Analysis of Incomplete Data using Multiple Imputation

This book gives a general overview of the theoretical background needed for the implicit and explicit assumptions which are required for ‘proper’ Multiple Imputation (MI), and provides insights as well as examples about how to handle incomplete-data analysis using MI techniques. Although the authors aim to motivate the usage of MI from a theoretical standpoint, they gradually leave the ‘laboratory world’ where distributional assumptions hold. As the book progresses it becomes increasingly applied in nature; current problems in empirical settings are addressed, and the current frontier in MI related research on empirical problems is described.

The first five chapters of the book build on each other and provide the fundamental information for the remainder of the book, but some of the later chapters can be read in random order, as they address topics such as diagnostics or special applications of MI methods.