Surveys of opinions on bilingual education: Some current issues

Surveys of opinions on bilingual education: Some current issues

Stephen Krashen


Krashen (1996) and Rossell and Baker (1996) have reviewed literature on attitudes toward bilingual education and have come to very different conclusions. This study reviews surveys included in Rossell and Baker as well as studies missed by both surveys and concludes that polls in which there is opposition to bilingual education typically present an extreme view of bilingual education that few of its supporters would endorse, with all subjects taught in the primary language and nothing in the second language. When subjects are asked about using both languages or are asked about bilingual education in general, responses are much more positive.


In the recent past, two surveys of public opinion on bilingual education have been published with contradictory conclusions. Krashen (1996) claimed that respondents generally approved of bilingual education and agreed with the principles underlying it, while Rossell and Baker (1996) argue that the results of polls overestimate support for bilingual education. Interestingly, there was no overlap in the studies covered by the two surveys.

This article reviews all available studies discussed in Rossell and Baker as well as several studies that both surveys missed, the 1983 Houston Metropolitan Survey, the Harris Poll (1993), de la Garza, DeSipio, Garcia, Garcia, and Falcon (1992), Krus and Brazelton (1983) and a recent report from the Center for Equal Opportunity (1996).

The first section of this paper reviews polls taken of the general public, the main focus of Rossell and Baker’s survey. Rossell and Baker have questioned the validity of these polls. The next section responds to their arguments. We then review studies of teachers and parents as well as Rossell and Baker’s comments on these polls.

Opinion Polls of the General Public

Krashen’s survey, even though it was titled “Is the public against bilingual education?” focused on surveys of parents and teachers. The only exception was Hosch (1984), in which there was clear support for bilingual education. While Hosch’s sample consisted of “40 individuals with a wide variety of backgrounds” (p. 19), nearly 25% of the sample had children who were in or had been in bilingual programs, and nearly 40% were Mexican-American or Mexican.

Table I presents the results of polls that attempted to get a representative sample, in which respondents were asked, in slightly different ways, whether they supported bilingual education.

It is important to examine each study in detail. For each study, we list the questions asked, as well as information about the sample.

Krus and Brazelton (1983)

In this study, “Students enrolled in an advanced class on theory of psychological measurement administered the questionnaire to their friends and members of their immediate families” (p. 249). Forty two subjects were interviewed.

The question represented in Table I was: “Does bilingual education (a) ultimately help, (b) ultimately harm Hispanic children?” In addition, the following statements were also presented to subjects:

“Bilingual education provides minority children with transferable skills which will allow them to be integrated into the dominant society.” Fifty-seven percent (57%) of the sample agreed with this statement.

“Paying for bilingual education with taxpayers money is wrong.” Sixty-six percent (66%) of the respondents disagreed with this statement.

“Placement of Hispanic children in an educational program in which they are taught in the Spanish language will prevent them from going beyond the twelfth grade, as they will not have the English skills necessary for college.” Twenty– eight percent (28%) of the respondents agreed with this statement.

(See also Krus and Stanley, 1985, for a comparison of this sample with a group of “persons identified as directly involved with the Bilingual and Multicultural educational program at Arizona State University” [p. 694]; for this pro-bilingual education sample, 100% supported bilingual education. For the additional questions discussed just above, 78% agreed that bilingual education provides transferable skills, none felt that using taxpayers’ money for bilingual education was wrong, and 20% felt that education through Spanish would prevent students from going beyond grade 12.)

20th Gallup Poll (1988)

The question was “Would you favor or oppose the local public schools’ providing instruction in a student’s native language, whatever it is, in order to help him or her become a more successful learner?” The sample (n = 2,118) was “designed to produce an approximation of the adult civilian population, age 18 and older, living in the U.S …” (p. 45).

Media General (1985)

This survey asked, “Do you think non-English-speaking students should be taught basic subjects in their own language while they learn English, or should they be placed in all English-speaking classes?” Their sample was ” … a representative sample of 1,462 adults across the nation living in telephone households” including listed and non-listed numbers. “The data projects to an estimated 161 million adults in telephone households” (information provided by Stephen Shaw, Director of Research, Media General, December 11, 1996).

Note that respondents in the Media General survey were asked to comment on a version of bilingual education which could be interpreted as teaching all subjects in the primary language until English is acquired. As discussed below, this may not be the best version of bilingual education.

Hakuta (1984)

The researchers asked “Do you think that bilingual education program is the best way for a Spanish-speaking child to learn English?” Additional details about this study are provided below.

Houston Metropolitan Area Survey (1983)

The researchers interviewed 1000 randomly selected residents in the Houston area. It could be argued that this survey overestimated support for bilingual education; respondents were asked whether bilingual education should be available in the public schools. A “no” response could indicate that the respondent felt that bilingual education should be illegal, an extreme position that most critics of bilingual education would disagree with.

In Huddy and Sears (1991) subjects were asked “How do you feel about bilingual education?”. The sample consisted of 1170 “Anglos,” which was supplemented by 100 “Non-Hispanic households” in areas of the country under represented in the first sample (n = 400). On a scale of -10 to +10, respondents’ mean support for bilingual education was +2.29.

Gallup 23 asked 995 adults:

“Bilingual education programs teach children who do not speak English basic subjects such as math and science in their native language, while also teaching them to speak English. Some people feel these bilingual programs should only be used until the child learns English. Others feel bilingual education should continue to be used in order to maintain the native language of children. Which opinion comes closer to your view?” (I use Rossell and Baker’s citation for this study as well as their description. I was unable to find it in the 1991 or 1992 Gallup Poll results as published in the Phi Delta Kappan.) Until a child learns English = 54%. Maintain native language = 37%.

In two other polls, subjects were given a choice of different options, with bilingual education as one of the options. This kind of poll can underestimate support for bilingual education; this was clearly the case in one such poll, the 1993 Gallup Poll (#25).

In Gallup 25 (Elam, Rose, & Gallup, 1993), only 27% of those interviewed (total n = 1306) appeared to support bilingual education, but subjects were asked to choose among three alternatives. Note that the option dealing with bilingual education asks whether all subjects should be taught in the students’ primary language. The following question was asked:

Many families who come from other countries have school-age children who cannot speak English. Which one of the following three approaches do you think is the best way for the public schools to deal with non-English-speaking students?

The Time Magazine Poll (1995) does not give a clear picture. The question asked was:

Which of these statements is closest to your views on bilingual education?

The first option could be chosen by advocates of bilingual education if it does not exclude the use of other languages. But if one chooses that option, one cannot choose the others. It is not clear whether those who reject number two think that transitional bilingual education is too little or too much. Finally, those who rejected the other options might have felt that they entailed teaching the entire curriculum in the first language.

Questioning the Validity of Polls

Rossell and Baker note that respondents in many polls support bilingual education, but question the validity of these results for two reasons: We refer to these arguments here as the Ignorance Argument and the Trade-Off Argument.

The Ignorance Argument

Rossell and Baker argue that many respondents do not understand what bilingual education is. They note, for example, that in Huddy and Sears (1990),

“only 22 percent … of 1,170 non-Hispanic adults were able to give a roughly accurate description of bilingual education. Almost 40 percent described it as bilingualism or foreign language instruction, and 29 percent were unable to give any description at all. Despite the fact that three-quarters of respondents could not accurately describe bilingual education, a majority supported bilingual education” (p. 171).

Huddy and Sears'(1990) classifications along with the percentage who supplied each definition from the sample of 1,170 is presented in the box below:

Even if we accept Huddy and Sear’s classification (it can be argued that “teaching English to foreign students” is too vague to categorize as accurate or inaccurate), what is crucial is that nearly all groups of subjects were mildly positive about bilingual education. The only negative subgroup was the one that defined bilingual education as “teaching foreign students in their own language,” which was very close to neutral, and consisted of a small subgroup of the sample:

They note that about 40% said they didn’t know if bilingual education was successful, but only 18% “lacked an opinion on whether non-English-speaking students should learn in their native language in school. In other words, it appears that some of the respondents who had no opinion on the efficacy of bilingual education were nevertheless willing to express a preference” (p. 166).

There is a simple explanation: Those who did not know whether bilingual education was successful or not probably did not support it, nor did those who felt it was not successful; those who felt it was successful did support it. In other words, those who were in the “don’t know” column for the first two questions probably migrated to the “no” column for the third question. If this is true, results of this poll underestimate potential support for bilingual education; with more knowledge, more might have supported it.

In support of this analysis, note the close correspondence in the “yes” column among the three questions; the percentage who felt that bilingual education was successful in teaching basic subjects was not significantly different from the percentage that supported bilingual education, while the difference between the percentage who felt bilingual education was successful in teaching English was just barely significantly larger than the percentage supporting bilingual education:

The Trade-Off Argument

Rossell and Baker argue that polls overestimate the support for bilingual education because they do not always ask “the trade-off question”: They do not ask whether the respondent would support bilingual education if it meant fewer resources or less time devoted to other things, maintaining that “more of something usually means less of something else” (p. 163).

Rossell and Baker support this point by noting that in several studies respondents appear to agree that the aim of bilingual education should be to encourage students to enter English-only classes as soon as possible, but also agree that programs should aim to maintain the Spanish language and culture. As an example of this, Rossell and Baker discuss Hakuta (1984):

In Hakuta’s 1984 survey of 216 adults in New Haven (37 of them Spanish-speaking), 76 percent of the respondents agreed that the emphasis of bilingual education should be to encourage students to enter English-only classes as soon as possible, yet 58 percent agreed that the emphasis should be on maintaining the Spanish language and culture of the children … it is clear that unless respondents are explicitly asked to consider the trade-off, they will not. In this case they should have been asked whether they wanted to emphasize maintaining the Spanish language even if it meant that a student’s entry into an English-only classroom would not occur quickly (p. 171).

Rossell and Baker assume that these goals are contradictory, that more time devoted to Spanish will necessarily mean less progress in developing academic English.

The point of bilingual education, however, is that more education in the first language, when done correctly, can mean faster development of English literacy: When students learn to read in their primary language, it is much easier to learn to read in English, and because the first language will be more comprehensible, it is easier to learn to read in the first language. In addition, background knowledge gained through the first language can greatly enhance comprehension of material presented in the second language. There is no trade-off. Respondents who want both English and Spanish literacy are not contradicting themselves.

Rossell and Baker’s characterization of Hakuta’s study is slightly inaccurate and incomplete. The total number of subjects surveyed was 216, but they were divided into two groups – a general sample and a Spanish-speaking sample. The data summarized by Rossell and Baker was from the general sample only. In addition, both groups were very positive about bilingual education, which was not mentioned by Rossell and Baker. Here is some detail from the study:

An Additional Teacher Study: The 1993 Harris Poll

The Harris poll was a national survey of teachers, 97% of whom were not Hispanic. Responses to a question dealing with bilingual education appear to be non-supportive:

Do you think government policy should promote bilingual education programs that teach English and teach other substantive subjects in a child’s native language, or should policy mandate that substantive subjects be taught in English?

Note that respondents could be rejecting a version of bilingual education in which all subjects are taught in the first language, with nothing ever taught in English. In one plan, the gradual exit plan (Krashen, 1996), subject matter is taught in the primary language only until the child knows enough English to follow instruction in English. “Transition” occurs gradually, a few subjects at a time, as they become comprehensible. The plan also includes sheltered subject matter teaching in English as a transition between all primary language and the mainstream. It is quite possible that many of those who felt that “substantive subjects should be taught in English” would have agreed with this kind of approach.

Additional Studies of Parents de la Garza, DeSipio, Garcia, Garcia, and Falcon (1992) was not covered either by Rossell and Baker or by Krashen. de la Garza et al. surveyed adults whose national origins were Mexican, Puerto Rican, and Cuban. The items discussed here were answered by a subgroup of the sample, citizens of the United States (Mexican origin, n = 878; Puerto Rican = 587; Cuban = 312). Attitudes toward bilingual education were very positive:

Are Parents Confused? A Discussion of the ETS Study

The ETS survey (Baratz-Snowden, Rock, Pollack, & Wilder, 1988) of parents of language minority students shows wide support for bilingual education. Rossell and Baker, however, claim that the report contains “fascinating inconsistencies in parental support for various options” (p. 172), and these inconsistencies suggest that respondents were confused about bilingual education.

Parents were asked whether they thought maintenance bilingual, transitional or immersion programs were a good idea. As Rossell and Baker note, all three programs were supported, which prompted Rossell and Baker to conclude that “regardless of the type (of program),…an overwhelming majority of parents think language minority children should get some kind of help, particularly in learning English, and they are not clear about differences between types of help” (p. 177). It is important to point out, however, that parents were not asked to choose among these programs: one-third of the sample was asked about maintenance bilingual education, one-third about transitional, and onethird about immersion. Agreeing that a certain program would be a good idea for helping students who don’t speak English might simply mean that respondents thought that doing the program would be better than doing nothing (submersion). Supporters of bilingual education would probably respond positively to all three options, when compared to submersion (sink or swim).

Rossell and Baker claim that “most of these parents do not think native language proficiency should be taught in school” (p. 174). They base this statement partly on responses to the following question: (If child does not speak or understand English very well) would it help the child if classes were taught using the non-English language? Twentynine percent of respondents categorizing themselves as Mexican– Americans responded positively as did twenty-eight percent of the Asian respondents.

Rossell and Baker note that this question was not actually asked in this way, but is a result of their reanalysis of the data in the survey. After reviewing Baratz-Snowden et al. in some detail, I was unable to discover how Rossell and Baker arrived at these figures. In addition, Rossell and Baker claim that “almost half of Mexican-American and 60 percent of Asian parents think that teaching in the native tongue interferes with English.” The full table of results, from table 19 of the survey, is presented below:

Do you think teaching in non-English language interferes with learning English?

Clearly, quite a few parents do not think that teaching in the primary language interferes with acquiring English (80% of the Cuban parents and 62% of one of the Puerto Rican samples.) Nevertheless, quite a few do; it is possible that this question was interpreted as referring to programs in which all teaching is in the primary language.

Baratz-Snowden et al. also asked parents “In what language should non-English students be taught?” for reading and writing and for basic subjects. Their results, expressed in percentages, are presented below:

This table shows clear support for bilingual education, especially when used for developing literacy. Rossell and Baker argue that it is another indication of confusion: Because 70% of Mexican-American parents preferred that reading and writing be taught in both languages, and 43% thought that teaching in the non-English language interferes with learning English (see above), then 27% of them must be confused. There is, however, an easy explanation: As noted earlier, some parents might have interpreted the question about teaching in the first language interfering with English as meaning teaching exclusively in the first language, an option that very few people support.

Rossell and Baker’s next “confusion argument” is based on responses to this question: “Do you think the schools should teach non– English language speaking children the non-English language if it means less time for teaching them (English, Math, Science, Art, Music)?” Parents responded negatively to all of these options. Many supporters of bilingual education would probably also respond negatively to this question, as discussed earlier. The point is that good bilingual programs do not force students to sacrifice in this way. Unfortunately, the wording of the question presupposes that instruction in the primary language inevitably means less time for other things, and suggests less English language development.

Rossell and Baker’s final “confusion argument” is based on responses to an open question in which parents were asked the three most important things they wanted children to learn from school. The most popular answers were mastery of academic subjects, English language competence, and “general education,” while “learn about child’s ethnic heritage” and “learn both languages” were not mentioned often. They conclude from these results that “there is, for all practical purposes, no desire on the part of…parents for the schools to teach ethnic heritage” (p. 181). The survey results are, however, reasonable. The point is that the proper use of the primary language in bilingual education programs leads to the attainment of the goals of mastery of academic subjects and English language competence. Note also that the low rank of “ethnic heritage” and “learn both languages” does not reflect lack of concern with these goals. Subjects were asked the three most important things the parents wanted their children to learn; it is unlikely that heritage culture and heritage language development would be in the top three.

Center for Equal Opportunity (1996)

The CEO survey included 600 randomly selected Hispanic parents with children currently in school. About 80% of the sample said that their children were not in “a program in school for children who need help with English” (p. 10) and the same percentage said their children were never asked to be in such a program. The programs the children participated in were hardly industrial-strength first language programs: Parents reported that about 10% had no Spanish at all, 23% were only a “small part” in Spanish, 29% were half in Spanish, and 27% were “most in Spanish.” Of the entire sample of 600 parents, 93 had children who were in or who had been in a special program, and of those 93, only 25 of them were in programs conducted mostly in Spanish, or 4% of the total. This is contrary to the view of some critics, who claim that Hispanic children throughout the US are taught in Spanish-only programs.

Of those whose children had been in such a program 74.5% reported that their children were in the program three years or less. While this latter figure underestimates the length of time a typical child participates, because it includes children still in the program, it suggests that children do not stay in special programs very long, also contrary to the claims of critics.

Not surprisingly, parents were pro-English: 51% rated “learning to read, write and speak English” as the most important thing children might learn in school and 19% rated it second. Many supporters of bilingual education would, of course, also rate this goal very highly.

The controversial questions were these:

(1) In your opinion, should children of Hispanic background, living in the United States, be taught to read and write Spanish before they are taught English, or should they be taught English as soon as possible?

English as soon as possible = 63%

Spanish before English = 17%

Same time = 17%

This is a difficult question to answer. One would expect parents to respond that children should be taught English as soon as possible. Bilingual education, it has been argued, is the best way to make this happen. The way the question is phrased, it suggests that learning to read and write in Spanish will not help children to learn to read and write English “as soon as possible.”

Shin asked a similar question, but did so more precisely, asking whether respondents felt that “developing literacy through the first language facilitates literacy development in English.” The language minority parents she questioned supported this position:

Hispanic parents = 53% (Shin & Gribbons, 1996)

Korean parents = 88% (Shin & Kim, in press)

Hmong parents = 52% (Shin & Lee, 1996)

Shin’s question does not presuppose that learning to read in the first language slows down the acquisition of English literacy; CEO’s version “In general, which of the following comes closest to your ion?

1. My child should be taught his/her academic courses in Spanish, even if it means he/she will spend less time learning English.

2. My child should be taught his/her academic courses in English, because he/she will spend more time learning English.

Spanish = 12%

English = 81%

As noted previously, such questions are flawed: First, they give the impression that in option 1, all courses will be taught in Spanish with nothing in English. Second, they suggest that learning content through Spanish will not help English language development. It is thus no surprise that respondents vote for English.

Shin also attempted to ask this question more precisely, asking whether parents agreed that “Learning subject matter through the first language helps make subject matter study in English more comprehensible.” Support was not as strong as it was for the question on literacy, but Shin found more support for first language content teaching than CEO did:

Hispanic parents = 34% (33% were “not sure”) (Shin & Gribbons, 1996)

Korean parents = 47% (Shin & Kim, in press)

Hmong parents = 60% (Shin & Lee, 1996)

The Effect of Political Orientation

Some of the opposition to bilingual education is not, it appears, based on whether it is good pedagogy. As Huddy and Sears noted, “There was considerable anti-Hispanic sentiment with the sample” (p. 128), and reported, not surprisingly, that indicators of this kind of attitude were predictive of objection to bilingual education, as shown by the multiple regression analysis presented below:

In addition, in the Media General report, 45% of those who classified themselves as “liberal” supported bilingual education, while only 33% of those who classified themselves as “conservative” did so (36% of those who were neither liberal nor conservative supported bilingual education).

It must be emphasized that anti-Hispanic sentiment and political orientation did not explain a large percentage of the respondents’ attitudes toward bilingual education (note that r^sup 2^ = .259 for all predictors combined in Huddy and Sears’ study).

Summary and Conclusions

The results of the polls discussed here are summarized in Table 2 on the following page.

Responses are clearly more negative when statements and questions can easily be interpreted as supporting an extreme version of bilingual education in which only the primary language is used (Harris Poll, Baratz-Snowden et al., Time, CEO) or one in which all subject matter is taught in the primary language until English is acquired (Media-General, Gallup Poll #25).

When subjects are asked about using both languages (Baratz– Snowden et al.) or are asked about bilingual education globally, they are much more positive (Krus and Brazelton, Hakuta, Huddy and Sears, Houston, de la Garza et al.). Baratz-Snowden et al. provides evidence that this generalization is correct: When it is made clear that both English and the primary language are to be included, subjects are supportive of bilingual education; when this is not clear, they react differently. This is a more plausible explanation than Rossell and Baker’s, who maintain only that respondents were confused. The only exception to this generalization is Gallup #23, which was supportive despite the vagueness of the question.

These results converge with those of Krashen (1996) who reviewed studies of parents, teachers and school administrators, and found consistent support for bilingual education when the question was asked globally. Polls that appear to present counterevidence typically present a view of bilingual education that few of its supporters would endorse.


Baratz-Snowden, J., Rock, D., Pollack, J., & Wilder, G. (1988). Parent preference study. Princeton, NJ: Educational Testing Service. Center for Equal Opportunity (1996). The importance of learning English. Washington: Center for Equal Opportunity. de la Garza, R., DeSipio, L., Garcia, F. C., Garcia, J., & Falcon, A. (1992). Latino voices: Mexican, Puerto Rican, and Cuban perspectives on American politics. Boulder, CO: Westview Press. Elam, S., Rose, L., & Gallup, A. (1993). The 25th Annual Gallup Poll of the public’s attitude toward the public schools. Phi Delta Kappan, 95, 137-152.

Gallup, A., & Elam, S. (1988). The 20th Annual Gallup Poll of the public’s attitudes toward the public schools. Phi Delta Kappan, 70, 33-46.

Hakuta, K. (1984). Bilingual education in the public eye: A case study of New Haven, Connecticut. NABE Journal, 9, 53-76. Harris, L. (1993). National survey of teachers. New York: Louis Harris and Associates.

Hosch, H. (1984). Attitudes toward bilingual education: A view from the border. El Paso: Texas Western Press.

Houston, Texas (1983). Metropolitan area survey. Houston: University of Houston Center for Public Policy.

Huddy, L., & Sears, D. (1990). Qualified public support for bilingual education: Some policy implications. Annals of the American Academy of Political and Social Science, 508, 119-134. Krashen, S. (1996). Under attack: The case against bilingual education. Culver City, CA: Language Education Associates. Krus, D., & Brazelton, J. (1983). Contributions to psychohistory: VIII. Perspectives on bilingual education in the Austrian Empire and the United States of America: Is the assumption of temporal contention of linguistic and territorial separatism valid? Psychological Reports, 53, 247-254.

Krus, D., & Stanley, M. (1985). Validity of the attitudes toward bilingual education scale with respect to group discrimination. Educational and Psychological Measurement, 45, 693-698. Media General Poll, November, (1985). Richmond, VA: Media General. Rossell, C., & Baker, K. (1996). Bilingual education in Massachusetts: The emperor has no clothes. Boston: The Pioneer Institute for Public Policy Research.

Shin, F., & Kim, S. (In press). Korean parent perceptions and attitudes of bilingual education. In R. Endo, C. Park, J. Tsuchida and A. Abbayani (Eds.), Current issues in Asian and Pacific American education. Covina, CA: Pacific Asian Press. In press. Shin, F., & Gribbons, B. (1996). Hispanic parent perceptions and attitudes of bilingual education. The Journal of Mexican American Educators, 16-22.

Shin, F., & Lee, B. (1996). Hmong parents: What do they think about bilingual education? Pacific Educational Research Journal, 8, 6571.

Time Magazine (1995). Rise of English-only. October 6, 1995: 49.

Stephen Krashen

University of Southern California

Copyright National Association for Bilingual Education Summer 1996

Provided by ProQuest Information and Learning Company. All rights Reserved