bilingual education - reason foundation · director of education studies, ... and the impact of...

R P P I P o l i c y B r i e f

by richard c. sederby richard c. sederdirector of education studies, rppidirector of education studies, rppi

RPPI l 3415 S. Sepulveda Blvd., Suite 400 l Los Angeles, CA 90034 l 310-391-2245RPPI l 3415 S. Sepulveda Blvd., Suite 400 l Los Angeles, CA 90034 l 310-391-2245

reading, writing, and rhetoricreading, writing, and rhetoricbilingual education

RPPIThe Reason Foundation is a national re-search and educational organization thatexplores and promotes the twin values ofrationality and freedom as the basic un-derpinnings of a good society. Since 1978,the Los Angeles-based foundation has pro-vided practical public policy research,analysis, and commentary based upon theprinciples of individual liberty and respon-sibility and limited government.

about RPPI

contents

about the author

Reason Public Policy Institute (RPPI), a di-vision of the Reason Foundation, fusestheory and practice to influence public poli-cies. The Institute's research draws uponeconomics, science, and institutional analy-sis to critique public policies and advancenew policy ideas in a variety of policy ar-eas, including education, infrastructure andtransportation, the environment, urbanland use and local economic development,social services, and privatization and gov-ernment reform. To that analysis, the Insti-tute brings a political philosophy that sup-ports rule of law, marketplace competition,economic and civil liberty, personal re-sponsibility in social and economic inter-actions, and institutional arrangementsthat foster dynamism and innovation.

The Reason Foundation is a tax-exempt edu-cational organization as defined under IRScode 501(c)(3). The Reason Foundation nei-ther seeks nor accepts government fund-ing, and is supported by individual, founda-tion and corporate contributions. Nothingappearing in this document is to be con-strued as necessarily representing theviews of the Reason Foundation or its trust-ees, or as an attempt to aid or hinder thepassage of any bill before any legislativebody.

Photos used in this publication are Copyright© 1996. Photodisc, Inc. Copyright © 1998.Reason Foundation. All Rights Reserved.

Richard C. Seder is the director of education studies for the Reason Public Policy Institute, a national public policyresearch organization. Mr. Seder has done extensive work in the area of quantitative and qualitative research on thestructure of educational systems; school choice programs; accountability issues; and the impact of expenditures onstudent achievement. Before joining RPPI, Mr. Seder worked as a research assistant to the executive director for theCato Institute in Washington, D.C., where he assisted in a study analyzing the relationship between public and privatepost-secondary institutions in three states. Prior to that, he was the recipient of a Charles G. Koch Fellowship throughthe Center of Market Processes examining the effects of immigration on the U.S. economy. Additionally, Mr. Seder hasserved as a project director for the Allegheny County School System in Pennsylvania, where he conducted extensive research on 43public school districts. Mr. Seder holds a Masters of Science in Public Policy and Management from the Heinz School at Carnegie MellonUniversity in Pittsburgh as well as a double bachelors degree in Government and Economics from Beloit College in Wisconsin.

1ExecutiveSummary

1Introduction Background

2SpecialInstructionalPrograms

7ResearchResults

8Conclusion

13Glossaryof Terms

16Endnotes

17

bilingual education 11

ExExExExExecutive Summaryecutive Summaryecutive Summaryecutive Summaryecutive Summary

How best to serve limited-English proficient (LEP) students has been at thecenter of both the pedagogical and

political struggle known as bilingual educationfor more than 30 years. The eyes of the coun-try are focusing on California and what manyconsider the future of bilingual education.

Over 1.3 million students are classified as lim-ited-English proficient in California, nearly 40percent of the nation’s LEP population. Upuntil February 1998, the state mandated theuse of native language instruction to educateLEP students. However, for a variety of rea-sons, approximately 30 percent of LEP stu-dents received the majority of their academicinstruction in their native language, while 16percent received no special instructional ser-vices at all.

Because of data limitations and the generallack of evaluation, dropout rates and othereducational measures do not differentiate be-tween those students who received specialinstructional services and those that have not.The lack of comprehensive data lends itself tomisrepresentation, by both bilingual educationproponents and opponents.

Research on the effectiveness of bilingual edu-cation is at best inconclusive. Both sides ofthe debate over the use of native-languageinstruction cling to the research results thatsupport their particular argument. However,few research studies go without questionssurrounding methodology, sample, or in theinterpretation of results.

Consensus of research conclusions in favor ofone methodology over another is unlikely giventhe diverse nature of these students — lan-guages, family and community backgrounds,prior education and English exposure — andthe diverse circumstances that school districtsface. To expect such a result would ignorethe complexities surrounding the issue.

Given the conflicting research results of bilin-gual education and the diverse population oflimited-English proficient students and schools,one “best” instructional program should not bemandated across all schools and districts.

Instead, state and federal policymakers shouldprovide districts the opportunity to innovatewith instructional programs to properly meetthe needs of LEP students. At the same time,policymakers should incorporate evaluation andaccountability components into their policies toensure that those instructional programsadopted are effective.

IntroductionIntroductionIntroductionIntroductionIntroduction

The questions of whether limited-Englishproficient children should receive spe-cial educational services, to what ex-

tent school districts are responsible for providingthose services, and what type of special servicesshould be provided has been a matter of publicdiscourse for more than 30 years. Limited-Englishproficient students, while first protected and pro-vided for by Title VI of the Civil Rights Act of1964, and later by Title VII of the Elementaryand Secondary Education Act of 1968, have beenat the center of both the pedagogical and politi-cal struggle known as bilingual education.

With over three million students designated aslimited-English proficient (LEP), and nearly 40percent of those students located in California,how to meet the educational needs of thesestudents remains controversial. What instruc-tional practice, or practices, work? And whatdoes “work” mean? Should schools concen-trate solely on English proficiency as quickly aspossible? If so, are sacrifices made and timelost in the other academic subjects, and whatis the quickest way to get these students En-glish-proficient? What methods get LEP stu-dents at a par with native-English speakersacross all subject areas? Is it appropriate tothink in terms of one-best program at all, givendiverse goals and needs?

117

Reason Public Policy Institute82

Obscured in the political rhetoric is the termbilingual education. What constitutes bilin-gual education instruction? The various sidesof the debate clutch to their anecdotal andempirical evidence that definitively proves thatbilingual education is either a rousing successor a dismal failure. Estimates of the expendi-tures on providing special services range fromnot-nearly-sufficient to an extraordinary andgrowing burden on state and local resources.

Can both sides of the argument be correct intheir assertions?

What is clear in the debate is the need to meetthe special circumstances of this diverse andgrowing population.1 Both bilingual educationproponents and opponents agree that account-ability in the current system is almost non-existent. Both sides agree that poorly imple-mented programs are resulting in the failureto provide a meaningful education to thesestudents.

This policy brief provides a brief overview of:

n the status of limited-English proficient stu-dents;

n the nature and definition of services beingadministered;

n the policies that govern how these stu-dents are served; and

n a synopsis of research evaluating the ef-fectiveness of special instructional servicesprovided to LEP students with an empha-sis on California.

The purpose is not a pedagogical endorsementof one form of instruction over another. Rather,the purpose is to provide an objective under-standing of the issues for sound policymakingat the local, state, and federal levels, and byparents concerned about the nature of servicesbeing provided to their children.

BackgroundBackgroundBackgroundBackgroundBackground

A. StudentsA. StudentsA. StudentsA. StudentsA. Students

During the 1994-95 school year, over threemillion students across the nation were clas-sified as limited-English proficient, or LEP. Cali-fornia alone serves over 1.3 million LEP chil-dren, nearly a 50 percent increase from 1990and nearly four times as many as in 1980,constituting more than one of every five stu-dents in California schools, and representing39.9 percent of the nation’s total LEP popula-tion.2 While the issue of serving limited-En-glish proficient students faces every state, over75 percent of the nation’s LEP population arelocated in four states, California, Texas, NewYork, and Florida (see Table 1).

In California, over 100 languages and dialectsare spoken by students attending elementaryand secondary schools. There is a heavy con-centration of LEP students in elementaryschool, with over 69 percent in grades K-6.More than 40 percent of these LEP studentscan be found in the early kindergarten throughthird grade (Figure 1).

While the task seems daunting, over 80 per-cent of these students speak Spanish, andnearly 92 percent speak Spanish, Vietnam-ese, Tagalog (Phillipines), Korean, Hmong,Khmer (Cambodia), and Cantonese.3

Given the growth in this segment of the stu-dent population, comparison to previous gen-erations of immigrants and the type of edu-cation that they received proves difficult. FullEnglish immersion may have been the onlyoption available, and had some success withLEP students, given the relatively small num-bers in the schools. Given the small num-bers, more personalized attention may havebeen given to these students both within theclassroom and through peer coaching. Re-gardless, in the 1960’s, nearly 50 percent ofMexican-American LEP youth did not com-plete the eighth grade.4


table 1

BBBBB. Federal Court Decisions, Laws,. Federal Court Decisions, Laws,. Federal Court Decisions, Laws,. Federal Court Decisions, Laws,. Federal Court Decisions, Laws,and Pand Pand Pand Pand Policyolicyolicyolicyolicy

The issues of whether or not to provide spe-cial services and what kind of services to pro-vide to limited-English proficient students arespelled out in federal and state laws, courtrulings, and policies and regulations.

Limited-English proficient students were firstrecognized under federal law under Title VIof the Civil Rights Act of 1964. Title VI en-sures equal educational opportunities for stu-dents by requiring that there be no discrimi-nation on the basis of race, color, or nationalorigin in the operation of any federally as-sisted programs.

To meet the specific needs of LEP students,Title VII of the federal Elementary and Sec-ondary Education Act of 1968, also knownas the Bilingual Education Act, was estab-lished to provide funding for various educa-tion programs to meet the needs of this stu-

dent population. The act was recently reau-thorized in 1994 under the ImprovingAmerica’s Schools Act (IASA).

The basis for the continued reauthorization ofthe Bilingual Education Act rests with federaland U.S. Supreme Court decisions. Perhaps thetwo most influential decisions regarding the pro-vision of special services to LEP students arethe 1974 Lau and 1981 Castaneda decisions.

Based on the inclusion of limited-English profi-cient students under Title VI of the Civil RightsAct, the U.S. Supreme Court, in 1974, de-cided in Lau v. Nichols that an equal educa-tion is not provided if non-English-speakingstudents cannot understand the teachers, text-books, or curriculum, in this case Chinese stu-dents in San Francisco.5

In the Lau case, the Court did not detail a spe-cific remedy or pedagogical solution. Instead,it deferred to Department of Health, Educa-tion, and Welfare guidelines:

1. California 1,262,982 39.9% 39.9%

2. Texas 457,437 14.5% 54.4%

3. New York 236,356 7.5% 61.8%

4. Florida 153,841 4.9% 66.7%

5. Puerto Rico* 143,769 4.5% 71.2%

6. Illinois 107,084 3.4% 74.6%

7. Arizona 98,128 3.1% 77.7%

8. New Mexico 84,457 2.7% 80.4%

9. New Jersey 52,081 1.6% 82.0%

10. Washington 51,598 1.6% 83.7%

STATES WITH THE LARGEST LEP ENROLLMENTS, 1994–95

LEP Enrollement Nat’l Enrollment % Cumulative %

Source: Data is from the SEA Survey forms submitted by the SEAs. *Limited SpanishProficient is used in place of Limited-English Proficient for Puerto Rico.


Where inability to speak and understand theEnglish language excludes national origin-mi-nority group children from effective participa-tion in the educational program offered by aschool district, the district must take affirma-tive steps to rectify the language deficiencyin order to open its instructional program tothese students.6

By leaving the specifics out, the Supreme Courtallowed for flexibility on the part of school dis-tricts to meet the needs of its student popula-tions. However, further clarification wasneeded to determine whether or not the ser-vices being provided were adequate by a school.

In 1981, a federal court set out in Castanedav. Pickard guidelines for determining whethera school had met its obligations under federallaw to provide adequate special services.These guidelines became the federal three-prong test, which included:

1. The school must create a program for non-English-speaking students based on aneducational theory that is recognized assound by at least some experts in the field

or that is recognized as a legitimate edu-cational strategy.

2. The school must have programs, policies,and resources in place that could be rea-sonably expected to implement effectivelythe chosen educational theory.

3. The school’s program must demonstratethat students are making progress in over-coming language barriers. No matter howreasonable a school’s original choice ofprogram may be or how exhaustive arethe resources dedicated to the program,the failure of students to make progressobligates the school to revise its program.7

According to interpretation by California’sLittle Hoover Commission, “This decision alsomade it clear that while schools must havetwo goals — helping students attain Englishproficiency and ensuring that they make aca-demic progress in the overall curriculum —the schools are free to pursue the goals se-quentially rather than simultaneously.”8 Thisallowed districts to choose and implementprograms that may sacrifice a student’s over-

GRADE LEVEL OF STUDENTS NOT FLUENT IN ENGLISH

K-3 44.93%

4-6 24.24%

7-9 18.25%

10-12 11.29%Ungraded 1.29%

figure 1Source: CA State Department of Education, 1997.

15

all academic learning with respect to native-English speakers while learning English. How-ever, parity with native-English speakers byLEP students must be achieved within somereasonable time upon entering the school.

Title VII instructional programs have focusedon the dual purpose set out in Castaneda:

The Congress finds that (8) it is thepurpose of this title to help ensure thatlimited English proficient students mas-ter English and develop high levels ofacademic attainment in content area;9

In 1994-95, approximately 300,000 limited-English proficient students nationwide re-ceived their education through Title VII-fundedinstructional programs, roughly nine percentof the total LEP population. Five major in-structional programs are designed to serveLEP students under Title VII (the number inparentheses denotes the percentage of totalLEP students receiving that particular programin 1994-95): Transitional Bilingual Education(6 percent), Developmental Bilingual Educa-tion (0.3 percent), Special Alternative Instruc-tional Program (2 percent), Family LiteracyProgram (0.1 percent), and the Special Popu-lations Program (1 percent).10

Title VII grants are provided on a competitivebasis. These grants have tended to favorprograms that utilize native-language instruc-tion, but since 1984, a portion of Title VIIfunds — 25 percent — may go to supportalternative, all-English approaches (includedin Special Alternative Instructional Plans).11

These funds serve a capacity-building purposethat allows districts to implement new pro-grams that eventually become self-support-ing. However, it must be noted that the capof 25 percent to non-native-instructional planscreates undesirable incentives when design-ing new instructional plans for language-mi-nority students.

C. California LawC. California LawC. California LawC. California LawC. California Law

From 1872 until 1967, California law requiredall instruction to be in English under Section

71 of the California Education Code. In 1967,this statutory mandate was repealed by Sen-ate Bill 53. In 1976, the state enacted theChacon-Moscone Bilingual/Bicultural EducationAct, followed by the Bilingual Education Im-provement and Reform Act of 1980.

In 1987, the California bilingual education lawwas allowed to “sunset” with the intention ofproviding school districts the flexibility to de-sign and implement instructional programs forlanguage-minority students otherwise not al-lowed under the strict standards of the exist-ing law. While the law was allowed to sunset,the California Department of Education issued“advisories” to school districts. These adviso-ries stated that school districts must continueto comply with the “general and intended pur-poses” of the original law, Education Code52161, thus rendering the sunset policy moot.12

The California Department of Education andBoard of Education set requirements that schooldistricts were to meet, similar to the three-prongtest outlined in Castaneda and allowing for flex-ibility at the district level outlined in Lau. Inshort, the guidelines required districts to pro-vide an instructional program that taught LEPstudents English, that provided equal academicopportunity, and to provide the necessary re-sources to effectively implement the program,both fiscally and through personnel.

However, the department enforced a specifictype of program that emphasized native-lan-guage instruction as the preferred method ofeducation for LEP students. This was donethrough the Coordinated Compliance Reviewprocess in which districts had to satisfy 12requirements. Below are some of the require-ments that essentially mandated native-lan-guage instruction:

n Place students who do not speak Englishfluently into a program of instruction inEnglish language development.

n Give each student primary-language accessto the core curriculum (based on the level ofproficiency in English and native language).


n Make specially designed academic instruc-tion in English available to those studentswho are advanced enough in English towarrant it.

n Assign an adequate number of “qualified”teachers to implement the English lan-guage-development program.

n Assign an adequate number of “qualified”teachers to implement the primary-lan-guage-instruction program.13

Districts that did not meet the twelve require-ments set forward by the department wereprovided six options:

1. Demonstration of educational results—al-lows a school to adopt an alternative-in-structional program, but the school mustbe able to prove that students perform ata level equal to or greater than the state-wide averages. The burden of proof is uponthe district and requires additional re-sources dedicated to the department’sevaluation process;

2. Assignment of teachers with Commissionon Teacher Credentialing (CTC) authoriza-tions—deals with qualifications of teach-ers, all staff have special certificates is-sued by the Commission on TeacherCredentialing. Five hundred districts chosethis option in 1991-92;

3. Local designation of other qualified teach-ers—districts may set up their owncredentialing program for staff under de-partment review;

4. Plan to remedy shortage of qualified teach-ers—if a school district has made every at-tempt to obtain the specified number of quali-fied instructors and comes up short, the dis-trict can devise a plan to remedy the situa-tion over a specified period of time. Some490 districts chose this option in 1991-92;

5. General waiver authority—if Option 4 fails,the district may request a waiver from thatspecific requirement, or for an alternative-

instructional program. The waiver doesnot dismiss the district from federal re-quirements. Approximately 300 districtsqualify for this option;

6. Small and scattered LEP populations—ifa district has fewer than 51 and no singleschool has more than 20 students in aparticular language group, the district isexempt from meeting the compliance re-view for that language group. About 850districts qualified for this option for oneor more language groups.14

Two major developments occurred in 1998that have dramatically changed the way dis-tricts provide special educational services tolimited-English proficient students.

In February 1998, a Sacramento SuperiorCourt judge ruled, in a case involving an Or-ange County school district, that the intentof the state’s expired bilingual education lawwas to “effectively and efficiently as possibledevelop in each child fluency in English(Chacon-Moscone Bilingual/Bicultural Educa-tion Act)” and that primary-language instruc-tion was not required and that the state boardwas wrong in requiring districts to obtainwaivers from native-language instruction.15

On March 12, 1998, the California StateBoard of Education, following the lead ofJudge Ronald Robie’s decision, voted unani-mously to rescind its long-standing policiesof requiring native-language instruction to LEPstudents. The vote allows school districts todecide the instructional programs for LEP stu-dents at the local level, including English-onlyprograms, without obtaining waivers from thestate. While districts are no longer requiredto comply with California mandates, they arestill required to comply with federal law.

DDDDD. California Funding. California Funding. California Funding. California Funding. California Funding

Failure to comply with the state requirementsor satisfy the department through one or moreof the available options, in the past, resulted

in delays, and potential loss, of Economic Im-pact Aid (EIA) funding targeted for educa-tionally disadvantaged and limited-English pro-ficient students.

Over $300 million is allocated through theEIA program. Many times, LEP students alsoqualify for educationally disadvantaged fund-ing as well. Because of this designation, manyLEP students qualify for federal Title I fund-ing that targets at-risk students.

Costs for providing special instructional ser-vices to LEP students are estimated to beroughly equal to those costs of regular in-structional programs.16 Little is actuallyknown about the true costs of providing spe-cial instructional services to LEP studentsbecause of the block grant nature of fundsdisseminated to the schools, such as thestate’s EIA program. In many instances, thesefunds are pooled together in the school’s op-eration budget without specification to par-ticular program areas.17 This allows schoolsthe flexibility to meet the specific needs ofits student body.

Special InstructionalSpecial InstructionalSpecial InstructionalSpecial InstructionalSpecial InstructionalPrograms for Limited-Programs for Limited-Programs for Limited-Programs for Limited-Programs for Limited-English ProficientEnglish ProficientEnglish ProficientEnglish ProficientEnglish ProficientStudentsStudentsStudentsStudentsStudents

A. California PA. California PA. California PA. California PA. California Participationarticipationarticipationarticipationarticipation

While there is some overlap in definitions be-tween some of the program models, there isalso overlap in the implementation of theseprograms as well. However, in most stud-ies, schools often identify their program un-der one of the general categories. Table 2illustrates instructional services provided toCalifornia LEP students.

Dropout rates for Hispanic students in theUnited States are higher at 12.4 percent in1995 compared to 4.5 percent for whites and6.4 percent of blacks (California dropoutrates: 5.6% for Hispanics; 2.4% for whites;

and 6.6% for blacks).19 While this figure rep-resents LEP and English-fluent Hispanics, thosestudents speaking Spanish at home (again notan accurate indication of English fluency) weremore likely to have repeated at least onegrade.20 However, we should be hesitant todesignate the dropout rate as an indication thatbilingual education has failed. Hispanics tendto drop out more, not because of academicachievement (4 percent). Rather, 38 percentof Hispanic dropouts gave economic reasons(desire to work, financial difficulties, home re-sponsibilities) for dropping out.21 Also, verylittle research has disaggregated the dropoutfigure according to instructional practices re-ceived.

In 1996-97, 16 percent of limited-English pro-ficient students did not receive any special in-structional services. This is primarily due tothe scattered language populations, exempt-ing schools from offering special instructionalservices. Only 29.7 percent of California LEPstudents received the majority of their aca-demic instruction in their native language. Anadditional 21.6 percent of LEP students re-ceived some native-language support fromsome type of aide. The remaining students(32.7 percent) received English instruction asthe primary mode of academic instructioncoupled with ESL instruction with no native-language assistance.22

According to an analysis of instructional pro-grams implemented to serve LEP students doneby Development Associates, Inc., it is prima-rily districts with large numbers of LEP stu-dents (1,000 or more) that tend to have moreintensive native-language-instructional pro-grams, where native-language instruction isused 50 percent of the day or more.23 It isprimarily through state funding programs (75percent or more) that these same districts pro-vide their chosen instructional program.24

For example, the way that California’s LEPprogram is funded, through the Economic Im-pact Aid program, districts have extra incen-tive to classify students as non-proficient in



English. This is also true for many other stateand federal programs, such as the federalgovernment’s Title I program. Because of thisincentive structure, closer analysis must bedone with regards to the number of studentsclassified as limited-English proficient, as wellas the reclassification, or redesignation rate,the rate at which LEP students are no longerconsidered limited-English proficient (see Fig-ure 2). Classification may sometimes proveto be a financial decision made on the parts ofschool and district administrators rather thana pedagogical decision.

RRRRResearch Research Research Research Research Resultsesultsesultsesultsesults

A. �Bilingual Education Programs DoA. �Bilingual Education Programs DoA. �Bilingual Education Programs DoA. �Bilingual Education Programs DoA. �Bilingual Education Programs DoNot WNot WNot WNot WNot Workorkorkorkork��

Strong criticism is being made of bilingual edu-cation programs across the country. Anti-bi-lingual education stories center around thetransitional and developmental bilingual edu-cation program models, those that utilize na-tive-language instruction. Stories abound ofstudents graduating without any proficiencyin English, or Spanish, for that matter, of stu-dents being placed in inappropriate native-lan-guage classes, of parents unable to take ad-vantage of the voluntary nature of the bilin-gual education programs, etc.

There is good reason for concern. The num-ber of students being classified as non-En-glish or limited-English proficient continuesto grow as the number of students being clas-sified as English-proficient does not.

While Figure 2 is alarming, bilingual educa-tion proponents claim that traditionaltransitioning takes place anywhere from twoto seven years into the program. However,the Little Hoover Commission noted,“Whether one expects students to transitionin two, three, four, five or even six years,there is no “bulge” in the redesignation fig-ures that accounts for the eventual transi-tion of the hundreds of thousands of studentswho are not fluent in English.”25 Given thispattern of program achievement, thousandsof students would eventually graduate with-out fluency in English. But again, there maybe perverse financial incentives involved thatmay limit the official redesignation rate, al-lowing for actual numbers of students identi-fied as proficient being much greater.

Opponents of native-language instructionalprograms turn to the evaluative research ofChristine Rossell. Rossell, and her colleaguesJ. Michael Ross and Keith Baker, reviewed theresearch done on bilingual education, andfound that of the 300 program evaluations per-

table 2

English Language Development (ELD) Only 158,640 11.5%

ELD and “Sheltered English” Instruction 274,845 19.9%

ELD, sheltered English and primary 298,395 21.6%language support by paraprofessionals

ELD and academic subjects through 410,127 29.7%the primary language

Not receiving any special service* 239,386 17.3%

California totals 1,381,393 100.0%

LEP INSTRUCTIONAL SERVICES

Number % of Students

Source: California Department of Education, 1997. *includes withdrawn at parents request

19

formed up to the point of their research, only72 proved to be methodologically acceptable— true experimental design with random as-signment to control and treatment groups,controlled for student and family characteris-tics, the comparison group included LEP stu-dents of the same ethnicity and language back-ground, outcome measures did not includegrade equivalents, and additional educationaltreatments were eliminated or controlled for.

Table 3 shows the comparison of transitionalbilingual education (TBE) to submersion, En-glish as a Second Language instruction, anddevelopmental bilingual education (DBE) onsecond language reading, language, and math.

The percentages in Table 3 indicate the per-centage of studies showing a program doingbetter than the alternative program being com-pared to, the percentage showing no difference,and the percentage of studies that showedworse performance than the alternative. In-cluded in the table are the number of studiescomparing programs. This method of analysisis often called the vote-counting method.

When comparing transitional bilingual educa-tion to submersion, the method of doing noth-ing for LEP students, 22 percent of studiesshowed TBE to be superior, 33 percent to beinferior, and 45 percent to be no different intests measuring reading. When comparingTBE and submersion in language arts, 7 per-cent showed TBE to be superior, 64 percentshowed TBE to be inferior, and 29 percentshowed TBE to be no different.

When comparing TBE to structured immersion,taking the multiple forms of full immersion andtwo-way immersion programs, no study showedTBE to be superior to structured immersion inreading, language arts, or math. From a policystandpoint, because of the multiple forms clas-sified under structured immersion, we shouldpay careful attention to which of the structuredimmersion programs did better than the TBE pro-grams and which did no better. This delineationis unclear from the results given.

Researchers continue to debate the method-ological validity which Rossell used in herchoice of studies to include in her analysis.26

We should also pay close attention to the na-ture of the evaluations done, who was tested,at what point were they tested, what werethe measured differences between programsbefore coming to the definitive conclusion thatnative-language programs are wholly ineffec-tive. Another source of debate is the methodof vote-counting which continues to be ofdebate in the scientific community in makingconclusive determinations about the effective-ness of competing programs.27

Though Rossell’s analysis has some problems,it should not be completely discounted, andresearch cited as showing that native-languageprograms are superior in providing educationalservices to LEP students, as measured by stu-dent test scores in multiple subject areas, areequally problematic.

BBBBB. Native. Native. Native. Native. Native-language instruction not-language instruction not-language instruction not-language instruction not-language instruction notthe only waythe only waythe only waythe only waythe only way

Another evaluative research study that native-language program opponents point to is theRamirez study conducted for the U.S. Depart-ment of Education. The Ramirez study com-pared two types of native-language-instructionprograms (transitional and developmental bi-lingual education) and immersion programsfrom districts in California, Florida, Texas, NewYork, and New Jersey.

All three programs proved to be effective inproviding educational services to limited-En-glish proficient students, with all studentsreaching comparable skills levels in reading,language arts, and math. “Students in all threeprograms realized a growth in English-languageand reading skills that was as rapid or more sothan the growth that would be expected forthese children had they not received any in-tervention.”28 Because of this, native-languageproponents also point to the Ramirez study asjustification for their program models. At thevery most, the Ramirez study put forward the

bilingual education

notion that native-language instructional pro-grams are not the only effective method ofteaching LEP students.

California also performed a study similar tothe Ramirez study, performed by BW Associ-ates, and known as the BW study. The BWstudy compared five different instructional ap-proaches for LEP students: ESL, ShelteredEnglish, Late-Exit Bilingual Education, Early-Exit Bilingual Education, and Two-Way Bilin-gual Education programs.

BW studied five schools, each considered tohave an optimally implemented program inplace. The conclusion was very similar to theRamirez conclusions, that the success of eachprogram is greatly dependent on how they areimplemented by the individual schools. BW,realizing the complexities surrounding the edu-cating of limited-English proficient students,concluded, “The public debate about this ques-tion has too often been cast as a choice be-tween bilingual or English-only programs. Thechallenge of educating [English learners] is

much too complex to be reduced to such asimplistic formulation.”29

While the Ramirez and BW studies do not de-termine which program is most suitable tothe education of LEP students, bilingual edu-cation opponents should point to the notableconclusion that all optimally implemented pro-grams appear to provide LEP students withproper academic instruction and English-lan-guage development, not just native-languageinstructional programs.

C. Bilingual education and laborC. Bilingual education and laborC. Bilingual education and laborC. Bilingual education and laborC. Bilingual education and labormarkmarkmarkmarkmarket earningset earningset earningset earningset earnings

Finally, the question must be asked whetherbilingual education is a worthwhile investmentin the form of labor market outcomes by partici-pating students. Mark Lopez and Marie Mora,utilizing a national sample of students, discov-ered that Hispanic LEP students who have re-ceived some bilingual education at some pointin their schooling do not appear to earn sig-nificantly more or less than similar peers whowere given English-immersion instruction.


NUMBER OF STUDENTS TRANSITIONING FROMNONFLUENT TO FLUENT STATUS (1982�97)

Source: State Department of Education

82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 970

200

400

600

800

1000

1200

1400

1600Students (thousands)

Total Non-English Fluent Students

Redesignated as Fluent

figure 2Source: CA State Department of Education, 1997.

However, when they disaggregated the stu-dent population by how recent their familiesimmigrated to the United States—first-gen-eration, second-generation, third-generation— the authors found that the more recentthe immigrant, the worse off they were inearnings if they were provided bilingual edu-cation. For first-generation immigrants whoreceived bilingual education, wages were ap-proximately 50 percent less than their English-immersed peers. Second-generation studentsearned approximately 30 percent less. Therewas no significant difference in wages withinthe third-generation student population basedon bilingual education exposure.30

The authors note, however, that the samplethat they are analyzing may be biased. Thestudents in this sample were first added whenthey were sophomores in high school. Thiseliminates all students that dropped out priorto this stage in schooling. If non-bilingualeducation Hispanic LEP students drop out ata greater rate than those receiving bilingualeducation, the study results could be skewed.

Secondly, Lopez and Mora’s analysis does notdistinguish the type of bilingual education ser-vices provided, nor the duration that the stu-

dents were involved in these programs. Theauthors refer to bilingual education as any in-structional program that utilizes the child’sprimary language. This includes English as aSecond Language (ESL) programs where in-structional assistance in the primary languagevaries from one class to half the school day.The authors note the heterogeneity of bilin-gual education programs offered to students.

Finally, the sample of students analyzed wasfirst questioned in 1980, six years after theLau decision. It could be argued that veryfew schools offered a thorough and well-de-signed bilingual education program for stu-dents in their native language at that time.

DDDDD. �Native. �Native. �Native. �Native. �Native-Language Instruction-Language Instruction-Language Instruction-Language Instruction-Language InstructionPPPPPays Off in the Long Runays Off in the Long Runays Off in the Long Runays Off in the Long Runays Off in the Long Run��

Researchers from George Mason University,Wayne Thomas and Virginia Collier, com-pleted perhaps the largest evaluative studyof bilingual education programs to date, in-cluding over 700,000 student records from1982 to 1996. Their analysis was done bothcross-sectionally (different students at differ-ent points in time) and longitudinally (samestudents over time).


table 3

TBE v. Submersion (Do nothing)TBE Better 22% 7 % 9 %No Difference 45% 29% 56%TBE Worse 33% 64% 35%Total N 60 14 34

TBE v. Submersion/ESLTBE Better 19% 6 % 11%No Difference 48% 35% 55%TBE Worse 33% 59% 34%Total N 67 17 38

TBE v. Structured ImmersionTBE Better 0 % 0 % 0 %No Difference 17% 100% 63%TBE Worse 83% 0 % 38%Total N 12 1 8

EFFECTIVENESS OF BILINGUAL EDUCATION BY TEST OUTCOMES

Reading* Language Math

Source: Is Bilingual Education an Effective Educational Tool?, Center for Equal Opportu-nity, Table 1, p. 23. *Oral English achievement for preschool programs.

Their research questions centered around:“Which instructional practices lead to even-tual achievement parity between English learn-ers and native-English speakers?” This is thedefinition of “works” that the authors use intheir evaluation.

Their research followed students in “well-de-signed and implemented” programs of two-way bilingual education, developmental bilin-gual education with content-based ESL, tran-sitional bilingual education with content-basedESL, transitional bilingual education with tra-ditional ESL instruction (typically remedial styleinstruction, structure versus content), ESL pro-grams taught through academic content, andESL pullout programs (structure-based versuscontent-based instruction). The authors de-fined “well implemented” as those programs

that have been in effect for a number of years,with teachers that have been thoroughlytrained to fully deliver the chosen instructionalprogram completely, and with substantial ad-ministrative support to help deliver the instruc-tional program in its entirety. These condi-tions fully satisfy the Castaneda tests to meetthe needs of LEP students.

Collier and Thomas’s findings concurred withmany of the other evaluations of bilingual edu-cation programs; there are no short-term (twoto three years) differences in academicachievement between traditional ESL and de-velopmental bilingual education programs.

Note that the researchers defined the TBE andDBE programs as those that provided native-language instruction and English-content in-


PATTERNS OF K-12 ENGLISH LEARNERS� LONG-TERM ACHIEVE-MENTS IN NCEs ON STANDARDIZED TEXTS IN ENGLISH READ-ING COMPARED ACROSS 6 PROGRAM MODELS

figure 3Grade

Source: Wayne P. Thomas and Virginia P. Collier, 1997.

struction “equally” during the school day andwhere English-language development was pro-vided throughout the curriculum rather than ina limited and separate capacity. Also impor-tant is the emphasis given to each language.Those programs that stressed the importanceof each language equally in the curriculum werethe most successful — that is, primary lan-guage was as important as English, but alsoEnglish was as important as the native lan-guage.31

The two-way and developmental bilingualeducation programs evaluated provided stu-dents with long-term achievement parity withnative-English speakers in four to seven years.However, LEP students in the other programsmake only enough progress relative to En-glish speakers to maintain the initial achieve-ment gap, but do not achieve academic par-ity with their native-English-speaking peers.Collier and Thomas point out that transitionalbilingual education, as it is often implementedin schools, fares no better than the best ESLprograms implemented.

Part of this disparity may come from the factthat those students classified as English-pro-ficient are what some researchers label con-versationally English proficient and not aca-demically English proficient.32 Jim Cumminsasserts that a significant level of conversa-tional fluency can be achieved in two to threeyears, whereas fluency in a language to fullygrasp cognitively demanding material may re-quire five to seven years to attain. The au-thors assert that If students are redesignatedtoo soon, they typically do not achieve aca-demic parity with native-English speakers.

Collier and Thomas conclude their study byrecognizing some of the limitations to fullyimplementing many of the instructional pro-grams described (too few students, shortageof bilingual staff, etc.) and the failure of manybilingual education programs currently imple-mented to serve limited-English proficient stu-dents. They recommend that districts teachLEP students with a cognitively challengingcurriculum, both in the native language andin English — to treat the program as an en-

richment program rather than a remedial pro-gram. They also recommend that districtswhose instructional programs are failing to de-velop a new program that better meets stu-dent needs, both short-term and long-term.

One of the major criticisms of Collier andThomas’s work involves the sample of schoolsand students studied. Five urban and subur-ban districts were chosen based on the au-thors’ assessment of which programs werewell implemented to avoid inclusion of imple-mentation problems in their analysis. How-ever, this non-random choice of districts raisesselection bias questions in the analysis.

Collier and Thomas’s analysis also limits theirstudy to following students within the speci-fied districts. The authors recognize the rela-tively high mobility rates within this studentpopulation, so those students that leave thegiven district are lost from the sample andanalysis.

The authors also do not disaggregate the per-formance of recent immigrants and second andthird generation LEP students in the variousprogram methods. Can we assume that oneprogram model is equally effective across allLEP student sub-populations?

Conclusion and PConclusion and PConclusion and PConclusion and PConclusion and PolicyolicyolicyolicyolicyConsiderationsConsiderationsConsiderationsConsiderationsConsiderations

The term bilingual education brings withit many connotations and also manyemotions. Native-language instruction

opponents point to the number of studies thatshow that native-language instruction is worse,or at best no better, than English-intensivemethods, and they point to detrimental labor-market effects for recent immigrant childrenexposed to bilingual education. Native-lan-guage proponents point to the long-term ef-fects of native-language instruction as the ra-tionale for continued implementation. Yet, onlythrough well-implemented programs do LEPstudents reach academic parity with their na-


tive-English peers while gaining full cognitivefluency in English.

Neither side of the argument is necessarily in-correct in its assertions. One element thatboth sides acknowledge is the importance ofinstructional programs, no matter what levelof native-language use, being well imple-mented in order to properly serve students withlimited-English proficiency. Of course, this cir-cumstance is no different from any instruc-tional program. Success depends on actualimplementation details.

The original intent of both the Lau andCastaneda decisions was to allow for experi-mentation with several models to serve LEPstudents. Federal law and states — until re-cently, California included — that favor oneinstructional program over another ignore thediversity of the limited-English proficient stu-dent population. An emphasis on “one-best-way” also ignores the specific circumstancesfaced by individual school districts as they at-tempt to meet the educational needs of thisspecific and the entire student population.

In California, legislation has been proposed thatwould allow school districts to make all deci-sions locally with regards to curriculum, in-cluding curriculum to serve LEP students.California has been without legislation pertain-ing to serving LEP students since 1987. Inthis void, the state’s school board issued itsown policies and advisories, which favoredcertain bilingual education programs. Schooldistricts many times deferred to the state boardthus reducing their own initiative and com-mitment. The same void, along with the stateboard’s favoring of native-language instruc-tion, has also prompted popular legislative ini-tiative activity that would institute an alterna-tive state policy on how districts are to serveLEP students. The recent court ruling andCalifornia state board vote are moves towardssolidifying state policy with regards to edu-cating this student population by saying therole of the state is not to prescribe specificprograms.

One-size-fits all policies that favor one instruc-tional approach at the expense of alternatives,at both the federal and state levels, have failedand will continue to fail to serve the instruc-tional demands of this student population.Native-language instruction and English-im-mersion programs can both succeed whenschools and parents dedicate themselves toeach particular program. Local officials shouldproperly assess the needs of their LEP stu-dent populations and define programs to meetspecific needs of the students. Favoring ormandating one instructional program at theexpense of another, given the diverse needsof the LEP student population, does not al-low districts and schools to innovate with pro-gram models that best match the circum-stances that they face. State and federalofficials should concentrate on district com-pliance with the Lau and Castaneda decisionsrather than compliance with a favored modelthat may be impractical, inappropriate, or in-feasible for a particular district.

Policymakers and architects of instructionalprograms to meet the needs of LEP studentsshould consider several goals central to thedesign of effective policies and programs.First, they should evaluate how quickly stu-dents achieve English proficiency and to whatlevel of proficiency. Secondly, full academicparity across all academic subjects with na-tive-English speakers should be a goal ofwhatever program is implemented. The nec-essary means must be undertaken to meetboth of these goals.

As an example, Collier and Thomas assertthat LEP students, without thorough native-language instruction, do not typically closethe gap with their English-speaking peers.This conclusion, however, assumes thatschools teach all students the same way.While their assertion may lead some districtsto move towards greater native-language in-struction, others may move to a longer schoolyear or after-school and summer enrichmentprograms to bridge the achievement disparity.


Lastly, overall program evaluation and account-ability is essential to properly meet the needsof the LEP population. If a program is not suc-ceeding according to the first two goals above,then an alternative instructional plan shouldbe implemented. The role of the federal andstate governments should shift from ensuringdistrict compliance with mandates to a purelyevaluative role. Districts should be evaluatedon their demonstration of academic and En-glish-language progress with these studentsrather than on adherence to administrativeguidelines or specific programs.

School districts, especially large school dis-tricts with many schools, are also cautionedagainst adopting a one-size-fits-all instruc-tional approach across all district schools.Doing so ignores the diversity of needs invarious LEP sub-populations. While institut-ing multiple instructional methods acrossschools within a single district make evalua-tion more difficult, we should not assume thatstudents residing within the same geographicboundaries all learn in the same manner. Weshould assume that instructional programs willvary across districts throughout the state, butalso recognize that instructional programsmay vary throughout schools within districts.

The recent California state board decision al-lows districts to develop local instructionalplans for LEP students, moving away fromfavoring native-language instruction. How-ever, this does not exempt districts from com-pliance with federal law with regards to thesestudents. Additionally, current state boardpolicy does not include any requirements forassessment at either the state or local dis-trict levels. Along with providing districts theflexibility to design local programs to meetthe full educational needs of all LEP students,policymakers should include evaluation andaccountability requirements to ensure thatthose needs are being met.

Another policy consideration relates to thesupplementary funding of programs servingLEP students. The funding mechanisms cur-rently in place institute a perverse incentive

structure for school districts in both their clas-sification as LEP students and reclassificationof students as they achieve English-languageproficiency. The new California state boardpolicy does not change this situation.

While no specific pedagogy is pushed on Cali-fornia districts, district and school administra-tors still face financial incentives in their clas-sification and redesignation policies becauseof the additional resources garnered by hav-ing additional LEP students. A funding mecha-nism should be instituted that provides schoolswith the necessary resources to effectivelyserve this student population without relyingupon the classification and redesignation ofLEP students.

There is also legislation in Congress that wouldchange the way that federal Title VII fundswould be distributed. Grants should beawarded to instructional programs that effec-tively address the needs of LEP students, re-gardless of native-language use. Fundingmechanisms that continue to favor one teach-ing method over another limit the innovationin designing effective district programs.

The breadth of research surrounding the issueof how to instruct limited-English proficientstudents does not warrant concluding that anyone method of instruction is best. Consensusof research conclusions in favor of one meth-odology over another is unlikely given the di-verse nature of these students — languages,family and community backgrounds, prior edu-cation and English exposure.

What is consistent across the major evalua-tive studies is that those programs that arewell implemented, designed to meet the needsof the local student population, with clear goalsand objectives, with teachers and administra-tors in support of the designated program, andwith resources well utilized can succeed inmeeting the needs of these children. Thus,policies that facilitate diverse local programdevelopment and implementation, along withthorough evaluation should ultimately be pur-sued. v



Glossary of TGlossary of TGlossary of TGlossary of TGlossary of Termsermsermsermserms

Additive Bilingualism: English language development builds on native language develop-ment.

Bilingual education is a term that carries a tremendous weight and several connotations. Toavoid confusion, specific program names will be used to describe the effectiveness (or lackthereof) of programs implemented and evaluated. Otherwise, when referring to the generalform, the term special instructional services/programs for LEP students will be used.

Developmental Bilingual Education (DBE, also known as Late-Exit Bilingual Education): focuson developing academic proficiency and literacy in native language before making full transi-tion to all-English classes. Program length typically five to seven years.

English as a Second Language (ESL): often referred to as ESL pull-out programs where LEPstudents are given special English development instruction away from the mainstream classes.Pull-out varies from one-period to half-day instruction. Instructors are trained specialists.

Immersion: LEP students taught entire curriculum in English with specific teaching methodsemployed to overcome language barriers. Little to no native language usage in the instruc-tional program. An aide is sometimes present.

Language Majority Students: native English speakers.

Language Minority Students: non-native English speakers, limited-English proficient.

Structured Immersion (also known as Sheltered Immersion): LEP students provided segre-gated English instruction by certified instructor or aide. In some instances, students areallowed to ask content-specific questions in their native language, but instruction remains inEnglish. Also considered under ESL programming.

Submersion: no instructional support provided by trained specialists. Not considered aprogram model under Lau and Castaneda. Traditionally known as the “sink-or-swim” method.

Subtractive Bilingualism: English language development replaces native language develop-ment.

Transitional Bilingual Education (TBE, also known as Early-Exit Bilingual Education): aca-demic curriculum taught in student’s native language while they are learning English, gradu-ally transitioning into all-English classes. Program length typically two to three years. Fulldevelopment of native language not anticipated.

Two-Way Bilingual Education: language majority and minority students are schooled to-gether, each serving as peer teachers. Curriculum is taught in both English and designatedminority language. In 1997, there were 202 two-way programs in 22 states and the Districtof Columbia, with 184 of 202 operating in English and Spanish.

end

no

tes

1 Given the high drop out rates, low achievementon standardized test scores, high incidences ofpoverty, and limited formal schooling on the partsof parents and other family members.

2 U.S. Department of Education, Summary Reportof the Survey of the States’ Limited English Pro-ficient Students and Available Educational Pro-grams and Services 1994-95, (National Clear-inghouse for Bilingual Education, 1996).

3 California Department of Education, EducationalDemographics Unit, Language Census (LEP/FEP).Spring 1997; CBEDS (enrollment), October 1996.

4 National Assoc. for Bilingual Education website,www.nabe.org/BE.html, as of January 1998.

5 Lau v. Nichols, 414 U.S. 563 (1974).

6 Department of Health, Education, and Welfare,35 Federal Regulation 11595, 1970.

7 Castaneda v. Pickard, 648 F. 2d 989, 1981, UnitedStates Court of Appeals, Fifth Circuit.

8 Little Hoover Commission, A Chance to Succeed:Providing English Learners with Supportive Edu-cation, July 1993, p.17.

9 Improving America’s Schools Act (IASA), P.L. 103-382, 1994.

10 U.S. Department of Education, Summary Reportof the Survey of the States’ Limited English Pro-ficient Students and Available Educational Pro-grams and Services 1994-95 (National Clearing-house for Bilingual Education, 1996).

11 Public Law 103-382, Sec. 7116.

12 Little Hoover Commission, A Chance to Succeed:Providing English Learners with Supportive Edu-cation (July 1993), p.18.

13 Ibid., pp. 56-57.

14 Ibid., pp. 57-58.

15 Quiroz v. California State Board of Education,Sacramento Superior Court, February 27, 1998.

16 J.A. Cardenas, “Bilingual Education: Its Analy-sis, Needs and Cost,” Multicultural Education: AGeneration of Advocacy (New York: Simon andSchuster, 1995).

17 BW Associates, Meeting the Challenge of LanguageDiversity: An Evaluation of Programs for Pupilswith Limited Proficiency in English, February 1992.

18 Center for Applied Linguistics, Two-Way Bilin-gual Education: Students Learning Through TwoLanguages, shown on CAL’s webpage,www.cal.org.

19 National Center for Education Statistics, The Con-dition of Education 1997, U.S. Department of Edu-cation, pp. 60–61; California Department of Edu-cation, Educational Demographics Unit, 1995-6.(Event dropout rate: percentage of those stu-dents in grades 10-12 who were enrolled theprevious October, but were not enrolled and hadnot graduated the following October.)

20 National Center for Education Statistics, The Con-dition of Education 1997, U.S. Department of Edu-cation, p. 54.

21 R. Rumberger, “Dropping out of high school: Theinfluence of race, sex, and family background,”American Educational Research Journal, 20 (2),1983, pp. 199–220.

22 California Department of Education, EducationalDemographics Unit, Language Census. Spring1997; CBEDS (enrollment), October 1996.

23 Howard Fleischman and Paul Hopstock, Descrip-tive Study of Services to Limited English Profi-cient Students, Volume 1, Table IV-4, Develop-ment Associates, Inc., 1993.

24 Ibid., Table IV-3.

25 Little Hoover Commission, A Chance to Succeed:Providing English Learners with Supportive Edu-cation, July 1993, p.37.

26 Wayne Thomas and Virginia Colliera, School Ef-fectiveness for Language Minority Students,National Clearinghouse for Bilingual Education,December 1997, pp. 20-25.

27 Light, R.J. and Pillemer, D.B., Summing up: Thescience of reviewing research, Cambridge, MA:Harvard University Press (1984).

28 J. Ramirez, et al., Final Report: LongitudinalStudy of Structured Immersion Strategy, Early-Exit and Late-Exit Transitional Bilingual Educa-tion Programs for Language-Minority Children,Volume I, U.S. Department of Education, p. 145.

29 BW Associates, Meeting the Challenge of Lan-guage Diversity: An Evaluation of Programs forPupils with Limited Proficiency in English, Feb-ruary 1992, Executive Summary.

30 Mark H. Lopez and Marie T. Mora, Bilingual Edu-cation and Labor Market Earnings Among His-panics: Evidence Using High School and Be-yond, January 1998.

31 Wayne P. Thomas and Virginia Collier, SchoolEffectiveness for Language Minority Students,disseminated by National Clearinghouse for Bi-lingual Education, December 1997, p. 60.

32 J. Cummins, Negotiating Identities: Educationfor empowerment in a diverse society, Califor-nia Association for Bilingual Education, 1996.


Reason Public Policy Institute [RPPI]3415 S. Sepulveda Blvd., Suite 400 l Los Angeles, CA 90034

310-391-2245 l 310-391-4395 [fax]www.reason.org

bilingual education - reason foundation · director of education studies, ... and the impact of...

Documents