MULTILEVEL INFLUENCES ON SCHOOL DISORDER: A COMMENT ON WELSH, GREENE, AND JENKINS*

Abstract

Hoffman and Johnson comment on research by Wayne N. Welsh et al regarding a multiuse statistical model to examine whether school and community characteristics affect school misconduct at an individual level. The research using a multilevel model is admirable, but the conclusions about the primacy of individual predictors are not well-supported by either the sample or models used.