So let's do it. Our dataset, the dataset that we shared with you has all these variables here. And you'll be able to come back to this analysis if you decided to download this dataset. If you have your own dataset, and you have different moderators, feel free to dot that dataset as well, and you can run now your own models. The steps will be the same, just the variable names will be different. The first step is to test for the mediation. Remember, so we are looking at the indirect effect of meaningfulness on performance via job engagement. Let's run the analysis, click on Analyze > Regression > PROCESS. And don't forget, don't forget to select model number 4, we are conducting a mediation analysis here. And then we just enter dependent variable is our performance, independent variable is our meaningfulness, and engagement is our mediator. Don't forget to select the Total effect, and also the Sobel test. We don't use Sobel test anymore because we know based on Monte Carlo simulations that the indirect effect coefficient is not normally distributed. So, the Sobel test has this assumption that the indirect effect is normally distributed. So, instead of using the Sobel test, we look at the bootstrapping analysis. But for completeness, we are getting the Sobel test here. Click on OK, and you should see the output. Again, we are not covering the output, you can watch this session about mediation to get more details on how to conduct mediations, and how to interpret the output from mediation models. The second step is to run this moderation model in which we have the interaction term between meaningfulness and organization identification. And we look for the significance of that interaction term on our mediator, job engagement. We go to Analyze > Regression> PROCESS. Our outcome variable will be engagement, our independent variable will be meaningfulness, and our moderator will be organization identification. Because we are conducting a moderation model analysis, model number 1. And don't forget to mean center the products to avoid multicollinearity issues, and print the matrix that we get covariance coefficients. Click on OK, you'll get the output. And now, we have our step 3 in which are conducting a different moderation test. And we are looking for the interactive effects of meaningfulness and the task complexity on performance. And it's pretty similar to what we just did in our step two. The only difference is that we are changing our outcome variable, now we have performance as our outcome. We have engagement as our independent variable, yes, our independent variable. And a task complexity as our moderator. We still have model number 1. We need still need to mean center the products, and get all the coefficients for our covariance matrix, and just clicking on OK, you'll get the output. For these two moderation models, you can go back and watch this session on moderation, and you have a more thorough explanation on how interpret and plot the interaction terms. Finally, we are conducting our model and testing for this dual stage moderated mediation models. We needed to select model 29. Remember, model selecting the right model is extremely important when you are running PROCESS macro. So, go to PROCESS, and now you have to select model 29. And what's interesting and different here is that you have the same performance, meaningfulness, and engagement. But now you have your stage 1 moderator in the W position, and you have your stage 2 moderator in the V position. It's important to put these variables in the right places otherwise, you'll get error terms or output that is not interpretable. Mean center the products, and print the covariance matrix. Click on OK, and you'll get the output. So, in the output, first, double check that you select the correct model, model 29. And then you can check the variable names, performance, meaningfulness, engagement, organization identification, and task complexity. Yeah, everything is there. Scroll down on your output, and you'll see the two different steps there. Or actually, in this case, you'll see just one step with the covariance matrix. So, in step one, you'll be looking at the interaction term, one into that impact on job engagement. And the interaction term one should be meaningfulness times organization identification. We do find a significant relationship between this interaction term and engagement, p is less than 0.05. And here you have the covariance matrix for these parameters. You will need these coefficient to plot the interaction and actually run simple slope analysis. And then, we have our second step, in which we have a bunch of different coefficients, and we'll go through them. So, what we are looking for here is, if the interaction term between our mediator and our second moderator, if that interaction terms is significantly related to our dependent variable. And you'll see that there are three different interaction terms here. We have interaction term number two, which is engagement times task complexity. We have interaction term number three, which is meaningfulness times organization identification. And we do have interaction term number four, which is meaningfulness times task complexity. Which one should we be looking at here? We should be looking at interaction term number two. Job engagement, and task complexity, and that effect on job performance. If we look here, we do find that interaction term number two has a significant relationship with performance. P is less than 0.05. The other two interaction terms are controls in this model. We are partialling out the effects of these other two interaction terms on our dependent variable, task performance. And down here we have the covariance matrix for all of these coefficients. So, yeah, this gets a little more complicated now. And this is why I said at the beginning that we should look for, we should aim to a parsimonious model. So, dual stage models are cool to run, but it can get a little bit more complex. And why? Let's take a look at the output here. So, remember, we are looking at the indirect effect of job meaningfulness on performance via job engagement. And we are also testing if that indirect effect changes based on values of organization identification and task complexity. So here, we have our mediator, and here we have different levels of our first mediator or moderator, I should say, and our second moderator, organization identification and task complexity. When we look at high levels of organization identification, we do find a significant indirect effect for those individuals who are at the mean, or perceived complexity as low. So those individuals who are high organization identification, and to see task complexity as being low, the indirect effect of meaningfulness on performance via engagement is significant. There is no zero in the confidence interval. For those individuals who are low on organization identification, well, we don't have a significant relationship or indirect effect, I should say. Because regardless of the level of task complexity, zero, it is in the confidence interval of our bootstrap analysis. So, if you are low in organization identification, it doesn't matter the level of task complexity. If you are high in organization identification, yeah, if task complexity is not too complex, if you will, the indirect effect of meaningfulness on performance via job engagement is significant. In this final session of our second workshop on conditional indirect effects, we covered dual stage models. Next, we will be talking about, we have a different workshop on multilevel analysis. And we'll be looking at the relationship of independent variables on dependent variables that are nested within departments, within leaders, within organizations. So we needed to take that nesting into consideration. I hope to see you in our third workshop, on multilevel analysis.