Overview of Linear Models for Longitudinal Data
Garrett M. Fitzmaurice, Nan M. Laird, James H. Ware · Wiley series in probability and statistics · 2011
This chapter introduces some vector and matrix notation and present a general linear regression model for longitudinal data. It considers some elementary descriptive methods for exploring longitudinal data, especially trends in the mean response over time. With longitudinal data, the covariance among the repeated measures can be expected to have certain features or patterns; with more general multivariate data, there is rarely any indication of structure to the covariance matrix. The chapter presents several approaches for modeling the mean of a vector of longitudinal responses. For the case of linear models for continuous longitudinal responses, the area under the curve (AUC) for the mean response over time coincides with the mean of the AUCs for the individuals in the population of interest. The chapter also considers distributional assumptions concerning the vector of random errors.