Multiple imputation of multilevel data

Abstract

This chapter discusses critical issues associated with imputation of multilevel data. Section 10.2 introduces the notation used and outlines how two formulations of the same model are related. Section 10.3 dissects the multilevel missing data problem into five main questions that need to be addressed. Section 10.4 outlines six different strategies for dealing with the missing data problem. Section 10.5 describes a multilevel imputation method for univariate data, and discusses its properties. Section 10.6 describes a method to apply the univariate method iteratively to multivariate missing data. Finally, Section 10.7 sums up the major points and provides directions for future research.