Chapter 13 Mediation and Moderation
Sometimes, the hypotheses we want to test are ones where the effect of interest works in a very specific way. In some cases, we might hypothesise that the relationship between two variables is actually best explained by a third variable that acts between them. For instance, we might predict that a relationship between sugar levels in food and happiness is best explained by food intake - we might eat more sugar, and eating more makes us happier (or something like that).6 This kind of hypothesis is called a mediation, and is common in psychological research as a lot of our work deals with processes.
In other instances, we might expect interaction effects like what we saw in two-way ANOVA, but between continuous predictors. Note that up until this point we have only considered categorical x categorical interactions within an ANOVA context, so now is a good time to explore continuous x continuous interactions. We describe these effects as moderation effects.
This chapter will cover both mediation and moderation.
I’m no dietitian or food psychology expert so I make no claims as to the veracity or accuracy of this statement.↩︎