External validity is a challenge in treatment effect estimation. Even in randomized trials, the experimental sample often differs from the population of interest. If participation decisions are explained by observed variables such differences can be overcome by reweighting. However,...