Moderated mediation PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Tony Reed posted on Thursday, April 21, 2011 - 4:12 pm
Hi Dr.Muthen,

I am using SEM to test a moderated mediation model. However, the Mplus did not produce model indices such as CFI, TLI, RMSEA but HO value instead. Is there any way that Mplus could produce the model fit for the structural model of my moderated mediation model?

 Linda K. Muthen posted on Thursday, April 21, 2011 - 5:44 pm
If chi-square and related statistics are not given, they are not defined for the model. There is no way to obtain them.
 Tony Reed posted on Thursday, April 21, 2011 - 6:00 pm
So, you are saying that there is no way to obtain the model fit for the structural model if I do moderated mediation? Or, it is just not possible to obtain such information in Mplus?

I am confused because there is an AMJ paper by Zhang & Bartol (2010, Vol. 53, No. 1, 107128.) in which the authors reported the model fit (CFI, RMSEA) for their structural model which is a moderated mediation one.

 Bengt O. Muthen posted on Thursday, April 21, 2011 - 6:10 pm
If it is possible to obtain such information, Mplus reports it. There may be some reasons for not reporting:

If your model is using XWITH, there is not an overall chi-2 measure of model fit defined yet for ML estimation.

If you don't have XWITH or Type=Random and still don't get a chi-2 fit measure, perhaps your model is just-identified.

If neither applies, please send output to support.
 Tony Reed posted on Thursday, April 21, 2011 - 6:50 pm
Drs. Muthen,

Thank you very much for your explanation!!

 Heike B. posted on Saturday, January 21, 2012 - 8:22 am
I would like to implement a first stage moderated mediation in an observed path model:

A-> B -> C -> D

The first path (A -> B) is moderated by M.

Would the following code be sufficient in MPLUS:

B on A M AxM;
C on B;
D on C;

Or would I have to incorporate the variables of the moderation also in the other paths, i.e.

B on A M AxM;
C on B A M AxM;
D on C A M AxM;

Thanks a lot in advance.
 Linda K. Muthen posted on Sunday, January 22, 2012 - 2:04 pm
I would use:

B on A M;
C on B A M;
D on C A M;
 Heike B. posted on Monday, January 23, 2012 - 9:38 am
Thank you, Linda.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message