This paper describes in detail the initial analyses of SACMEQ II data, using multilevel modelling techniques, to develop general models of school effectiveness for SACMEQ II member countries and specific models for Tanzania (including Zanzibar). Differences between schools in Grade 6 pupils? reading and mathematics achievements are explored and the percentage of variance in pupil outcomes attributable to school and country levels is estimated before and after adjusting for various pupil and school factors outside the control of the school. A wide range of key factors have been examined such as school location and resources, school leadership, school and community/parent relationships, school inspection, homework and extra tuition, teacher and pupil behaviour problems, grade repetition, and student and teacher absence, and a subset of these were found to be statistically significantly related to student academic achievement. The findings will be discussed in the light of previous research on school effects in sub-Saharan Africa.