ABSTRACT
Often in clinical, epidemiological, or laboratory studies the response of each individual is recorded repeatedly over time. In such experiments, comparisons among treatments (or other characteristics of primary interest) are being assessed over time. This type of experiment is often called a repeated measures experiment. Repeated measures designs can be employed when, for example, each subject receives many different treatments so that different subjects may receive different sequences of treatments. Such studies are often referred to as “crossover” designs and will not be our focus in this chapter. Two excellent references for readers who have an interest in crossover designs are Fleiss [49] and Brown and Prescott [14]. In this chapter, we’ll consider repeated measures data where each subject, animal, or experimental unit receives the same treatment and is measured repeatedly or is repeatedly observed under the same condition. We’re interested in either characterizing a single group of such subjects or possibly comparing two or more groups of subjects where the individuals within a given group each receive the same treatment over time. Treatments or conditions across groups of subjects will be different. Because each subject is associated with only one treatment, the subject effect is said to be nested within the treatment effect. The general data structure for such studies is depicted below.
