When Evidence Supporting Bilingual Education is Interpreted as Evidence for Dropping Bilingual Education: The Case of Bali (2001)

Submitted for publication

Consider these results: There are two groups of children in elementary school. One group had bilingual education, varying from one to three years. A second group did not have bilingual education, but was always in all-English classes. Both groups are put in the same all-English program for one year. The group that had bilingual education makes better gains. What does this mean? Obviously, it means that bilingual education was beneficial, that it did a better job of preparing these children for an all-English curriculum. This study was done, but the author’s interpretation is the exact opposite.

Bali (2001) examined the immediate impact of Prop. 227 in the Pasadena school district, comparing gains over one academic year made by English learners who had previously been in all-English programs and those who had been in bilingual programs and who were shifted into all-English classes. She controlled for socio-economic status, as well as for the fact that fewer children in bilingual education were tested on the pre-test.

The former bilingual education students made better gains. In 1998, just before Prop 227 passed, students in bilingual education scored 2.4 points lower than English learners students in all-English on SAT9 reading. But in 1999, after a year of all-English, the now former bilingual students scored only .37 points lower. In other words, they caught up to those who had done only all-English. (In math, there was no change. In 1998 bilingual students did .5 points better, in 1999 .49 points better.)

Instead of giving bilingual education the credit for these gains, Bali concludes that "abruptly" stopping bilingual education had no ill effects and actually helped; it was a "small step toward equalizing educational outcomes between Hispanic and white students" (p. 312).

Bali may have been influenced by the fact that in 1998 those in bilingual education had slightly lower scores. This was probably because children with less English proficiency when entering school tend to be placed in bilingual rather than all-English classes. The only data in Bali’s study that is properly statistically controlled is the gains made after all ELLs were put in all- English classes. Those with previous bilingual education made better gains. This suggests that bilingual education prepared them better for all-English than all- English.

We must be cautious. The gains were small, and Bali presents no description of the kind of bilingual program Pasadena had, nor does she tell us very much about the all-English classrooms. What is clear is that her data does not support her conclusion that stopping bilingual education "actually helped." Rather, the data suggests that doing bilingual education helped.

Bali, Valentina (2001). Sink or swim: What happened to California’s bilingual students after Proposition 227? State Politics and Policy Quarterly Spring 2001: 295-317.

Thanks to Philip Leeman for informing me about this paper.