A Regression-Based Strategy for Defining Subgroups in a Social Experiment

| James J. Kemple, Jason Snipes

Many evaluations of education, employment and training, and other programs ask not only “what works?” but “what works for whom?” Addressing the latter question calls for analyses of program effects on different subgroups of people. Using data from MDRC’s evaluation of Career Academies, this paper from the MDRC methodology series describes a regression-based strategy for defining subgroups that is more systematic than traditional accumulation strategies.