Extending from the prior episode on categorical by continuous interactions in regression, in the current episode, Dan explicates how to test, probe, and plot an interaction between two continuous predictors...

Dan compares models with and without an interaction to show how the inclusion of a product term for the two predictors allows the effect of one predictor to vary as a function of the value of the other. He then re-arranges the regression equation to highlight the simple intercept and slope that characterizes the relationship between the outcome and one predictor (the focal predictor) at particular levels of the other predictor (the moderator). Probing the interaction typically involves testing these simple slopes at high, medium, and low values of the moderating predictor and generating a conditional effects plot to communicate the results. Continuing the example from the prior episode, Dan considers the prediction of course quality ratings from the easiness of a class and the level of attractiveness of the professor (now rated on a continuous scale as compared to Episode 7).