How Do We Know if a Program Made a Difference? A Helicopter Tour of the Econometrics of Non-Randomly Allocated Community-Level Programs

Gustavo Angeles, University of North Carolina at Chapel Hill
Peter M. Lance, University of North Carolina at Chapel Hill

This paper offers a survey of popular econometric methods for addressing statistically community-level characteristics that might otherwise prevent us from obtaining valid estimates of the impact of human resource policies or initiatives (for instance, health or family planning initiatives). We begin by characterizing the problem and then provide a survey of the various econometric methods typically applied in response. Our survey focuses on various ‘classes’ of estimation strategy. In each case, we emphasize data requirements and assumptions, sources of identification and interpretation of results. We conclude with a brief coverage of a few techniques that have received significant attention in the recent literature (such as regression discontinuity design). Our goal is to offer empirical researchers from a variety of disciplines an accessible and intuitively appealing guide to the tools available from the econometrics tradition