Confirmatory Factor Analysis WITH st... PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 Scott Smith posted on Monday, August 20, 2012 - 2:02 pm
We have a 4 item factor and have specified 3 degrees of freedom as we are allowing two of the residuals to covary at the within level. The fit dramatically regresses if we don't include the "WITH" statment (4 df), per the MOD INDICES. Can we report the fit statistics for the 4 item /3df model?
 Bengt O. Muthen posted on Monday, August 20, 2012 - 2:19 pm
Not enough information. Please send to Support.
 Matthew Podlogar posted on Wednesday, February 01, 2017 - 7:39 pm
I believe the following syntax is poor practice, but I am unable to find citation for or against it. Specifically, correlating an indicator with a variable not in the model:

USEVARS are: f1 f2 a1 a2 a3 b1 b2 b3 c1;

f1 BY a1 a2 a3;
f2 BY b1 b2 b3;
b1 WITH c1;

Is this model flawed? or can it be properly interpreted?
 Bengt O. Muthen posted on Thursday, February 02, 2017 - 5:39 pm
You should not have the factors f1 and f2 on your USEV list. The model is ok if you think that c1 correlates only with b1 and not the other 5 observed variables.
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