Estimating treatment effect heterogeneity in randomized program evaluation

Kosuke Imai, Marc Ratkovic

When evaluating the efficacy of social programs and medical treatments using randomized experiments, the estimated overall average causal effect alone is often of limited value and…