ABSTRACT
The majority of the literature on cross-over designs assumes a very simple model for the direct and carry-over effects and that is the model we used in the previous chapters. In this model there is a unique carry-over effect for each treatment and that carry-over effect does not depend on the following treatment, if there is one, or on the magnitude or nature of the treatment which produces it. Let us label the t treatments as A, B, C,…, and the corresponding direct effects as . In this simple model the corresponding carry-over effects are denoted by . So, for example, if a subject received the treatment sequence BAC, the simple model for the direct and carry-over effects in these three periods would be and , respectively.
