5.3. k Populations

Suppose there are k different treatments and the ith treatment, say Ai, is applied to a group of ni subjects which are observed over time for p time points. In such a situation, the p dimensional observations on these groups can be thought of as the respective random samples from the k populations corresponding to various treatments. As a first step, we can look at the problem of comparing various treatments as a problem of multivariate one-way classification. Clearly, the p dependent variables, namely, the observations taken over time, are correlated and may have different means for different time points as well as for the different treatments.

For the jth dependent variable (that is, response at jth time point) j = 1,..., ...

Get APPLIED MULTIVARIATE STATISTICS: WITH SAS® SOFTWARE now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.