This presentation addresses the question of what language experiences are most appropriate for the academic instruction of children who do not speak English as a native language. The question is explored through a reanalysis of data gathered in a previous 6-year study that varied English and Tagalog instruction across grades one through six in 18 Philippino classrooms. It was concluded that (1) proficiency in English is directly related to the number of years in which it is used as the medium of classroom instruction, and (2) the average level of literacy in Tagalog is not closely related to the number of years in which it has been used as a medium of classroom instruction. Throughout the reanalysis of the Philippino data, methodological considerations which should be applied when choosing the unit of analysis, covariates, and variables for use in a multivariate analysis of covariance procedure are demonstrated. The conclusions of the original study were supported by the reanalysis of the data. On the basis of these results, it is recommended that English be used as the language of instruction for non-native speakers of English and that such instruction be accompanied by appropriate instruction in native language arts and culture. (Author/RH)