ABSTRACT
The multilevel regression model is more complicated than the standard single-level
multiple regression model. One difference is the number of parameters, which is much
larger in the multilevel model. This poses problems when models are fitted that have
many parameters, and also in model exploration. Another difference is that multilevel
models often contain interaction effects in the form of cross-level interactions. Interaction
effects are tricky, and analysts should deal with them carefully. Finally, the multilevel
model contains several different residual variances, and no single number can be
interpreted as the amount of explained variance. These issues are treated in this chapter.
