Causal inference for average treatment effects of multiple treatments with non-normally distributed outcome variables

摘要

In this paper, we conduct a simulation study to investigate the performance of three popular weighted estimators, namely the Imbens' weighted estimator, the propensity score weighted estimator, and the generalized doubly robust (GDR) estimator for estimating average treatment effects of multiple treatments when the outcome variable is generated from one of the seven commonly seen non-normal distributions: lognormal, student's t, exponential, uniform, gamma, beta, and Weibull. When outcome variables are non-normally distributed, simulation results show that the GDR estimator has a better performance compared to the other two estimators.