Moderated Mediation - Interpreting ou... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 KJ Research posted on Wednesday, November 12, 2014 - 4:34 am
Hello, I have run a moderated mediation model using the syntax from the UG - Ex 3.18. I have produced the plots, saved the data from these plots, and am able to interpret this output fine.

Is it also possible to ascertain from the .out file whether the overall indirect effect is significantly moderated? I.e., something comparable to the 'Index of Moderated Mediation' that you get from running PROCESS in SPSS?
 Linda K. Muthen posted on Wednesday, November 12, 2014 - 11:30 am
How is the Index of Moderated Mediation defined?
 KJ Research posted on Wednesday, November 12, 2014 - 11:48 am
"The slope of the line relating the indirect effect to the moderator is the index of moderated mediation...An inference as to whether this index of moderated mediation is statistically different from zero is a formal test of moderation of the indirect effect by the moderator in the model"

Is this enough info? I'm happy to send a copy of the PROCESS output if that helps to clarify.
 Bengt O. Muthen posted on Friday, November 14, 2014 - 10:53 am
The Mediation page on our website has the document

Loop plot for ex 3.18

where you find an example where the X->M relation is moderated by Z. Here, the definition of the indirect effect is

indirect = b*(gamma1+gamma2*mod);

where mod assumes different Z values. From this perspective, I don't understand the SPSS statement you quote: "The slope of the line relating the indirect effect to the moderator is the index of moderated mediation.." -- does this refer to gamma2 or to b*gamma2, or something else? There must be a formula somewhere. Is this Index published?
 KJ Research posted on Friday, November 14, 2014 - 1:10 pm
Many thanks for your reply.

I have the output for the plot and can interpret this fine for different values of the moderator. I'm wondering if the .out file can tell you whether the 'overall' indirect effect/mediation is significantly moderated (before looking at the breakdown at low/med/high levels of the moderator)?

The formula for the PROCESS model is given in the table at the top of the third page of this pdf (I've been looking at Model 8 on PROCESS): http://www.afhayes.com/public/docaddendum.pdf
 Bengt O. Muthen posted on Saturday, November 15, 2014 - 2:20 pm
The Hayes paper that this refers to shows the Mplus input for getting this index on page 51. In terms of our Loop plot for ex 3.18 document I pointed to where the X->M relation is moderated by Z and the definition of the indirect effect is

indirect = b*(gamma1+gamma2*mod);

the Hayes' index of moderated mediation is obtained as the NEW parameter

index = b*gamma2;

On page 51 this is referred to as "a3b1".
 KJ Research posted on Sunday, November 16, 2014 - 6:55 am
Thank you so much for your help with this - it all makes sense now.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: