A Multilevel Model Of School Effectiveness In A Developing Country

The comparative effectiveness of schools in developing countries has become the center of a lively debate. Of particular concern is the appropriate analytic method to employ when examing school effects. This paper uses a multi-level approach to examine determinants of growth in grade 8 mathematics achievement in Thailand. Results of the analysis showed that schools in Thailand were equally effective in transforming pretest scores into posttest scores, and that schools and classrooms contributed 32 percent of the variance in posttest scores. Higher levels of achievement were associated with a higher proportion of teachers qualified to teach mathematics, an enriched curriculum and frequent use of textbooks by teachers. Individual characteristics, however, contributed 68 percent of the variance, with achievement higher for boys, younger students, and children with higher educational aspirations. The model developed in the paper was able to explain most of the between school variance, but significantly less of the within school variance. The implication of these results is that schools in Thailand are much more uniform in their effects than previous research in developing countries would have suggested.