Using hierarchical linear modeling to study social contexts: The case of school effects

This article focuses on a particular educational context, the school, and how characteristics of the structure and organization of high schools influence students' academic development. The emphasis is on a type of quantitative inquiry called school effects research. It describes a methodology that is most appropriate for conducting studies of school effects in particular and educational contexts in general: hierarchical linear modeling (HLM). Two previously published studies are used as heuristic examples of school effects studies conducted with HLM methods. Both studies use large and nationally representative longitudinal data from the National Education Longitudinal Study of 1988 to explore school effects on learning and its social distribution by student socioeconomic status. Study 1 focuses on the effects of high school size on learning. Study 2 focuses on how teachers' attitudes, taken as a collective property of the social organization of schools, influence both learning and its social distribution. Implications for both policy and research are discussed.