Path Analysis Fit PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Daniel Lee posted on Sunday, May 12, 2013 - 8:27 pm
Hello,

I just ran a very simple mediation model (1 IV, 1 Mediator, 1 DV). The sample size is 879. The fit indices came out like this:

RMSEA: 0.000
90% CI for RMSEA: 0.000, 0.000
TLI: 1.000
CFI: 1.000
Chi-Square Test of Model Fit
Value: 0.000
Degrees of Freedom: 0
P-Value: 0.0000

I'm not sure if I executed the code wrong, but was VERY surprised to see this kind of fit...

Here is my input:
Title:
Model 1 BGCP 2013
Data:
file=bgcp2013_aim1.dat;
Variable:
names = disc hope anxi;
usevariables = disc hope anxi;
Analysis:
estimator=ml;
Model:
anxi on hope disc;
hope on disc;
Model Indirect:
anxi ind disc;
Output:
stand;

Thank you so much for your help Dr. Muthen. I would be happy to send you my code and output if necessary.

DL
 Linda K. Muthen posted on Sunday, May 12, 2013 - 8:34 pm
Model fit cannot be assessed for a model with zero degrees of freedom. The model is saturated.
 Daniel Lee posted on Monday, May 13, 2013 - 12:42 pm
Thank you Dr. Muthen!

I had one question about moderated mediation. Using Haye's technique + code (for full mediation, when path "a" and "b" are both moderated by a certain variable), the code indicates that you should center the moderator at conditional value of interest when probing interactions. I'm wondering if I center my moderator in the data set and use the standard deviation as my conditional value of interest? In other words, I'm stuck in choosing the condition value of interest for the moderator. Insight on how you select this value would be very helpful!

In addition, if I center the moderator, it seems that you should center the IV and mediator, as well. Would you say that's correct? Or, is it okay to only center the moderator in this case?

Thank you so much!

DL

Thank you so much Dr. Mut
 Daniel Lee posted on Monday, May 13, 2013 - 12:43 pm
*Dr. Muthen

*Apologies for misspelling your name in the above post.
 Bengt O. Muthen posted on Monday, May 13, 2013 - 8:31 pm
It helps to center the moderator as well as standardize it so that +1 is 1 SD above the mean and =1 is 1 SD below the mean. Those could be values of interest to condition on. Centering other IVs is not necessary but doesn't hurt.
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: