Multiple Imputation in Practice

TLDR

Missing data complicates scientific analysis, prompting decades of research into methods such as Rubin’s multiple imputation, a broadly applicable general‑purpose approach. This review examines multiple imputation as an analytic strategy for missing data. The authors evaluate and compare several software packages implementing multiple imputation, contrasting interfaces, features, and results. The comparison reveals package shortcomings and useful features, and the authors suggest additional functionalities while outlining limitations and cautions for using multiple imputation.

Abstract

Missing data frequently complicates data analysis for scientific investigations. The development of statistical methods to address missing data has been an active area of research in recent decades. Multiple imputation, originally proposed by Rubin in a public use dataset setting, is a general purpose method for analyzing datasets with missing data that is broadly applicable to a variety of missing data settings. We review multiple imputation as an analytic strategy formissing data. Wedescribe and evaluate a number of software packages that implement this procedure, and contrast the interface, features, and results. We compare the packages, and detail shortcomings and useful features. The comparisons are illustrated using examples from an artificial dataset and a study of child psychopathology. We suggest additional features as well as discuss limitations and cautions to consider when using multiple imputation as an analytic strategy for incomplete data settings.