I was wondering if anyone could tell me if this is right/ wrong (and if so, how can I fix it). So, I have 3 variables: X (binary), M (continuous moderating variable) and Y (continuous). I conducted a linear regression with interaction and it was all significant. I am now trying to plot it. I read this paper (Baron & Kenny, 1986) and it showed a conceptual graph of a binary X and continuous M.
I have tried to base my own graph on this:
So, I put the moderator on the X axis and my outcome variable on the Y axis. The two lines then show the performance of the 2 groups (which is my X binary variable). However, I am not sure if this conveys the effect of a moderation/ interaction? Does it not look like I just used my binary X variable as the moderators?
Best Answer
My answer is yes. The slop (combined red and green) reflects the effect of moderation which appears very strong significance from the plot (as linear lines showed). The gap/distance between red and green lines reflects the effect of the binary X (it appears also very significant). The interaction effect will be reflected by whether the two lines, red and green, parallel or not. If parallel, it means no interaction effect. Statistically, the more of the two lines non-parallel, the stronger effect of interaction (the plots shows there is some level of interaction between X and moderation).