On the Analysis of Repeated-Measurements Experiments

Suppose there are I levels of some treatment, with ni subjects exposed to the ith level, and each subject measured on some characteristic periodically for K times post treatment. Let yi ik be the measurement made at the kth time on the jth individual at the ith treatment level. Then there is associated with each subject for a given ith treatment level a vector of K measurements [yiik], k = 1 to K. Denote the average over the population from which the subjects were drawnl by E(yiik) = pi, . This average represents all fixed effects and may be decomposed into ,u + a, + yk + bik . The random departure of Yiik from vik may be denoted by m jk so that