Replication data for: Estimating Individual Causal Effects

Most empirical work focuses on the estimation of average treatment effects (ATE). In this dissertation, I argue for a different way of thinking about causal inference by estimating individual causal effects (ICEs). I argue that focusing on estimating ICEs allows for a more precise and clear understanding of causal inference, reconciles the difference between what the researcher is interested in and what the researcher estimates, allows the researcher to explore and discover treatment effect heterogeneity, bridges the quantitative-qualitative divide, and allows for easy estimation of any other causal estimand. The framework I develop for estimating ICEs starts from the potential outcomes framework and then combines existing methods for matching in causal inference with a Bayesian model to impute missing potential outcomes. Researchers can use the resulting posteriors for the ICEs to derive the posterior for any other causal estimand by simple aggregation. In my dissertation, I first lay out the basic framework and estimation strategy. I then compare various models via simulation to test the effectiveness in recovering the true ICEs. Finally, I apply the model for estimating ICEs to two applications: a randomized field experiment on monitoring corruption from Olken (2007) and an experiment on the effectiveness of job training programs. I show the flexibility of the model in estimating ICEs for different types of outcome and treatment variables as well as with two-stage models using instrumental variables. I also show the various ways one can use the model to detect treatment heterogeneity and estimate a large number of different causal estimands. Complete date fields below for: time period covered; and date of collection