New York Times

Bilingual Education Wasn't a Cure 

To the Editor: 

An Aug. 20 front-page article about the improvement in standardized test scores in California suggests that Proposition 227, which eliminated bilingual education, was a success. 

But as the article notes, it is not clear whether dropping bilingual education was responsible for these gains because other changes may also have occurred, including increased test preparation and reduction in class sizes in the lower 
grades. 

A less obvious factor is selective testing. An article in The San Francisco Chronicle on July 22 documented several cases in the Bay Area in which increases in test scores coincided with decreases in the number of students tested. 

Controlled studies comparing groups that differed only with respect to the use of the first language have consistently shown that children in properly organized bilingual education programs acquire English at least as well and usually better than children in all-English programs. 

STEPHEN KRASHEN 
Malibu, Calif., Aug. 20, 2000 
 

Note: Krashen's original letter (before editing) follows:

Jacques Steinberg (August 20, 2000) states that the increase in California's SAT9 test scores suggests that Proposition 227, which eliminated bilingual education, was a success. 

Using SAT9 scores to evaluate bilingual education is bad science. As Steinberg notes, it is not clear that dropping bilingual education was responsible for these gains, because other changes may also have occurred. Obvious changes include increased test preparation and reduction in class sizes in the lower grades. A less obvious factor is selective testing. A recent article by Nanette Asimov in the San Francisco Chronicle documented several cases in the Bay Area in which increases in test scores coincided with clear decreases in the number of students tested. 

The only way to really determine the effect of bilingual education on English language development is to examine the results of controlled scientific studies. In these studies, groups are compared that differ only with respect to the use of the first language. This research consistently shows that children in properly organized bilingual education programs acquire English at least as well and usually better than children in all-English programs. In addition, students in bilingual programs drop out less. 
 

Stephen Krashen 
Professor of Education 
University of Southern California