Model building and testing for the change‐over design

The two‐period cross‐over design is discussed within the framework of univariate and multivariate analysis of variance technique; the relations between both procedures are explained. It is shown that all hypotheses of interest can be tested if the design is regarded as a special case of a repeated measurement design. Some features of the n‐period change‐over design are explained by discussing the model and hypotheses of a three‐period design.