Anti-Bilingual Initiative
Bilingual Education
Impact of Unz
Initiative Text
LA Times Polls
Prop. 209
Responding to Unz
Sophistry 101
USC Poll

Anti-Bilingual Initiative

A Researcher's View of Unz

By Stephen Krashen 
University of Southern California

October 1997

"English Language Education for Children in Public Schools" is a proposal submitted to California voters by Ron Unz and Gloria Matta Tuchman. Several aspects of the Unz-Tuchman proposal are not in question at all: "Immigrant parents are eager to have their children acquire a good knowledge of English," English is an important language, and we certainly have an obligation and duty to provide children with English literacy. The rest of the proposal is extremely problematic. Its claims and proposed changes are completely unsupported. 

A Poor Job?
The proposal asserts that "the public schools of California currently do a poor job of educating immigrant children ..." 

This attack on public education is unjustified. In the last decade, the burden on California's school system has increased incredibly, and schools have done as well as could be expected with the resources they have: 

    1. California has the worst school libraries in the country. California is next to last in terms of books per student (White 1990) and dead last with respect to the number of librarians per student (Snyder and Hoffman 1995). 

    2. The number of limited English proficient (LEP) students in California has increased from about 600,000 in 1987 to about 1,300,000 today. 

    3. California's public libraries have declined in access and quality. McQuillan (1995) notes that book budgets in public libraries in California have been cut by 25 percent since 1989, and the number of hours public libraries are open has declined by 30 percent since 1987. Children's services have been hit the hardest by these cuts. 

    4. McQuillan (1995) also notes that the absolute number and percentage of school age children in poverty have increased dramatically in California. California ranks 41st out of 50 in the percentage of children living in poverty, and there was a 25 percent increase from 1989 to 1993 in the number of children in poverty in California. The National Assessment of Educational Progress (NAEP) reported that California ranks near the bottom of the country in the percentage of homes with more than 25 books. 

Despite these problems, reading scores in California have held steady (McQuillan 1995), and dropout rates among Hispanic students decreased from 1988 to 1992 (Little Hoover Commission 1993). 

Failed Programs?
"[California schools have been] wasting financial resources on costly experimental language programs whose failure over the last two decades is demonstrated by the high dropout rates and low English literacy levels of many immigrant children."

The "experimental" language program referred to is bilingual education. Empirical studies of bilingual education show that children in these programs, when they are set up correctly, do well, acquiring English literacy better than children in all-day English programs (Krashen and Biber 1988; Krashen 1996).<1>

One cannot blame bilingual education for the high dropout rate of Latino students. First, not all Hispanic youngsters are limited-English-proficient. In California, less than half are classified as LEP. Second, as noted above, of those who are limited in English, not all are in bilingual education. In California, about 50 percent of LEP children are in programs that have any kind of non-English language support, and 21 percent receive no special instructional services at all (Han, Baker, and Rodríguez 1997; Snyder and Hoffman 1996; Macías 1996). Third, well designed bilingual programs produce better academic English, which suggests that bilingual education is part of the cure, not the disease. 

Finally, 40 percent of all Chicano children live in poverty in the U.S., and economic factors have been linked to dropout rates (Rumberger 1991; Portes and Rumbaut 1996; Rumbaut 1997). Poor children have far fewer books in the home, in their school libraries, and in the public libraries in their neighborhoods, and go to schools that are often poorly funded, among other problems. 

"Sheltered English Immersion"
"Children who are English language learners shall be educated through sheltered English immersion during a temporary transition period not normally intended to exceed one year." 

The initiative defines sheltered English immersion as follows: "an English language acquisition process for young children in which nearly all classroom instruction is in English but with the curriculum and presentations designed for children who are learning the language." 

It defines bilingual education as follows: "a language acquisition process for students in which much or all instruction, textbooks, and teaching materials are in the child's native language." 

There are many problems with this plan: 

    1. "Sheltered English immersion" is not a term in current use in the language education profession but is a confusing combination of terms. "Sheltered subject-matter teaching" is a method used with intermediate, not beginning, second language students in which content is made comprehensible (Edwards, Krashen, Wesche, Clement, and Krudinier 1983; Krashen 1991). 

    Sheltered subject-matter teaching was inspired by the success of Canadian immersion programs, but the "immersion" aspect of these programs was not the factor we adopted. Rather, it was the fact that subject matter was presented in a comprehensible way. (The term immersion has been used in several different ways, some of them contradictory. See "Appendix: A Note on Immersion" below for more details.) 

    2. Note that structured English immersion, as defined above, allows some first language use, contrasted with bilingual education, in which "much or all" of instruction is in the first language. Since no bilingual education program is all in the first language, these two programs differ only in how much first language use is allowed. Just how much is too much? Who decides? Will there be language police to determine when a teacher is using too much of the child's first language?

    The "gradual exit" program used in El Paso and California uses both sheltered subject-matter teaching and instruction in the first language, with varying amounts depending on the level of the child. Its status under the Unz-Tuchman proposal is not clear. Sheltered subject-matter teaching is a valuable part of programs for LEP children. It is used as a transition between the first language component and the mainstream. 

    3. The proposal sets one year as the standard for special English instruction. After one year, it is expected that students will participate in mainstream classes. This criterion is wildly optimistic, one that Unz has noted is based on "common sense" and not research. Even those who are opposed to or are highly critical of bilingual education note that it takes more than one year to acquire academic language, the kind of English language competence children need to succeed in school (e.g., the heavily anti-bilingual education Little Hoover Commission report [1993] notes that "some experts believe that English can be academically comprehensible for children in as little as two years [emphasis mine], while others believe that six or more years of assistance is necessary"; p. 36).

    The evidence we have suggests that one year of intensive English is not enough to bring children to the level where they can do grade-level work. 

    Goldberg (1997) published in the READ Journal, a publication whose emphasis has always been anti-bilingual education describes an all-English program for LEP students. In kindergarten, there was no instruction at all in the primary language, and no English as a second language (ESL), but children "receive a language-rich curriculum [in English] based on thematic units" (p. 64). While 90 percent of the students showed some growth in English during the year, most still scored in the "beginner" range on the PreLAS in English, clearly nowhere near ready for a full academic program in English. 

    Kreuger and Townsend (1997) describe a program for LEP 1st graders in Quebec who were given a great deal of help in English literacy: small-group work (three to four students per group) for two hours daily devoted exclusively to literacy development, based on Reading Recovery. Nineteen of the 23 went on to grade 2, but the students were still well behind native speakers of English, scoring at the middle of grade 1 in reading at the end of the year, "still well below the class average (p. 127). These students, in addition, had already had a full year of kindergarten entirely in English, in a semi-sheltered situation 75 percent of the class consisted of second-language acquirers. 

    Thus focusing only on English in very young children has been tried. Focusing on the primary language works better, even for English language development. When the Carpenteria preschool program drastically reduced the amount of English and increased Spanish, focusing on language and cognitive development, dramatic gains were seen in school-readiness measures as well as in English! As the designers noted, English improved because the children were better able to understand the English they were exposed to outside of school, thanks to their improved cognitive development (Keating 1984).

    4. The proposal encourages grouping by English language proficiency alone. If "sheltered English immersion" is in fact sheltered subject-matter teaching, this means one class with a group of beginners in English, studying subject matter that can be at very different levels even if the program is limited to children younger than age 10 (as the Unz-Tuchman plan proposes). This is a nightmare for the teacher and will result in a great deal of dead time for children exposed to lessons that are completely inappropriate for them. 

    Children who "know English" are excused from sheltered English immersion. To demonstrate that they know English, they must score "at or above the state average for [their] grade level or above the 5th grade average, whichever is lower ... " on standardized tests of reading comprehension. This is statistical nonsense: 49 percent of the native speakers of English score below the state average, by definition. 

Why Bilingual Education?
Bilingual education is sensible. A LEP child who knows the subject matter will understand more in a class taught in English than one who does not. The more math the child knows, the easier it will be to understand a math class taught in English. The more the child understands, the faster he or she will acquire English. Bilingual education gives children subject-matter knowledge and thus speeds their English language development. Also, children who develop literacy in their primary language have a much easier time developing literacy in English. It is easier to learn to read in a language you already know, and once you can read, this ability transfers to English. Thus, in addition to making sure children do not fall behind in subject matter, bilingual education makes a powerful contribution to their language and literacy development. 

Published research supports bilingual education (Krashen 1996). Children in well designed programs, programs that give them sold subject-matter instruction in the primary language, literacy in the primary language, and ESL, acquire academic English as well and usually better than children in all-English programs. 

In addition, it is a shame not to promote primary or heritage language development. Children who develop their primary language in addition to English, not instead of show superior cognitive development in some areas, typically do better academically than their monolingual peers from the same group, and do better in the work world (see papers in Krashen, Tse, and McQuillan, in press). Our country and economy benefit from bilinguals.<2>

The best solution for our LEP children is a solid academic foundation in the first language, with ESL beginning immediately, and with subjects being taught in English as soon as they can be made comprehensible for the child. 

Appendix: A Note on Immersion
English for the Children explicitly supports "sheltered English immersion" and argues that "learning a language is much easier if the child is immersed in that language." As I noted in Krashen (1996), there is a great deal of confusion about this term. One definition of immersion is "submersion" or "sink or swim" (doing nothing), an approach that English for the Children does not support. 

A second use of the term refers to the program used in Canada for French as a foreign/second language development for English-speaking children. Canadian-type immersion is bilingual education. It satisfies the three requirements for proper bilingual education (Krashen 1996): (1) comprehensible input in the second language; (2) literacy development in the first language; and (3) subject-matter learning in the first language. In addition, because the vast majority of the children in these programs are middle class, they do a considerable amount of reading outside of school (Eagon and Cashion 1988). Much of the curriculum is in the first language (English), and the goal is bilingualism development of both languages. 

A third type of immersion is "structured immersion." Structured immersion uses the first language only minimally and includes direct teaching of grammar and preteaching of vocabulary (Gersten and Woodward 1985). The results of structured immersion research are not at all convincing. Gersten and Woodward (1985) report that children in structured immersion in Uvalde, Texas, reached the 30th percentile of the reading comprehension subtest of the Metropolitan Achievement Test at the end of grade 3. After leaving the program, however, they dropped to the 15th and 16th percentiles in grades 5 and 6 (Becker and Gersten 1982). While this performance was better than that of a comparison group, it is still dismal. In addition, we have no idea what the comparison group did. (Uvalde children did somewhat better on the WRAT reading test, which emphasizes "decoding skills.")

In a second study, Gersten (1985) claimed that more LEP children in a California school district in structured immersion performed at or above grade level than comparison children in bilingual education. There were several serious problems with this study. First, no details were provided about the bilingual education program. Second, the sample size was small (28 children in bilingual education, 16 in structured immersion). If performance of just a few children had varied slightly, the results of this study would have looked very different. Third, the study only followed children to grade 2. Gersten and Woodward reported that the structured immersion children did extremely well in a follow-up study, reaching the 65th and 78th percentiles two years later, but only two groups of nine children each were studied.

A fourth definition of immersion is sheltered subject-matter teaching, which as noted above, was inspired by Canadian-type immersion. These are classes, used in both second- and foreign-language teaching situations, in which intermediate-level language students get subject-matter instruction with the second language used as a medium of instruction. Native speakers and beginners are not included in these classes, in order to make sure that the input is comprehensible. Studies show that students in sheltered classes make at least as much progress in the second language as students in regular intermediate language classes and usually more and learn subject matter at the same time (Krashen 1991). In addition, they acquire more "academic language." 

As noted above, the gradual-exit program uses sheltered subject-matter teaching as an intermediate stage between classes taught in the first language and the mainstream. 


1. The Rossell and Baker (1996) critique of bilingual education research is not convincing. I reviewed their survey in Krashen (1996). Briefly, I found that those studies considered to be not in favor of bilingual education suffered from vague labeling, i.e., some successful programs labeled "immersion" were actually bilingual education, with a significant amount of first language use. And in many cases, unsuccessful programs labeled "bilingual education" were clearly bad programs in which the critical components of properly designed bilingual programs were not provided. Also, several studies interpreted as not supporting bilingual education were very short term not nearly long enough for bilingual education to show an effect. Interestingly, the vast majority of studies Rossell and Baker considered to be scientifically unacceptable support bilingual education. The flaws in these studies were not threats to the validity of the conclusions, in my view. 

2. Rumbaut (1995) examined the progress of over 15,000 high school students in San Diego from language minority groups. Predictably, those classified as LEP had lower grade-point averages and were more likely to drop out. What is very interesting, however, is that those classified as "fluent English proficient" (in other words, bilingual) had better grades and slightly lower dropout rates than those rated "English-only." This was the case even though parents of English-only students were of higher socioeconomic status than parents of the bilingual students. 

Rumbaut also found that parents' "ethnic resilience and reaffirmation" was a positive predictor of grade-point average. Those parents who felt that "their ethnic group must stay together as a community to preserve their own culture and identity even as they adapt to the U.S. economy to 'make a living,'" but were not planning to return to their homelands, had children who performed better in school (p. 51). As Rumbaut points out, this result provides support for the concept of "additive" rather than "subtractive" acculturation (p. 52). 


Becker, W., and Gersten, R. 1982. A Follow-Up of Follow Through: The Later Effects of the Direct Instruction Model on Children in the Fifth and Sixth Grades. American Educational Research Journal, 19: 75-92. 

Eagon, R., and Cashion, M. 1988. Second Year Report on a Longitudinal Study of Spontaneous Reading in English by Students in Early French Immersion Programs. Canadian Modern Language Review, 44: 523-526. 

Edwards, H., Wesche, M., Krashen, S., Clement, R., and Krudinier, B. 1984. Second Language Acquisition through Subject-Matter Learning: A Study of Sheltered Psychology Classes at the University of Ottawa. Canadian Modern Language Review, 41: 268-282. 

Gersten, R., and Woodward, J. 1985. A Case for Structured Immersion. Educational Leadership, 43: 75-79. 

Gersten, R. 1985. Structured Immersion for Language Minority Students: Results of a Longitudinal Evaluation. Educational Evaluation and Policy Analysis, 7: 187-196. 

Goldberg, A. 1997. Follow-Up Study on the Bethlehem, Pa., School District's English Acquisition Program. READ Journal, 4: 59-84. 

Han, M., Baker, D., and Rodríguez, C. 1997. A Profile of Policies and Practices for Limited English Proficient Students: Screening Methods, Program Support, and Teacher Training. Washington, D.C.: U.S. Department of Education, NCES 97-472. 

Keating, J. 1984. An Assessment of the Carpenteria Preschool Spanish Immersion Program. Teacher Education Quarterly, 11, 3: 80-94. 

Krashen, S. 1991. Sheltered Subject Matter Teaching. Cross Currents, 18: 183-188. 

Krashen, S. 1996. Under Attack: The Case Against Bilingual Education. Culver City, Calif.: Language Education Associates. 

Krashen, S., and Biber, D. 1988. On Course: Bilingual Education's Success in California. Los Angeles: California Association for Bilingual Education. 

Krashen, S., Tse, L., and McQuillan, J. (Eds.) In press. Heritage Language Development. Language Education Associates. 

Kreuger, E., and Townsend, N. 1997. Reading Clubs Boost Second-Language First Graders' Reading Achievement. Reading Teacher, 51: 122-127.

Little Hoover Commission, 1993. A Chance to Succeed: Providing English Learners with Supportive Education. Sacramento, Calif.

Macías, R. 1996. 1996 LEP Ed Services Little Changed from 1995. Linguistic Minority Research Institute News, 6, 1. 

Portes, A., and Rumbaut, R. 1996. Immigrant America. 2nd ed. Berkeley: University of California Press. 

Rossell, C., and Baker, K. 1996. Bilingual Education in Massachusetts: The Emperor Has No Clothes. Boston: Pioneer Institute for Public Policy Research. 

Rumbaut, R. 1995. The New Californians: Comparative Research Findings on the Educational Progress of Immigrant Children. In R. Rumbaut and W. Cornelius (Eds.), California's Immigrant Children. San Diego, Center for U.S.-Mexican Studies, pp. 17-69. 

Rumbaut, R. 1997. Passages to Adulthood: The Adaptation of Children of Immigrants in Southern California. Report to the Russell Sage Foundation.

Rumberger, R. 1991. Chicano Dropouts: A Review of Research and Policy Issues. In R. Valencia (Ed.), Chicano School Failure and Success. New York: Falmer Press. 

Snyder, T., and Hoffman, D. 1996. Digest of Educational Statistics. Washington, D.C.: U.S. Department of Education, National Center for Education Statistics. 

White, H. 1990. School Library Collections and Services: Ranking the States. School Library Media Quarterly, 19: 13-26.

Copyright © 1997 by Stephen Krashen. Permission is hereby granted to reproduce this page for free, noncommercial distribution, provided that credit is given and this notice is included.