Moderated factor loadings PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Hillary Schafer posted on Wednesday, September 14, 2011 - 1:18 pm
Hello, I apologize if I have missed the answer to this topic in an earlier post.

I would like to allow a factor loading to be moderated by a continuous variable (age). I've tried labeling the factor loading (i.e., 'b1') and using the MODEL CONSTRAINTS command to regress the parameter on age: b1 on age.

However, I must be doing something wrong because I get the following error:
*** ERROR in MODEL command
Unknown group name CONSTRAINTS specified in group-specific MODEL command.

Advice on how to do this properly would be much appreciated!
 Bengt O. Muthen posted on Wednesday, September 14, 2011 - 1:34 pm
To moderate the loading b1, you shouldn't regress b1 on age (which implies a residual), but instead express a deterministic relationship. In Model Constraint you would use the NEW parameters a and b, say, and say

b1 = a + b*age;

where you have stated

constraint = age;

This estimates a and b, where a is the b1 slope value at age=0 (you may want to center age at its mean for a to be meaningful).
 Hillary Schafer posted on Wednesday, September 14, 2011 - 1:35 pm
Thank you!
 Hillary Schafer posted on Wednesday, September 14, 2011 - 2:21 pm
Hello,
I apologize for another question, but I am still having difficulty with the model that I described earlier. After making the fix that you suggested, age is not recognized in the MODEL CONSTRAINT command. I've tried keeping it on the 'use' command line and removing it, and I've tried using the DEFINE statement to crease a new age variables = 1* the old (centered) age, but neither of these strategies worked. It's not 'NEW' since it's a measured variable. Do you know what I am doing wrong? Thanks again for your help.
(Error message pasted below):
*** ERROR
Unknown parameter label in MODEL CONSTRAINT: AGE_C8
 Bengt O. Muthen posted on Wednesday, September 14, 2011 - 2:45 pm
You should say Constraint=age;

in the Variable command. See V6 UG ex 5.23 for an example.

If you are centering age, you may have to do this in a first step, saving the data, before the run with Constraint=age.

If you are still having problems, send output to support.
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: