What Works is a concise methods text that represents a new approach for policy program analysis. The authors, Meier and Gill, combine statistics with normative concerns. They consider how things might be, and they focus on subsets of cases that differ from the norm. Their approach uses regression and methods in a qualitative, yet rigorous manner. In What Works, the authors address questions such as the following: why do some agencies learn to perform missions faster than others? What factors influence this learning? In which states do criminal justice policies based on deterrence work? What do excellent school districts do differently from those that are simply better than average? Why do some firms comply with public policy quickly while others wait?The case examples the authors employ and evaluate are especially helpful. What Works will appeal to anyone seriously interested in policy analysis, and in learning about--and understanding--new approaches for policy program analysis.