MULTIPLY ROBUST NONPARAMETRIC MULTIPLE IMPUTATION FOR THE TREATMENT OF MISSING DATA

摘要

Imputation offers an effective solution to the problem of missing values. We propose a nonparametric multiple imputation procedure that uses multiple outcome regression models and/or multiple propensity score models. Our procedure leads to a multiply robust point estimator in the sense that it remains consistent if all models but one are misspecified. We obtain a variance estimator and establish the asymptotic properties of the proposed method. The results of a simulation study, that assesses the proposed method in terms of bias, efficiency, and coverage probability, support our findings.