ABSTRACT
The purpose of this chapter is to introduce a wide array of topics in the area of multilevel analysis that do not fit neatly in any of the other chapters in the book. We refer to these as advanced issues because they represent extensions, of one kind or another, on the standard multilevel modeling framework that we have discussed heretofore. In this chapter, we will describe how the estimation of multilevel model parameters can be adjusted for the presence of outliers using robust or rank-based methods; multivariate multilevel data problems, in which there are multiple dependent variables. We will look at multilevel generalized additive models, which can be used in situations where the relationship between a predictor and an outcome variable is nonlinear. We will finish with a discussion on predicting level-2 outcome variables with level-1 independent variables, and with a description of approaches for power analysis/sample size determination in the context of multilevel modeling.
