Data from an earlier study of a cognitively-based English-language training program for 67 Mexican-American Head Start children were re-analyzed using meta-analysis. A comparison of posttest and follow-up test results on various instruments showed that the use of Spanish as the language of instruction resulted in larger effect sizes than did the use of English, both languages, or a control treatment using arts, crafts, and music activities. Recently expressed doubts about the efficacy of beginning instruction in non-English speakers' native languages are questioned.
