MODEL statements in MCFA PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 Scott J. Peters posted on Thursday, February 12, 2009 - 4:10 pm
I am conducting a multi-group CFA (MCFA) in order to evaluate measurement invariance. I'm proceeding in order from equal form to equal factor loadings to equal indicator intercepts, etc. However, I'm unsure that I am correctly programming the MODEL statement in order to test each parameter correctly. I know there are to be two model statements (one general and one specific) for a two group comparison. However, I'm not clear on how you model the two statements and how those statements change to test for different parameter invariance.

Here is some code to test equal factor loadings:

GROUPING IS FR (0=No 1=Yes);
MODEL: f1 by H1@1*H3*H8*H9*H10*H13*H16*H17;
f2 by H2@1*H5*H6*H7*H11*H12;
H6 WITH H12;
MODEL YES: f1 by H3*H8*H9*H10*H13*H16*H17;
f2 by H5*H6*H7*H11*H12;
H6 WITH H12;

Is this testing an equal model for yes/no? Can you explain the differences in the MODEL statements when testing for different invariant parameters?

Thank you so much!
 Linda K. Muthen posted on Thursday, February 12, 2009 - 6:59 pm
The Topic1 short course handout on the website gives inputs for testing for measurement invariance and population heterogeneity. If the handout is not sufficient, there is also a video available that covers Topic1.
 Scott J. Peters posted on Friday, February 13, 2009 - 12:13 pm
Great! Thanks!

Because MEANSTRUCTURE is now the default in multi-group, do I have to include NOMEANSTRUCTURE when testing equal form and equal loadings in measurement invariance testing?
 Linda K. Muthen posted on Friday, February 13, 2009 - 2:16 pm
No, this is not necessary. They are held equal as the default. When you relax the factor loadings, the intercepts remain constrained.
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