A multi level model of school effectiveness in a developing country

What makes one school more effective than another - particularly which inputs and management practices most efficiently enhance student achievement - has become the center of lively debate in the literature. Which method to use to compare school effects particularly concerns analysts. The model developed by the authors is able to explain most variance between schools but significantly less within schools. Only one variable slope is observed: the relationship between educational aspirations and achievement. The authors apply multi level techniques to longitudinal data recently collected by the International Association for the Evaluation of Educational Achievement in Thailand. One question they try to answer is : how do estimates obtained from the new multi level techniques compare with those obtained from ordinary regression models?