ABSTRACT
Studies in healthy adults have shown that intranasal oxytocin increases trust. Does oxytocin have the same effect on individuals with borderline personality disorder (BPD) as it does on healthy individuals? At first glance, it would seem possible to answer this question by designing a study restricted to individuals with BPD. However, if the question of interest is literally whether the effect of oxytocin is the same for individuals with BPD as it is for healthy individuals, it is necessary to compare the effect of oxytocin for the two types of individuals in the same study. Notice the key concept here is comparing two effects to each other, namely comparing the effect of oxytocin for individuals with BPD to the effect of oxytocin for individuals without BPD. This type of comparison entails a factorial design, where we consider two factors in the same study. In this case, one factor is type of individual (i.e., BPD or healthy), and the other factor is administration of oxytocin or a placebo in order to assess the effect of oxytocin on trust. From a statistical perspective, the main benefit of a factorial design is that it allows us to assess the presence or absence of an interaction. In particular, if the effect of oxytocin is different for individuals with BPD than it is for others, we say that the two factors interact. We will see that we can also estimate the effect of oxytocin averaging over type of individual, which is called a main effect. More generally, in this chapter you will learn about various types of effects that can be tested in a factorial design.
