have an opportunity for continued discussion and reflection upon their practices as a group. These discussions foster perpetual growth and enrichment of teachers’ skills as well as provide a peer group network to allow teachers to better manage their ever-changing student body populations and resulting teaching methods (National Research Council, 1997).

The recently completed efforts of the National Board for Professional Teaching Standards to develop teaching standards for bilingual and ESL teachers should be applauded as the "deluxe" model for professional development, but the magnitude of the problem is staggering when we look at the other elements of professional preparation such as schools of education, professional development models, and Title VII incentives. In addition, current knowledge about the effectiveness of strategies for teacher education and the assessment of teacher knowledge and skills is very limited. With respect to teacher competencies, there has never been a study that analyzed the kinds of teacher competencies that result in ELLs meeting high skill levels. States should fund this type of research to allow for the tailoring of credentials/certificates that are tied to what we know about critical teacher competencies for ELLs. Policymakers should demand a systematic inquiry into ways to understand, support, and coordinate all of these efforts.

Summary for Standard 2: In order to be compliant with Standard 2 of Castañeda, districts need to ensure that the following is accomplished: 1) ELL programs (regardless of type) are not seen as "one size fits all", but rather are developed after reflection of a particular program’s population and their needs; 2) language and content standards specific to ELLs are developed; 3) additional quality materials for instructing ELLs are developed and equitably distributed; 4) more teachers are trained to work with the ELL population; and 5) these teachers be offered on-going professional development.


Standard 3: Does the district evaluate its programs and make adjustments where needed to ensure language barriers are actually being overcome?

Program evaluation is often gauged in terms of student performance, yet there are several key issues that are raised by assessing ELLs, including 1) the need for inclusion of ELLs in the standards-based reform process, 2) the difficulty in deciphering results of content area assessments (i.e., are errors based on a lack of English proficiency or a content area knowledge deficit?); 3) the testing accommodations necessary for ELLs, particularly for high-stakes assessments; and 4) the need for English language assessments that more accurately reflect the type of language that a classroom demands – often called "academic English." Each concern has implications for states’ policies, as assessments administered to ELLs are used for re-designation purposes and have ramifications that extend far beyond the determination of their language needs. In fact, because of the serious inadequacy of ELL assessments, they have the potential to erroneously impact—and subsequently render inequitable—state-provided educational experiences.

As mentioned previously, the state should develop academic standards for ELLs and, through aligned assessments, illustrate compliance with Castañeda’s requirement of comprehensive content area instruction. The natural progression of these standards would be to design research studies to test their validity, especially with respect to policy-relevent junctures, such as identification and reclassification.

Up to now, ELLs’ progress has been gauged almost exclusively by their rate of English language development. However, instead of looking simply at English language proficiency as a measure of success, ELLs’ progress should also be determined based on their achievement of academic content standards that are developed specific to this population. Despite the need for evaluating ELLs’ progress against content area standards that are tailored specifically for that population, trying to capture ELLs’ comprehension of content material when this is measured through an English language test is a true challenge for ELL teachers. How can one separate out the problems a student has that are related to his/her English comprehension from those that reflect his/her difficulty in mastering the content? Taken on its own, this dilemma does not bode well for ELLs; however, the issue becomes even more critical when these content area assessments are used for high-stakes purposes (i.e., for student grade promotion, etc.).

Using content area assessments for high-stakes purposes is precisely what some districts in California are doing with the SAT-9, despite the known difficulties in separating out content knowledge deficits from problems of English language comprehension and despite the fact that the SAT-9 was normed with a very different population and was never intended to be used as an ELL assessment tool. Even in districts in which the SAT-9 is not used as a determiner of grade promotion, the test is still used as one of the primary measures for determining whether or not a student is able to be "redesignated" from an "English Language Learner" to a "Fluent English Proficient" learner. While California has been singled out here as an example, other states are guilty of the same behavior.

Given the previously mentioned concerns over the SAT-9, ELL advocates have publicly questioned the validity of ELLs’ scores on the SAT-9 and sought to find ways to meaningfully interpret them. From a civil rights perspective, ELLs are potentially being foreclosed from a meaningful education if a state-administered assessment (that they are required by the state to take) cannot provide an accurate evaluation of their performance and thus prohibits their grade promotion or places them in a class or academic track below their true skill level. There needs to be an end nationwide to the use of standardized tests like the SAT-9 as high-stakes tests for ELLs and alternative assessments need to be developed for ELLs, particularly ones to be used specifically for redesignation and accountability purposes. In this vein, longitudinal research that tracks the educational progress of former ELLs that have been deemed English proficient could also prove useful in pinpointing areas that are critical to assess prior to redesignating and ending specialized instruction for an ELL (National Research Council, 2000).

It is clear, then, that assessing ELLs can be a double-edged sword. By not including this population in assessments, one creates a recipe for the provision of a sub-par education without accountability. Yet, including ELLs in assessments that are invalid measures and that could potentially be used against them is harmful to the very students that the policies intend to serve (National Research Council, 2000). To mitigate this, states must acknowledge that when ELLs are assessed (particularly in a potentially high-stakes context like the SAT-9), consideration must be given to the appropriateness of accommodations that might be made for this particular group of students. Oftentimes on assessments like the SAT-9, the level of English vocabulary reflected in both the reading and listening passages is quite advanced and well beyond the beginning English vocabulary that would typically be employed in materials in ESL classes for students learning English as a new language. While the advanced English vocabulary might reflect English language arts as taught to native English speakers from kindergarten through high school, that vocabulary would be much more difficult, indeed, literally and figuratively foreign to ELLs. Research has illustrated that, depending upon the age of the child, it can take up to seven years before a majority of ELLs would have enough exposure to such vocabulary to be able to understand the language used on assessments (like the SAT-9) and to have a reasonable or fair chance to pass it on an equitable basis as native English speakers (Hakuta, Butler, & Witt, 1999).

Translating the assessment into the student’s native language is one potential accommodation that is frequently discussed, but this is not as easy a solution as it might appear, since tests that are translated into other languages often fail to preserve the integrity of the original test (National Research Council, 1997). Another possibility that is currently being tested is the use of dual language test booklets. With these booklets, students’ first language can theoretically be used to scaffold, or provide support for facilitating their understanding of the English version of the test. Other possible accommodations are administering alternative assessments with a reduced language load, awarding additional time, allowing students to use dictionaries, and simplifying test instructions (National Research Council, 1997), but one also needs to ensure that accommodations do not merely introduce further problems (e.g., becoming distracted and thumbing through a dictionary) that could take the focus off the task (National Research Council, 2000). It is easy to see what a disadvantage ELLs are operating from without viable testing accommodations. By not intervening, states are failing to meet their legal obligation under Castañeda v. Pickard to educate language minority students and are promoting inequitable assessment environments in the nation’s schools.

In order to set realistic standards and appropriate testing accommodation guidelines that will enable ELLs to both attain and demonstrate their full potential, there is a continued need for objective assessments of ELLs’ English language abilities in all four skills areas as well, including reading and writing, which take considerably longer to develop than oral English. Additionally, the validity of the assessments currently used, for example in California schools, to measure ELLs’ oral English language ability is questionable. For instance, many of these assessments do not tap students’ academic oral proficiency but rather solely their basic, daily language—or even merely their grasp of fundamental grammar principles, therefore not offering accurate information about students’ oral abilities as they pertain to a classroom environment (National Research Council, 1997). As a result of their performance of these invalid measures, ELLs can end up in a classroom that is inappropriate for their level of instruction. In order to get an accurate picture of ELLs’ English language proficiency level for classroom placement purposes, states should support the supplemental use of more authentic assessments of students’ academic language that more closely mirror classroom demands.

Summary for Standard 3: In order to be compliant with Standard 3 of Castañeda, districts need to involve ELLs in the standards-based reform process and assess them in both their content area knowledge and in academic English performance. However, accommodations must be available for ELLs and evaluations of their performance that are gained by using assessments normed on native English speakers cannot be used for high-stakes purposes.

An Additional Standard: Using data to make program improvements.

It is important to make explicit the point that the third standard set by Castaneda is intended to trigger adjustments in either the educational theory or in its implementation. This is itself a de facto fourth standard, that the educational system, be it a local or state agency, be prepared to use appropriate information on program effectiveness to improve its programs. Local and state system capacity for addressing the needs of ELL students based on sound information should be a further requirement in addition to the three standards.


California Department of Education, Educational Demographics Unit, Language Census, 1999. Number of staff members providing primary language instruction to English learner students in California public schools, by language of instruction, 1998-99 [On-line]. Available: http://www.cde.ca.gov/demographics/reports/statewide/lctch99.htm

California Department of Education, Educational Demographics Unit, Number of teachers in California public schools by ethnic designation and gender, 1999-00 [On-line]. Available: http://data1.cde.ca.gov/dataquest/NumTeachCo.asp?cChoice=StateNum&Radio2=T&cYear=1999-00&submit1=Submit

California Department of Education, Educational Demographics Unit, Statewide classroom teacher credential and experience report for the year 1999-00 [On-line]. Available: http://data1.cde.ca.gov/dataquest/TchExp1.asp

California Department of Education, Educational Demographics Unit, Statewide enrollment in California public schools by ethnic group, 1999-00 [On-line]. Available: http://data1.cde.ca.gov/dataquest/StateEnr.asp?cChoice=StEnrEth&cYear=1999-00&submit1=Submit

Castañeda v. Pickard, 648 F.2d 989 (5th Cir. 1981).

Crawford, J. (1997). Best evidence: Research foundations of the Bilingual Education Act. Washington, DC: National Clearinghouse for Bilingual Education.

Darling-Hammond, L. (2000). Teacher quality and student achievement: A review of state policy evidence. Education Policy Analysis Archives, 8(1). [On-line]. Available: http://epaa.asu.edu/epaa/v8n1

Greene, J. P. (1998). A meta-analysis of the effectiveness of bilingual education. Austin, TX: University of Texas, The Tomas Rivera Policy Institute.

Hakuta, K., Butler, Y.G., & Witt, D. (1999). How long does it take English learners to attain proficiency? University of California Linguistic Minority Research Institute Policy Report 2000-1.

Moss, M. & Puma, M. (1995). Prospects: The Congressionally mandated study of educational opportunity and growth: Language minority and limited English proficient students. Washington, DC: U.S. Department of Education.

National Research Council. (1997). Improving schooling for language-minority children. August, D. & Hakuta, K. (Eds.). Washington, DC: National Academy Press.

National Research Council. (2000). Testing English-language learners in U.S. schools. Beatty, A. & Hakuta, K. (Eds.). Washington, DC: National Academy Press.

Rumberger, R. & Gandara, P. (2000, October). The state of education for English language learners in California. Presentation at the University of California All Campus Collaborative, Outreach, Research, and Dissemination (UC ACCORD) Conference, San Jose, CA.

The Proposition 227 Task Force. (1999). Educating English learners for the twenty-first century. Sacramento, CA: California Department of Education.

Walqui, A. (1997). The development of teachers’ understanding: Inservice professional growth for teachers of English language learners. Unpublished doctoral dissertation, Stanford University.

Willig, A.C. (1985). A meta-analysis of selected studies on the effectiveness of bilingual education. Review of Educational Research, 55(3), 269-317.