I am writing in response to issues raised in guest columns and letters to the editor in the Tucson Citizen on the issue of bilingual education. I am one of only two people - the other is my co-author, Keith Baker - in the United States who have read all the empirical research on bilingual education and classified the scientific studies into those where bilingual education does better than the alternative, the same, or worse.

Of the 300 program evaluations Keith Baker and I read, we found only 72 that were scientific. When we examined the 72 scientific studies, we found that with regard to second language reading, or oral proficiency for young students, 33 percent showed it to be no different from doing nothing. When we looked at language, a test of a student's understanding of grammatical rules, we found 64 percent showed bilingual education to be inferior, and 29 percent showed it be no different from doing nothing. In math, 35 percent of the studies showed bilingual education to be inferior, and 56 percent showed it to be no different from doing nothing.

No studies showed bilingual education to be superior in reading, language, or math to structured immersion - all -English instruction is self-contained classroom of second language learners.

With regard to other outcomes besides achievement, there is very little evidence. The studies showing that bilingual education reduces dropout rates are hard to evaluate because bilingual education at the secondary level is typically sheltered English and sheltered subject matter, although it may be labeled "bilingual" education increases the dropout rate is also unsubstantiated.

In conclusion, the research suggests that children who do not know English in Arizona are better off in structured immersion - an all-English program with academic support where the English is taught at a pace they can understand. Unfortunately, the research also indicates that an all-English program will not have a large effect on the achievement of these children. The most important causes of the low achievement in English of limited-English proficient children are their poverty and the fact that a non-English language is spoken in their home and community. A much broader approach is needed to address these issues. CHRISTINE ROSSELL, Political Science professor, Boston University, Boston, Mass.