ABSTRACT
The estimation and inference methods of Chapters 3 to 5 are mainly concerned with evaluating the mean time curve E[Y (t)] = µ(t) without incorporating the effects of potential covariates. In most practical situations, the scientific interests of a longitudinal study are often focused on evaluating the effects of time t and a set of covariates X (l)(t), l = 1, . . . , k, which may or may not depend on t, on the chosen time dependent outcome variable Y (t). The objective of this chapter is to present a series of methods for modeling and estimating the effects of a set of time-invariant covariates on a real-valued longitudinal outcome variable. Methods for the general case involving time-dependent covariates are presented in Chapters 7 to 9.
