ABSTRACT
In order to conduct a more complete analysis of data with interaction, it is helpful to determine where the interaction occurs in the data. For example, if it is known that all of the interaction in an experiment is caused by only one of the possibilities (often a control) of the T-treatments, then all of the remaining possibilities of the T-treatments could still be compared after averaging over all of the possibilities of the B-treatments. This results in more powerful tests for comparing the possibilities of the T-treatments that do not interact with the B-treatments.
