Up to now we've been talking about mediation. So, we covered mediation, parallel mediation, serial mediation. Now, we switch gears and start talking about moderation. And this is the agenda for today. First of all, what is moderation? And how can we conduct a moderation analysis in a sophisticated way? Okay, so how do we do this? How can we find results for our moderation models that we are, theoretically, developing through literature reviews and through all the conversations you have potentially with faculty, with friends? We'll get there, we'll get there. So first, what is moderation? So when we are talking about moderation, we are looking at the circumstances, or when or where our independent variable influences our dependent variable. We are looking at a moderator, a variable that alters, that changes the strength or direction of the relationship between our independent variable and our dependent variable. Job meaningfulness, does it always influence job performance? When is this relationship stronger? When is this relationship significant? Perhaps, if you have different levels of organization identification, the relationship between your job meaningfulness and job performance could change. Those are the type of questions that we are looking at when we are testing moderation models or adding a moderator to your model. And this is the graphical representation that you usually have when testing for moderators. So we have a main effect, the independent variable influencing the dependent variable, but that relationship is a function of a particular moderator. It can change based on the level of the moderator. How do we test for a moderator or moderation models? It's pretty simple. Again, we have a series of steps. So first, in Step 1, you have to mean center your independent variable and your moderator, especially if they are continuous variables. If they are categorical, like gender, you don't need it to center the moderator. Usually the moderator would be the gender. But if they are continuous variables, like job meaningfulness or organization identification, yes, you have to center. And the reason is we need to avoid multicolliearity issues. The way that we create a moderator variable is by multiplying your independent variable and the moderator. You create an interaction term. And when you do that, the correlations tended to be really high. So, to avoid multicolliearity issues, you needed to mean center your independent variable and your moderator. And in Step 2, you run a regression in which you have your independent variable, and your moderator, and also the dependent variable. In Step 3, you add the interaction term. So now, you have your independent variable, your moderator, and the interaction term, which is the multiplication of your independent variable and your moderator. And also you have your independent variable So next you'll do a simple slope analysis. And this is questionable because some scholars have brought to light, actually, that the purpose of a moderating model is not to show that the slopes at different levels of the moderator are significant. The purpose of a moderating analysis is to see if the effects of the independent variable on the dependent variable change base on the level of the moderator. Finally, we needed to plot, and the reason is, we needed to take a look at the pattern of this relationships. So it helps us to understand and interpret these moderating effects. I usually adopt a Excel spread sheet that you can download from the Jeremy Dawson website, it's pretty cool, it's pretty neat. And you just have to enter the coefficients in the spreadsheet and it will give you the graphical representation of your interaction. On top of that, you can also do simple slope analysis by using this spreadsheet. I would strongly recommend you to download this spreadsheet.