Missing Data, Imputation, and the Bootstrap

Abstract Missing data refers to a class of problems made difficult by the absence of some portions of a familiar data structure. For example, a regression problem might have some missing values in the predictor vectors. This article concerns nonparametric approaches to assessing the accuracy of an estimator in a missing data situation. Three main topics are discussed: bootstrap methods for missing data, these methods' relationship to the theory of multiple imputation, and computationally efficient ways of executing them. The simplest form of nonparametric bootstrap confidence interval turns out to give convenient and accurate answers. There are interesting practical and theoretical differences between bootstrap methods and the multiple imputation approach, as well as some useful similarities.