Chapter 7. Learning

Statistical Lesson: Interactions can be tricky to interpret since we have to take into consideration the combined influence of the main effects. A general rule of thumb in interpreting interaction graphs is to look at whether or not the lines connecting the groups of one of the main effects are parallel or not. The more parallel the lines are, the less likely there is an interaction. While the less parallel or more perpendicular the lines are (they cross each other), the more likely there is an interaction.

Graph A below shows an interaction (lines that cross) and Graph B shows no interaction (lines are closer to being parallel). In Graph A, we can see that a participant’s score depends on the learning method (Method 1 or 2) to which they were exposed and their style group (Group A or B). “A Style Learners” performed better when they utilized Method 1, while “B Style Learners” performed better when they utilized Method 2. It is the combined effect of method and learner style that makes this an interaction.

Meanwhile, in Graph B we see that the lines are close to parallel, which means that the same thing occurred for “A Style Learners” and “B Style Learners”: they increased their test scores regardless of the method to which they were exposed. In this example, we can see that those exposed to Method 1 always performed better than those in Method 2, and that “B Style Learners” always outperformed “A Style Learners.” Each main effect, method and learner style, was significant but there is no interaction or combined influence of both main effects on learning in Graph B.

Be aware that independent variables are referred to as “main effects” in some research designs but they are essentially referring to the same thing. When we state “there is a main effect,” we basically mean that the independent variable was found to have a significant effect on the outcome. If we state that “there is no main effect,” then the independent variable did not have an effect on the outcome, or was not significant.