moderated mediation
This is a topic that many people are looking for. thevoltreport.com is a channel providing useful information about learning, life, digital marketing and online courses …. it will help you have an overview and solid multi-faceted knowledge . Today, thevoltreport.com would like to introduce to you Moderated mediation in SPSS using Hayes Process macro (August, 2019). Following along are instructions in the video below:
hello in this video I provide a general introduction to moderated mediation and demonstrate how to carry out several of these analyses using Andrew Hayes process macro the templates I review in this video are drawn from his tax which provides a highly readable account of mediation moderation and conditional process modeling a copy of this PowerPoint and of the data utilized can be will be made available for download underneath the video description and the PowerPoint actually is going to contain a lot more information than I can really go through in a timely fashion in this video so if you find the video and supporting materials helpful please take time to like the video and share the link with others now before we get started with our demonstrations lets review a few concepts mediation analysis is utilized to test whether a proposed causal effect of X on Y may be transmitted through a mediating variable M the model record the model represents a general causal conception of the relationships among X Y and M below we see a simple mediational model involving three paths and all paths thats a B and C Prime and this model represent the direct effect a one variable on another in this conception path a represents the assumption that X is antecedent to M whereas path B represents the assumption that Pat that M is antecedent to Y the indirect effect or the mediational effect of X on Y is calculated as a product of paths a and B path C prime represents the direct effect of X on Y after controlling for the proposed mediation so within this conceptual framework all paths are unconditional that is the relationships among the variables are not moderated by other variables now standard mediational models assume that the indirect effect of X on Y is constant across levels of other variables nevertheless it may be the case that researchers seek to describe and test whether mediated effects vary vary across contexts groups of individuals or individuals that vary on the independent variable in other words a researcher might be interested in determining whether a mediational process is conditional on other variables this is referred to as moderated mediation Heys processed macro incorporates a variety of model specifications that allow you to test the effects both direct and indirect effects of x on y conditional on a moderator so given the sheer number of models that are available through process were going to focus on just a few of them in this presentation so youll notice that down below just kind of reviewing some of these models we have process model number seven and so in this model we have the moderator W and W is moderating the effect of the independent variable X on M next we have process model eight right here where we have the moderator W moderating the effect of X on the mediator so thats basically path a right here and X on Y which is the C prime path that you see right here so W is moderating both of those two paths next we have Process Model 14 where we have the moderator W moderating the effect of the mediator on Y so its basically having an influence on path B which is what we see right here then we have process model number 5 and in this particular case we have W moderating the the effect the direct effect of X on Y so thats moderating the C prime path Process Model 15 which we have down below we have W moderating the effect of the mediator on Y so thats moderating path B and the direct effect of X on Y which is the C prime path that you see if you look at process number model number 58 we have W moderating both paths a and B so its moderating the effect of X on the mediator and the mediator on Y and then process model 59 you have W essentially adding in the moderation of the direct effect of X on Y which is reflected in the C prime path so you have basically moderation of the paths a B and c-prime so for this presentation Im gonna cover process model seven and eight along with model
58 using examples from Paul Josas discussion of mediated moderation in the original presentation the author carried out a test of mediated moderation and Amos effectively testing whether the moderated effect of rumination on depression is mediated through anxiety he also provided a demonstration of moderated mediation through the use of multi group analysis using Amos and with some different data so nevertheless he stopped short of demonstrating how one might carry out moderated mediation involving a continuous moderator in a state he referred the reader to Hays process macro of the aim of ones analysis is to test for moderated mediation involving a continuous moderator although the macro does work with categorical moderators as well so our first model will include perceived control as a moderator of the path from rumination to anxiety and our second model will include perceived control as a moderator of the previous path and the direct effect of rumination on depression so for our first model we have again we have rumination as a predictor of anxiety anxiety as a predictor of depression and perceived control as a moderator of the effect of rumination on anxiety so lets go ahead and open up SPSS and run the analysis so here we have our data weve got rumination we have perceived control depression and anxiety and these variables actually were already centered or at least several of them were already centered so Im going to show you how to just request centering if you want it anyway but what well do is what to run our analysis were going to go through analyze regression go down to process and open it up Im going to reset this and walk you through the steps so now what well do is were going to add our X variable which is rumination were going to were going to move it over to the x-variable box so were gonna do move it over here then we have our depression variable thats our Y variable and were going to move it to this box right up here so there it is and then were going to move our moderator variable which is our variable W Im going to move this down to the moderator W box down here so sometimes sometimes it drags a little bit better than others and then we have our anxiety variable which is our mediator variable so Im going to move it to the mediator or M variable box right there under model number were going to set this as seven next well click under options and Im going to go ahead and ask for generate code for visualizing interactions then were also going to click on mean Center for construction of products although its not really going to make much of a difference since our variables were already mean centered Im going to click on conditioning values and Im going to use the traditional pick a point approach which is negative 1 standard deviation mean and positive 1 standard deviation for the moderator variable and Ill also ask for pairwise contrast of indirect effects so next well click on continue and then on ok and it takes a couple of seconds to run because its basically performing bootstrapping and so there we go so now we have our output and lets walk through the various pieces so Ill just kind of show you what this is looks like in general and then Ill go into our PowerPoint so the first portion of the output basically contains the regression of the mediator which is anxiety on to perceived control rumination and the interaction between those two variables and we see here that the interaction term was statistically significant suggesting that perceived control moderates the effect of rumination on anxiety next we get the conditional effects of the independent variable rumination on anxiety at different levels of the moderator so you can see right here that we have simple slopes based on the pickup a point approach and negative one standard deviation on perceived control the effect of rumination on anxiety was positive and significant at the mean of perceived control the effect of rumination was positive and significant and at positive 1 R at plus 1 standard deviation of perceived control rumination was again a positive
and significant predictor now we also see in our output that we can we have our code for visualizing the interaction so youll see on the left this actually is what appeared in our output and on the right this is something that I generated based on this code so let me quickly walk you through this so here we are back into our original output and you can see the code that is given right here so what were going to do is double click on the output file and I am actually going to scroll up to this lets say to our point right here and Im going to highlight everything that starts with data list free all the way down to the period at the end of the last line so then Im going to right click and click on copy and then we go to a file you may not see it on your screen but Im clicking on the file tab go to new and then Im going to go to syntax right here so Im going to click on that and the syntax file opens up and Im gonna paste our syntax in so now what Ill do is Ill highlight everything and click the green button to run it so youll notice that it already kind of gives you sort of a plot right here but you cant really draw the lines on this unfortunately and so to get the lines on it what Im going to do is notice that when I click on my my little icon at the bottom of the screen theres now a new data set thats been created which contains levels of our of our independent variable X and our moderator perceived control and then our predicted values on the mediating variable anxiety so now what we can do to plot out our simple slopes is we can go to graphs legacy dialogues I will go to line right here click on multiple define and Im going to move the X variable rumination over to the category axis box Im going to move concern R moderating variable to the define lines by box and Im actually going to click on other statistic and just kind of set that there and then move my outcome variable which is the anxiety variable to this box right here when I click on OK now you can see that I have the simple slopes that you will notice in the video I mean in the PowerPoint so basically looking at our slopes you can see that again this is our perceived control variable right here so this would be our low medium and high levels based on our pick a point approach so you can see that at high perceived control the slope for the relationship between rumination and anxiety is less positive than among those individuals where you have lower perceived control so thats this slope right here okay so moving on with our output you can see that the next portion contains our dependent variable depression and so in this particular case you can see that we have basically the direct effect of rumination and the direct effect of anxiety so both of these coefficients are positive and statistically significant so in this portion of the output we get information on the conditional indirect effect associated with the model so here with this portion of the output we have the index of moderated mediation and this provides an omnibus test of the conditional indirect effect of X on Y so if the null of 0 does not fall between the lower and the upper limit of the 95% confidence interval then we will infer that the indirect effect is conditional on the level of the moderator variable so in this example that seems to be the case where basically perceived control appears to be moderating the indirect effect of rumination on depression now assuming the index of moderated mediation is statistically significant then we may wish to probe the conditional indirect effects so the this portion of the output thats highlighted contains the conditional indirect effects of rumination on Y at minus 1 standard deviation the mean and plus 1 standard deviation on the perceived control variable and you do
in all cases the indirect effects were positive and statistically significant now finally we have the pairwise contrasts for the conditional indirect effects reflecting the indirect effects at different levels of the moderator and so basically what youll youll notice is that we have 95% confidence intervals that allow us to test those contrasts so youll notice that with these confidence intervals theres a lower limit and upper limit so you can see youve got effect 1 versus effect 2 and so these numbers that you have right here are coming from this part of the table above so you can see in all cases the no of zero does not fall between the lower and the upper bound in each of these cases so that basically signifies that all of our pairwise contrasts of the indirect effects are statistically significant so that is further evidence that we have moderation of the indirect effects okay so now lets look at another model so in this case we are using a model number eight right here and so we have perceived control serving as a moderator of the direct effect of rumination on anxiety and rumination on depression so were going to leave basically everything else the way it was just changing the model number and so when we look at our output you can see that the interaction effect of reflecting the moderation of perceived control on excuse me reflecting the moderation of perceived control on the relationship between rumination and anxiety we that is statistically significant so our interaction term is statistically significant again we have the same conditional effects of rumination on anxiety that we saw in our previous model and the same code is presented so were not really going to go into that portion of the output at this point now we have the outcome variable depression and so in this case we have depression being regressed on to rumination anxiety perceived control and then the interaction term reflecting the interaction between rumination and perceived control on depression and you can see that the interaction term is statistically significant so using the pick up pick a point approach you can see looking at our simple slopes all of these simple slopes at our low medium and high levels of the moderator are positive and statistically significant and when we plot out those simple slopes you can see what it looks like down here so you can see basically a very similar pattern that will we had previously when we were looking at the moderation of the effect of rumination on anxiety in this case we still see that at high levels of perceived control the slope is weakest at low levels that perceived control the slope is more positive or stronger once again with our index of meat of moderated mediation we can see that the moderation effect was statistically significant okay so for this model we have model number 58 so were not doing anything different in terms of what were requesting except changing the model number to 58 and so in this particular demonstration we have perceived control moderating the effect of rumination on anxiety and the effect of anxiety on depression so as we have seen before the effect of rumination and anxiety appears to be moderated by perceived control so our interaction term is statistically significant as weve seen before and in this case the interaction of anxiety and perceived control is statistically significant so this is also indicating that the effect the relationship between anxiety and depression is moderated by perceived control now unlike our previous models where we obtained an index of moderated mediation were not getting that in this particular output nevertheless we do see the conditional indirect effect of rumination on depression at low medium and high levels of the moderator are all basically positive and statistically significant and we also see that all pairwise contrast between the conditional indirect effects were statistically significant so this is essentially evidence favoring the notion that perceived control moderates the indirect effect of rumination on depression okay so that concludes our demonstration on the last page of the PowerPoint youll see the various references that I utilized in putting this together so be sure to check them out and I appreciate you watching
tags:
Moderated mediation, Moderated mediation with Process macro, moderated mediation spss, moderated mediation process model 7, moderated mediation process model…
Thank you for watching all the articles on the topic Moderated mediation in SPSS using Hayes Process macro (August, 2019). All shares of thevoltreport.com are very good. We hope you are satisfied with the article. For any questions, please leave a comment below. Hopefully you guys support our website even more.
