Longitudinal Data: Simple Univariate Methods of Analysis

These notes descibe various simple models and methods for analysis of longitudinal data and show how the analyses can be carried out using SAS. The descriptions require some knowledge of analysis of variance, also with random eeects as in split-plot experiments, in particular. Knowledge of multivariate analysis is not required, and the plain multivariate analysis of variance , treating each series as a single multivariate measurement, is not accouted for here. Other relevant methods that are not included are random coeecients regression and the so-called ante-dependence methods. General references for statistical analyses of this kind of data are 1] and 2]. Below is rst given a brief introduction to the problems and an example of multivariate data which will be used to illustrate the methods. Section 2 contains the description of the methods, while SAS-programs and output are collected in Section 3.