Testing Covariances and Mean Differences With Missing Data

This chapter integrates material covered in Chapters 3, 4, and 5 under a single framework. We present a first complete example of how to estimate statistical power to detect mean differences with incomplete data. We deliberately begin by extending the simple bivariate example first introduced in Chapter 4 to focus on ignorably missing data that are either nonsystematic (missing completely at random, MCAR) or systematic (missing at random, MAR).