ABSTRACT
Statistical analysis with missing data has always been a challenge. However, there have been tremendous advances in statistical theory related to analysis with missing data. Of equal importance for the applied researcher is the ready availability of these advances in a variety of software applications. This chapter outlines a general approach to analysis with missing data called multiple imputation (Rubin, 1987; Schafer, 1997). The chapter is divided into three parts. In the first part, we give a brief introduction to analysis with missing data. In the second part, we discuss the general principles of multiple imputation and its application, particularly with Schafer’s (1997) Windows-based multiple imputation program, NORM. In the last section, we illustrate the use of NORM with an empirical example.
