Causal Inference Using Propensity Score Methods Workshop:

Abstract: The estimation of causal effects is a central goal of social science research. In this workshop, I introduce propensity score methods that are widely used by empirical researchers to conduct reliable and efficient causal inference in both experimental and observational studies. Specifically, in this workshop, we discuss several uses of propensity score methods. First, we review the standard use of propensity score methods for a binary treatment. In this context, we introduce the covariate balancing propensity score (CBPS) methodology which estimates the propensity score such that the resulting covariate balance is optimized. Second, we show how to use propensity score methods in other contexts. In particular, we show that the propensity score methods can be used to estimate causal effects of non-binary treatments, extrapolate the experimental and instrumental variable estimates to a target population, and estimate causal effects of time-varying treatments in panel data settings. We illustrate these methods with social science applications and open-source software so that researchers can readily apply these techniques to their own substantive research questions.