Discussion of ”Dissecting multiple imputation from a multi-phase inference perspective: what happens when god’s, imputer’s and analyst’s models are uncongenial?”

Abstract

<p>We would like to first congratulate Drs. Xie and Meng on their excellent work on investigating the mystery of multiple imputation. Multiple imputation (MI) has been promoted as a general purpose estimation tool for missing data, but there are debates over its statistical validity in many practical situations. This article will certainly serve an important building block to address these debates from a multiphase inference perspective.</p>