Interactions: Factorial ANOVA

Visualization of the interaction

In contrast to the one-way ANOVA approach, the factorial ANOVA approach allows us to estimate the interactive effect directly along with a standard error. We can then use it to calculate a measure of confidence and formally test the interaction, if we want. Before we get to more formal statistical modeling, it’s useful to visualize the interaction. R has a handy interaction.plot() function that can help us with this, shown in the code widget below: