ABSTRACT

Just like dichotomous outcomes and proportions, some types of dependent variable will always violate the normality assumption. This chapter treats ordered (ordinal) categorical outcomes, which have a uniform distribution, or counts of rare events. Such data are preferably analyzed using the multilevel generalized linear model introduced in Chapter 6. This chapter describes the use of the generalized linear mixed model for multilevel ordered categorical data and for count data. The statistical issues in estimation and interpretation are the same as the issues discussed in Chapter 6.