|
Abstract:
|
Inferring a causal relationship is an important task in social science and health research. In a large population, different subgroups might respond differently to certain treatments. Identifying and estimating the heterogeneous effects can help researchers improve treatments or better allocate resources to meet the needs. In studies with real world data, propensity score is often used as a dimension reduction tool to aid the inference under the ignorable treatment assignment assumption. In this work, we develop a semiparametric Bayesian approach for efficient heterogeneous causal effect estimation in observational studies. Our model incorporate propensity scores and observable factors into potential outcome models via flexible Gaussian process regressions. We show that our model produces estimators take the same form as traditional matching estimators under certain prior specifications, and outperform the matching estimators with improved efficiency and better identification of heterogeneous effects. Furthermore, we apply our method to investigate the impact of college attendance on women fertility, which is known to suffer from the potential heterogeneous effects.
|
