ABSTRACT
To this point, we have concerned ourselves with the range of analytic contexts in which sufficient statistics such as means, variances, and covariances were all that was necessary to estimate statistical power. This approach, however, only begins to scratch the surface of situations that may be of interest to the applied researcher. By extending our consideration of statistical power with missing data to include Monte Carlo simulation studies with raw data, the range of applications and situations that can be evaluated is greatly expanded. In this chapter, we begin by very briefly considering some guidelines for planning and implementing a Monte Carlo simulation study, along with references to more detailed sources. Next, we present some of the different ways of generating raw data for use in a Monte Carlo study. The latter part of this chapter is devoted to exploring some applications of Monte Carlo methods with missing data, such as evaluating convergence rates, assessing model fit statistics, complex missing data patterns, and violations of model assumptions.
