Bilingual Education: A Failed Experiment on the Children (IP-6-1997)

June 1, 1997

Issue Paper

By Sheldon Richman

Executive Summary
[Note: The HTML/Web version of this document does not contain the many endnotes. A printed copy of the paper,
containing the endnotes, is available from the Independence Institute, at 303-279-6536, for eight dollars.]

• Bilingual education is based on the theory that the best way to make minority-language children proficient in English is to
first strengthen their skills in their native languages.
• A large body of research shows that native-language instruction is an inferior method of moving limited-English-proficient
children to full proficiency.
• Of the 18 million dollars spent on bilingual education in Colorado, 8 million is for illegal immigrants.
• In Denver, 80% of students in bilingual programs fail to make significant progress towards learning English even after two
years of bilingual education.
• Sixty percent of children who are forced into bilingual programs already have English as their dominant language.
• "Bilingual education" as it exists in most public schools is not really bilingual. Rather, students are instructed solely in one
language--typically Spanish. In fact, it is a gross perversion of the English language to label "bilingual" the forcing of
students into Spanish-only classrooms.
• There have been dozens and dozens of studies on the results of bilingual education. Hardly any of them show that bilingual
education is superior. Indeed, the weight of research finds bilingual education to be inferior to just putting students in classes
taught exclusively in English. Programs which aim to move students rapidly to full English proficiency (such as programs
which rely on native-language instruction for a year or less, and which include a good deal of English right from the start)
also outperform Spanish-only methods by a wide margin.
• Studies show that the longer a student stays in a segregated classroom, which does not use English as the primary means of
instruction, the worse his or her academic performance--even in subjects like math, where language is less important.
• The American Institutes for Research evaluated federally supported bilingual programs and found, startlingly, that only 16
percent of students spoke only Spanish. Eighty-six percent of project directors said that when children became functional in
English, they nevertheless stayed in the program.
• A 1996 survey of opinion was commissioned by the Center for Equal Opportunity asked six hundred Hispanics were asked
to rank five educational goals randomly given them.
•
◦ Fifty-one percent ranked as No. 1 "learning to read, write, and speak English."
◦ Only 11 percent put the highest priority on "learning to read, write and speak Spanish," and only 4.3 percent
selected "learning about Hispanic culture."
◦ When asked if children of Hispanic background should be taught Spanish before English or taught English as soon
as possible, 63 percent said that children should be taught English as soon as possible. Only 16.7 percent chose the
first option.
◦ In the final question, 81.3 percent of Hispanics agreed with this statement: "My child should be taught his/her
academic courses in English, because he/she will spend more time learning English." Only 12.2 percent agreed that
"My child should be taught his/her academic courses in Spanish, even if it means he/she will spend less time
learning English."
◦ Similarly, a 1988 survey by the Educational Testing Service for the U.S. Department of Education found that only
12 percent of Mexican-American parents wanted their children taught Spanish in school if it reduced the time for
English instruction. Only 11 percent of Asian-American parents wanted that. Moreover, less than one percent of
those parents said they want the schools to teach ethnic heritage.
• Bilingual education as is it currently exists is not really a program to help children learn English slowly. Rather, it is a
program designed by and run for the benefit of radicals who do not want children from Mexico, Puerto Rico, or other
 Spanish-speaking locales to assimilate and join the rest of American culture. Bilingual education is a subterfuge for
segregation, and therefore contrary to the spirit of Brown v. Board of Education, in which the United States Supreme Court
unanimously ruled that public schools could not deliberately segregate students by race or ethnicity.
• Bilingual education is a by-product of government monopoly schools. In a world of school choice and educational freedom,
hardly any families would choose it.

Policy Recommendations:

• The legislature and school boards should immediately adopt policies to prevent students from being forced into segregated
"bilingual" classrooms against their will.
• The legislature and school boards should adopt policies abolishing any "bilingual" program that continues longer than two
years.
• Any steps taken to increase educational freedom--including vouchers or bolder steps towards separation of school and state--
will help bring out the end of bilingual education as we know. The program exists its current size not because parents and
students want it, but because government school monopolists impose it.
Public schools in America have always been laboratories for social engineers. The children are the guinea pigs. The
people in the schools of education, departments of educational psychology, the teachers unions, and the bureaucracy
are the pseudoscientists who assure the parents of America that they know best while subjecting their children to
faddish theories and politically motivated practices.

This was true from the very beginning. Horace Mann launched the "common school" movement with the promise
that a national culture could be engineered by enlightened and well-meaning educational theorists who would put
the nations children through a scientifically conceived curriculum designed to create good citizens. The curriculum
was in no sense scientifically designed; the mantle of science was used to hide a well-meaning but nevertheless
politically motivated agenda. That approach to education necessarily interfered with the prerogatives of the family.
In other words, bureaucrats made decisions that should have been made by parents.

The history of public, or state, schools has been a series of similar episodes. Terms such as "life adjustment," "new
math," "whole language," "values clarification," and "family life" have become notorious shorthand for the social
engineering approach of the schools. Should "bilingual education" be on that list? It seems so.

What Is Bilingual Education?

Bilingual education is one of several approaches to the teaching of what are called limited-English-proficient (LEP)
children. Technically known as transitional bilingual education (TBE), this method is based on the theory that the
best way to make minority-language children proficient in English is to first strengthen their skills in their native
languages. In TBE, children learn to read and write in the native language, while theoretically getting an increasing
amount English instruction. The children are taught other subjects in the native language also. The ostensible
purpose is to keep the children from falling behind in those subjects while they are learning English. It is important
to understand that TBE is publicly defended as a way to ease the transition to English.

Christine Rossell of Boston University, who has studied bilingual education extensively, identifies three other basic
approaches. Besides TBE, Rossell lists

• "submersion," in which no special assistance is provided to LEP children, who, like earlier generations of immigrants, take
regular classes with English-speaking children;
• "English as a Second Language" (ESL), in which the children attend regular classes except for English instruction each day;
and
• "structured immersion," in which the children learn all subjects in English but in a special class for LEP children where the
teacher can take account of their English proficiency.
Note the difference between TBE and the others. In TBE, children are taught reading, writing, and other subjects in
their native languages. (Thus TBE programs are sometimes called "native-language instruction.") The teaching of
English occurs apart from other teaching. "The rationale underlying TBE differs depending on the age of the child,"
Rossell, a critic of TBE, writes. "For very young children, it is that learning to read in the native tongue first is a
necessary condition for optimal reading ability in the second language. For all children, it is argued that learning a
second language takes time, and children should not lose ground in other subjects, particularly math, during that
time period."

The University of Houstons Irma N. Guadarrama, a supporter of TBE, writes, "One of the most cogent arguments in
support of bilingual education is that children learn best when instruction proceeds from the known to the
unknown....Native-language-based instruction is aligned with learning theory, and as a crucial component of
bilingual education, contributes to the pedagogy that encompasses both the students culture and language as well as
the socialization and politics that result from the convergence of experiences."

In the other approaches, English is taught not only directly, but indirectly through its use in the teaching of standard
subjects.

Below we will examine the claims of success that proponents of the two broad approaches make. Preliminarily,
however, it should be noted that the question of what method is most effective in speeding the transition of limited-
English-proficient children to full proficiency is a pedagogical, as opposed to an ideological and political, matter. It
is ultimately to be decided empirically. Unfortunately, many participants in the debate forget that. Some advocates of
native-language instruction seem more concerned with cultural promotion and so-called group rights than with the
education of children.

Ultimately, the test of a teaching methods success or failure lies in the results, not whether it satisfies a theory. And
the test of the results lies in the objectives people wish to achieve. That brings us to a critical institutional question:
Who should determine the objectives of education? For the last 150 years in the United States, the major decisions
about a childs education have been made by government officials and their advisers from the education profession.
Conspicuously absent from the system have been the parents. That has been true in all aspects of education,
including in the matter of whether a child from a non-English-speaking home will be taught primarily in English or
in a foreign language.

What should not get lost in the argument over which teaching method is better and which objectives should be
served is the more important question: who should decide? The parents of Colorado and the other states have had
important decision-making power taken from them by "experts." In a free society that is intolerable. As we will see
later, parental freedom in education not only would restore lost liberty, it would also provide the best test of the
contending methods of teaching English to limited-proficient children.

Who Are the Limited-English Proficient Children?

Much of the attention to limited-English proficiency goes to Spanish-speaking children. But of course there are
many language minorities in the United States. According to the U.S. Department of Education, 73 percent of the
roughly 3 million LEP students speak Spanish. The next biggest groups were Vietnamese (3.9 percent), American
Indians (2.5 percent), Hmong (1.8 percent), Cantonese (1.7 percent), Cambodian (1.6 percent), and Korean (1.6
percent). Because of the relatively small number of non-Spanish-speaking LEP students and the difficulty in finding
teachers fluent in the other languages, most of those students are in instructed in English rather than their native
languages.

It is revealing, right off the bat, that Hispanic studentsthe only students consistently forced into TBE programshave a
lower rate of achievement and a higher drop-out rate than other groups, which are usually instructed in English. A
1988 study by the U.S. Department of Educations National Assessment of Educational Progress, for example, found
that Asian students performed better in reading and mathematics than their Hispanic (and white) classmates. It also
found that Hispanics competent in English did well on the tests, even if at home they spoke Spanish.

Federal Involvement in Bilingual Education

Bilingual education grew out of the civil rights movement. Concern about racial minorities quickly extended to
language minorities, particularly Hispanics, whose educational achievements were behind those of whites and blacks.
That disparity was interpreted as a violation by the states and localities of the civil rights of Spanish-speaking people.
Congress responded with two legislative acts:

• Title VI of the Civil Rights Act of 1964, which outlawed discrimination on the basis of race, color, or national origin in
federally assisted programs, and
• The Bilingual Education Act of 1968 (Title VII of the Elementary and Secondary Education Act), which created the first
federal bilingual education policy for language minority students and the mechanism for funding local school district
programs.
As a result, the Office for Civil Rights (OCR) and the Office of Bilingual Education and Minority Languages Affairs
(OBEMLA) were opened. Two bilingual education researchers, Christine H. Rossell and Keith Baker, point out that

bilingual education became an organizing principle for politically active Hispanics who considered themselves
uniquely excluded from the educational process by language and cultural problems not addressed in other programs.
Bilingual advocates argued that the reason for Hispanic children on average having lower achievement than white
children was the then-current practice of placing Spanish-speaking children in all-English regular classrooms and
"forcing" them to give up their native tongue.

The early legislation did not specify any particular method of teaching; native-language was not mandated. The law
merely authorized grants to local districts for research and experimental projects.

In 1974 the effort to involve the federal government in the issue of language minorities got a boost when the U.S.
Supreme Court ruled in Lau v. Nichols that for limited-English-proficient children, identical education was not equal
education. (The case was an interpretation of the Title VI of the Civil Rights Act.) The court said that a school district
that accepts federal money must take "affirmative steps" to neutralize the language barriers of students whose native
language is other than English. But the Court did not specify a remedy.

Nevertheless, Lau was critical in institutionalizing a systematic federal minority-language effort. The Office of Civil
Rights set up a task force to propose policies to carry out the Lau principles. "The OCR task force recommendations,
known as the Lau remedies, went well beyond the Courts requirement that school districts do something for LEP
children," Rossell and Baker write. "The task force insisted that transitional bilingual education was the best, if not the
only, instructional approach for providing equal educational opportunity to linguistic minorities." That policy
direction was chosen without public discussion.

From 1975 to 1980 OCR contracted with 500 school districts to set up TBE programs. Districts that wished to use
methods other than native-language instruction had to demonstrate the effectiveness of any alternative "even though
OCR itself had never proven that TBE was effective." After Lau, Congress passed the Equal Educational
Opportunities Act of 1974. In the same period, Congress amended Title VII to expand eligibility for the limited-
English-proficiency children and to permit the enrollment of some English-speaking children. The amendments
specified native-language instruction "to the extent necessary to allow a child to achieve competence in the English
language." Thus, transitional bilingual education, as the name implies, was primarily construed as a means to turn
limited-English-proficient children into fully English-proficient children. That fact will loom large when the results
of TBE are contrasted with those of English-oriented programs.

Shortly before the 1980 presidential election, the Carter administration made an aborted attempt to mandate bilingual
education in schools having a specified minimum percentage of minority-language children. When President Reagan
succeeded President Carter, the U.S. Department of Education published a survey of the research on bilingual
programs; the survey found no strong evidence for the efficacy of such programs. As a result, the federal
government backed off requiring local districts to use native-language instruction.

The 1984 Bilingual Education Act gave states and local districts even greater leeway, permitting up to 4 percent of
federal money (and in some cases 10 percent) to go to programs other than TBE. (Note the heavy tilt toward TBE
despite the loosening of the rules.)

More flexibility was permitted in the 1988 reauthorization. Twenty-five percent of funds could go to English-
language programs. "In addition, a three-year limit was placed on a students participation in a transitional bilingual
education program or in alternative instructional programs, although under special circumstances a student could
continue in a program for up to two additional years."

The Elementary and Secondary Education Act was reauthorized in 1994, dubbed the Improving Americas Schools
Act. The reauthorized Title VII gave this ringing endorsement to TBE:

the use of a child or youths native language and culture in classroom instruction can . . . promote self-esteem and
contribute to academic achievement and learning English by limited English proficient children and youth.
[ 7102(a)14A.]

The Act also declared that "the Federal Government. . . has a special and continuing obligation to ensure that States
and local school districts take appropriate action to provide equal educational opportunities to children and youth of
limited English proficiency." [ 7102(a)15.]

When the goals of Title VII of the 1994 federal law were listed, the development of English proficiency was placed
third, behind "developing systemic improvement and reform of educational programs serving limited English
proficient students through the development and implementation of exemplary bilingual education programs and
special alternative instruction programs" and "developing bilingual skills and multicultural understanding." [7102(b)
(1) and (2)] When English proficiency is finally specified it is combined with "developing . . . the native language
skills of such children and youth." [7102(b)(3)]

The critical section is 7116, which specifies the criteria under which federal grants to local districts will be made. The
act gives "priority to applications which provide for the development of bilingual proficiency both in English and
another language for all participating students." (Emphasis added) To emphasize that policy, the section continues
the 25 percent limit on alternative programs. The U.S. Secretary of Education may permit that limit to be exceeded
where the small number of limited-English proficiency students or the lack of qualified teachers makes TBE
impractical.

Thus, the weight of federal law is clearly behind native-language instruction, despite much rhetoric about local
flexibility to encourage proficiency in English.

What has this cost the American taxpayers? The American Legislative Exchange Council (ALEC) studied that
question for the years 1991 and 1992. Using data from the National Center for Education Statistics, the Office of
Bilingual Education and Minority Languages Affairs, and other government sources, ALEC found that of the more
than $200 million spent by the federal government in those years, only 20-30 percent went to English-language
programs. Of the nearly 2 million LEP students enrolled in some kind of special language program, only 22 percent
were in English as a Second Language programs. (Eighteen percent were in programs that the states could not
describe.) ALEC, conceding the difficulty of obtaining good statistics, estimated that in 1991-92, native-language
programs cost $5.5 billion overall; ESL cost nearly $2 billion; and "unknown" (undescribed) programs cost $2.4
billion, for a total of nearly $10 billion.

One caveat: the term "bilingual education" does not signify the percentage of time that native languages are to be
used in the classroom. The variation is great from class to class; it can depend on the teachers own philosophy and
proficiency in the native language. Such variation makes determining exactly how much "orthodox" TBE takes place
problematic. That might explain why so many people involved in bilingual education lament the lack of information
and accountability.
As a percentage of total spending, the federal contribution to bilingual education is small. Total federal spending on
education is only about 6 percent of all government spending on education. But what the federal government
chooses to fund is highly symbolic and sets the pace for the states and local districts. Former bilingual education
teacher Rosalie Pedalino Porter notes that even in states that have repealed TBE mandates, the state departments of
education often continue to pressure districts to maintain native-language instruction. Below is a table of Title VII
grants to all states since 1969.

Title VII Funding for States from FY69 to FY95

Year For Local For State For Total
Training
1969 101500 0 0 101500
1970 389290 0 0 389290
1971 756493 0 0 756493
1972 875522 0 0 875522
1973 1084803 0 0 1084803
1974 1464850 0 0 1464850
1975 549188 0 0 549188
1976 1089513 0 499000 1588513
1977 1803370 0 542386 2345756
1978 1919339 0 431895 2351234
1979 1561110 0 741803 2302913
1980 2076470 79727 610333 2766530
1981 1566243 93441 627414 2287098
1982 1609784 75217 671043 2356044
1983 900204 47931 537849 1485984
1984 748619 48791 411954 1209364
1985 1157585 50000 407729 1615314
1986 807502 54406 723959 1585867
1987 1228039 50000 952307 2230346
1988 1335636 51567 1107954 2495157
1989 1321364 75000 489259 1885623
1990 1175473 75000 384769 1635242
1991 1117454 75000 599742 1792196
1992 2694334 75000 1440765 4210099
1993 3171150 85009 1390287 4646446
1994 3215424 98391 1218693 4532508
1995 2540268 109260 1069080 3718608
Total 38260527 1143740 14858221 54262488
 1989 1321364 75000 489259 1885623
1990 1175473 75000 384769 1635242
1991 1117454 75000 599742 1792196
1992 2694334 75000 1440765 4210099
1993 3171150 85009 1390287 4646446
1994 3215424 98391 1218693 4532508
1995 2540268 109260 1069080 3718608
Total 38260527 1143740 14858221 54262488
(Source: National Clearinghouse for Bilingual Education, 1996)

Colorado Involvement in Bilingual Education

Colorados bilingual education law was passed in 1974 and was repealed in 1984. "The law mandating bilingual
education was replaced by the English Language Proficiency Act, which was intended to permit local districts more
flexibility in addressing limited-English students," said Roger Martinez, Colorados new Title VII coordinator.

Before 1984, English as a Second Language programs were not permitted if the number of LEP students surpassed a
specified threshold. Under the new law, local districts may use other methods if the state deems them effective.
According to Martinez, Colorado spends $2.8 million in state money on LEP programs, "mostly ESL." The
exception, he said, is in Denver Public Schools, which has the greatest concentration of LEP students. Most of those
students are in native-language programs.

The Colorado Department of Education reports that 13,129 students were involved in special language programs at a
cost of $2.6 million to the states general fund in the 1993-94 school year. The federal government puts in another
$2.8 million and the local school districts taxpayers kick in another $12.5 million.

A significant part of the expense of bilingual education in Colorado is created by bilingual instruction for illegal
immigrant children, as the table on the next page details. Of the 18 million dollars spent on bilingual education in
Colorado, 8 million is for illegal immigrants. To point out this fact is not to "bash" immigrants in general, or illegal
immigrants in particular. But it is notable when 44% of the spending for a particular government program goes to
people who are not supposed to be eligible for any programs funded by tax dollars taken from American citizens.

Illegal immigrant population of school age (5-17) 11,907

(Source: The Urban Institute and the US. Census Bureau)

Proportion of illegal immigrant children participation in school 83.98%

(Source: The Urban Institute)

Illegal immigrant children participating in school 9,999

Illegal immigrant children in bilingual programs 7,499

Average per-pupil cost for bilingual education programs $1,151

(Source: Colo. Dept. of Educ./ and local districts)

Total costs for illegal immigrant children in bilingual education programs $8,631,349
 Average per-pupil cost for bilingual education programs $1,151
(Source: Colo. Dept. of Educ./ and local districts)

Total costs for illegal immigrant children in bilingual education programs $8,631,349

The Results
Martinez laments that little is known about how well these programs serve the needs of students. "No one has kept a
running record" of success or failure in the districts, he said. He pointed out that federal grants are made directly to
local districts and that the state has no oversight authority.

Certainly the Denver Public Schools record on Hispanic students is not encouraging. The DPS has 31,000 Hispanics
students. The drop-out rate is nearly half, says Martinez. That high rate is consistent with rates across the nation.
According to the Census Bureau, 53 percent of Hispanic students graduated high school in 1994, versus 82 percent
of whites and 73 of blacks.

According to the National Clearinghouse for Bilingual Education, Colorado had over 26,000 limited-English
proficient K-12 students in 1993-94, the latest year for which it has figures. Those students constituted almost 4
percent of total enrollment.

Considering the amount of money spent on and the passion behind the nearly 30-year-old bilingual movement, one
would expect a volume of studies demonstrating the efficacy of the method. There appears to be no such
documentation, however. Dispassionate researchers have concluded that, on the contrary, native-language
instruction is an inferior method of moving limited-English-proficient children to full proficiency. Porter said that
this approach was started as an experimental project and that the supporting theories are after-the-fact
rationalizations. She has concluded from her own experience and from her examination of many such programs that
bilingual education is based on two "misleading propositions":

First, it is essentially dishonest to hold out the promise that development of native-language skills for several years
will lead to better learning of English. Second, pushing home-language-maintenance bilingual education programs
as a means of making disadvantaged, limited-English students "balanced bilinguals" draws classroom time and
resources away from the more urgent educational needs of these children. Fostering the notion that full bilingualism
can be easily achieved at no cost to the development of English or of subject matter learning is surely a deception
that effectively hardens class divisions at linguistic barriers.

Porter argues that neither in theory nor in the data can one find support for the view that the best way to teach
English to limited-proficient students is to spend several years teaching them their native languages (the so-called
vernacular argument). "Nothing in my fifteen years in this fieldfrom firsthand classroom experience to concentrated
researchhas begun to convince me that delaying instruction in English for several years will lead to better learning of
English and to a greater ability to study subject matter taught in English," she writes. "When all the rhetoric is
stripped away, the vernacular argument is still a hypothesis in search of legitimacy, and not a documented,
empirically proven, successful method of second-language learning."

In her book, Porter reproduced a startling table drawn from the survey of research performed by Rossell and Baker.
The table lists the percentage of "methodologically acceptable studies" that show the superiority, inferiority, or
equality of transitional bilingual education, English as a Second Language, submersion (essentially doing nothing
for LEP students), and structured immersion programs. The studies used reading, language, and math as indicators
of the success of the methods. In no case, did more studies show the superiority of TBE than showed its inferiority.
A portion of the chart is reproduced below.

TBE vs. Doing Nothing

 Reading Language Math

TBE Better 22% 7% 9%

No Difference 45% 29% 56%

TBE Worse 33% 64% 35%

# of studies 60 14 34

TBE vs. ESL

Reading Language Math

TBE Better 0% 0% 25%

No Difference 71% 67% 50%

TBE Worse 29% 33% 25%

# of studies 7 3 4

TBE vs. Structured Immersion

Reading Language Math

TBE Better 0% 0% 0%

No Difference 17% 100% 63%

TBE Worse 83% 0% 38%

# of studies 12 1 8
 TBE Better 0% 0% 0%

No Difference 17% 100% 63%

TBE Worse 83% 0% 38%

# of studies 12 1 8

Here is how Rossell and Baker summarize their findings:

After reviewing the results of these studies, we find no consistent research support for transitional bilingual education
as a superior instructional practice for improving the English language achievement of LEP children.

Elsewhere, Rossell summed up her research thus:

...substantially more studies show a harm from TBE, compared to all-English instruction, than show a benefit, and
this disparity increases when the all-English program is structured immersion. Thus, the risk of academic deficiency
in English is greater for TBE than for all-English instruction.

In other words, doing nothing tends to be less detrimental to children than native-language instruction. Rossell
suspects the results would be even worse for TBE except that "many bilingual education teachers are subverting the
theory. Rather than waiting until their students are proficient in reading and writing in their native tongue as the
theory advocates, they transition their students fairly quickly into English. Unfortunately, this cannot be said for all
of them."

Among the studies worth noting is one conducted by the El Paso, Texas, Bilingual Immersion Project of 2,500
Spanish-speaking students from grade one to three. The students, all from similar economic backgrounds, were
divided into native-language and English-immersion programs. The results were stark. Those in the immersion
program performed better than those in TBE on standardized tests in reading and language. They did as well in
math, and were above average in science and social studies.

Another indication of the inferiority of native-language instruction, according to researchers, is the exit rate from the
various programs. The studies tend to show that students in English-oriented programs are "mainstreamed" sooner
than students in native-language programs. For example, a New York City Board of Education longitudinal study
from 1990-94 found that of students who entered ESL programs in kindergarten, nearly 80 percent exited in three
years, versus only 52 percent of students in bilingual classes. Of those who entered ESL in second grade, 68 percent
left in three years, versus only 22 percent of those in bilingual programs.

In his introduction to the study, Chancellor Ramon C. Cortines wrote that it "showed that ESL-only students tested
out of LEP services more quickly and that in the short term, those in ESL-only programs appeared to have better
outcomes." Such results are reproduced in many places.

Writes Porter:

In both the El Paso and New York City longitudinal studies . . . students in the English language, structured
immersion programs met program goals in three to four years and were assigned to regular classrooms without
special help, while students in the traditional bilingual classrooms needed six to seven years to reach the same level
of mainstreaming.

Those studies undermine the theoretical claim that students will best learn English by first learning their native
languages and then transferring those skills to English afterward. At least one advocate of bilingual education has
trouble accepting that claim. The educational psychologist Kenji Hakuta has said, "What is remarkable about the
issue of the transfer of skills is that despite its fundamental importance, almost no empirical studies have been
conducted to understand the characteristics or even to demonstrate the existence of transfer of skills."

The proponents of transitional bilingual education tout studies purporting to show the success of the program. The
U.S. Department of Education points to one (done by D. Ramirez et al.) that seems to show that the longer students
stay with their native languages, the better they perform academically. That result is said to confirm the "facilitation
theory," which states that success in a second language is facilitated by high proficiency in the native language. The
study compared the achievement of students in English immersion, early-exit native-language, and late-exit native-
language programs. The Department summed up the study by stating:

Students in all three bilingual education programs realized growth in English language and reading skills that was as
fast or faster than the norming population. A higher percentage of late-exit students (about one-third) are reclassified
from LEP to fully English proficient (FEP) than are students in either immersion strategy (22 percent) or early-exit
(19 percent) programs.

Rossell and Baker, however, express doubts about the study. Although they concede that it may demonstrate that
bilingual education may be better than an all-English program when a student knows no English at all, they argue
that it does not follow that it is better once a student knows some English. At that point, they say, maximum time
("time on task") on English instruction is critical. They go on to write:

Ramirez et al., on the other hand, shows no support for the facilitation effect. Contrary to the exaggerated claims
made, the descriptive portion of the study shows that students who stayed in bilingual education the longest did the
worst. While this apparently negative finding for bilingual education is not reliable given the lack of statistical
control for student and classroom characteristics, it is definitely not positive evidence.

Porter notes that another problem with the study is that the three programs compared are not as distinct as they
appear to be. She quotes the National Research Councils comment:

Although the studys final report claims that the three programs represent three distinct instructional models, the
findings indicate that the programs were not that distinct. They were essentially different variations of the same
treatment: immersion and early-exit programs were in some instances indistinguishable from one another . . . . The
final designs of the longitudinal and immersion studies were ill-suited to answer the important policy questions that
appear to have motivated them.

Another indication of the inefficacy or harm of bilingual education is to be found in the words of its own defenders.
Rossell and Baker quote several, including Kenji Hakuta:

There is a sober truth that even the ardent advocate of bilingual education would not deny. Evaluation studies of the
effectiveness of bilingual education in improving either English or math scores have not been overwhelmingly in
favor of bilingual education. . . . An awkward tension blankets the lack of empirical demonstration of the success of
bilingual education programs. Someone promised the bacon, but its not there.

Another bilingual proponent, Thomas Carter, said in 1986, "Regardless of the many roots of the debate, one issue is
unresolved. Does bilingual education work?" Linguist Christina Bratt Paulson, also an advocate for bilingual
education, has said, "The rationale for bilingual programs is that they are more efficient in teaching English although
there is not much hard data to support such a view." Bilingual advocate Pertii Toukomaa wrote:

We wish to dissociate ourselves from those arguments, for teaching in the mother tongue, which attempt to frighten
parents into choosing mother tongue-teaching by threatening emotional and intellectual under-development in those
children who do not receive mother tongue-teaching. Teaching in the mother tongue does not seem to have a
magical effect on the childs development, for good or for ill, which it has sometimes been ascribed.

Rossell and Baker also quote the Association for Supervision and Curriculum Development (90,000 principals,
school superintendents, teachers, and other educationists) on its consideration of the respective merits of English
versus native-language instruction: "It is unclear which approach is better." Finally, a state commission in
Massachusetts, which had the first bilingual mandate in the nation, in 1994 confessed that it did "not know, on the
basis of measured outcomes, whether TBE programs in Massachusetts produce good results or poor results."
Bilingual Educations Failure in Denver
While it is true that Denver Public Schools (DPS) have not conducted an in-depth assessment of their TBE program,
Colorado law does give us a unique mechanism for looking at the program. The state of Colorado provides two
years of bilingual funding to its school districts through the English Language Proficiency Act. Section 22-24-106
(2) of this law requires all districts accepting state funds to provide a report to the general assembly assessing the
programs' performance. After examining this report, the only possible conclusion is that the DPS program has been a
failure.

Each year, students in the TBE program were given the Language Aptitude Survey (LAS), which measures the
students English reading levels. Each student is assigned a number from one to five, one being the lowest reading
level and five meaning the student no longer qualifies for bilingual funds. By tracking the students performance
over two years, it is clear that most students are not benefiting from the program.

Students were grouped into four categories: those who improved from levels 1-3 to 4, those who improved but
stayed within the 1-3 range, those who stayed the same, and those who regressed. In year one, 16% of DPS students,
regardless of age, were able to reach level 4. Unfortunately, 21% improved within the 1-3 range, while a full 57% of
students failed to show any improvement. Six percent of students regressed in the first year. In all, 84% of students
failed to reach level 4, which for many would still have qualified them for state bilingual funds. The second year
doesnt get any better.

Year two saw an improvement in those students who reached level 4, up from 16% to 20%. However, the number
who regressed also increased from 6% to 11%. Twenty-one percent of students improved within the 1-3 range, while
a staggering 48% of students still showed no improvement over two years of funding. In other words, 80% of
students in Denver showed no significant improvement after two years.

By any measure, this program must be seen as a failure. To the credit of the Denver School Board, several reforms
are being implemented, to a