ABSTRACT
So far, we have seen how to compare the means of a groups of individuals. This chapter continues this theme but in a more general context. We now consider designs where the groups are defined by two or more factors (independent variables). For example, suppose that a psychologist wants to evaluate the effectiveness of biofeedback and drug therapy for treating hypertension, that is, for lowering blood pressure. The psychologist might design a study with four groups: both biofeedback training and drug therapy, biofeedback but no drug therapy, drug therapy but no biofeedback, and neither biofeedback nor drug therapy. We will see later in this chapter that such a design provides efficient tests of the individual effects of biofeedback and drug therapy, as well as the effect of the two in combination. As before, each subject selected to participate in the study would be assigned to one of the four groups, ideally at random.
