Propensity Score Matching and Its Application

• In Event: 52.076 - Effects of Design and Methods in Evaluating Policies and Programs

Abstract

In social sciences, there are often cases where it is unethical or even impossible to randomly assign participants to treatment and control groups. Propensity score matching methods were developed by Rosenbaum and Robin in 1983 to evaluate treatment effects in cases when random assignment cannot be used. The three goals of this paper are to facilitate the use of propensity score matching by: (1) critically reviewing the available literature on propensity score matching; (2) demonstrating the use of propensity score matching in a non-randomized program evaluation; and (3) provide syntax for applying propensity score matching. This paper is designed to enhance researchers’ understanding of propensity score matching and offer guidelines for how to use propensity score matching.

Authors

• Ming-E Chen, Indiana University - Bloomington

• Allison Ann Howland, Indiana University

• Jeffrey A. Anderson, Indiana University

• Li-Ting Chen, Indiana University

• Heidi R Cornell, Wichita State University