Categorical Data Analysis Using The SAS® System, 2nd Edition

15.2. Methodology

15.2.1. Motivation

Correlated data come from many sources: longitudinal studies on health care outcomes, crossover studies concerned with drug comparisons, split plot experiments in agriculture, and clinical trials investigating new treatments with baseline and follow-up visits. You may have multiple measurements taken at the same time, such as in a psychometric study. You may also have clusters of correlated measurements: one example results from group randomization, such as randomizations of litters of animals to experimental conditions. Another example is sample selection of physician practices and the assessment of all of the patients in each practice, or cluster. Often, particularly with longitudinal studies, missing ...