Multiple imputation proposed by Rubin is a statistical technique for analyzing incomplete data sets. The goal of multiple imputation is valid inference. The aim of this paper briefly reviews multiple imputation and compare the software packages. Then,...
Multiple imputation proposed by Rubin is a statistical technique for analyzing incomplete data sets. The goal of multiple imputation is valid inference. The aim of this paper briefly reviews multiple imputation and compare the software packages. Then, using examples from two data sets, we apply the methods and illustrate that the MI procedure creates imputed data sets and MIANALYZE procedures can be used to combine results from the analysis.