Abstract:
|
We consider causal inference from observational studies when confounders are partially observed. When the confounders are missing at random, multiple imputation is commonly used; however, it requires congeniality conditions for valid inferences, which may not be satisfied when estimating average causal treatment effects. Alternatively, fractional imputation, proposed by Kim 2011, has been implemented to handling missing values in regression context. In this article, we develop fractional imputation methods for estimating the average treatment effects with confounders missing at random. We study the asymptotic properties of the proposed estimator, which provide a basis for valid inference. Via simulation study, we show our proposed estimator outperforms existing estimators that are widely used in practice.
|