Second-order CFA growth model PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Seth J. Schwartz posted on Friday, March 14, 2014 - 1:17 pm
I need your help with a second-order CFA growth model. No examples exist in the literature for how to do this. We have five timepoints.

There are problems with "identification" because I am not setting the model up correctly. I have to constrain factor loadings for the primary factors (measurement equivalence). I know how to do this. The question is whether you also have to do this for the second-order factor loadings as well. Growth is then based on the second-order factor.

I have to set the marker variable intercepts to equivalence and to zero across time [V1@0 V2@0 ... V10@0], but do I set the factor intercepts for the first-order latent constructs to zero to match the "marker variable" concept [F1@0 F2@0 ... F5@0], since they are the marker "variables" for the second-order construct.

Do I set the intercept for the second order factor using the usual Mplus convention [SecOrdV1]?

Thanks!
 Bengt O. Muthen posted on Friday, March 14, 2014 - 5:44 pm
Start with the specification for multiple indicator growth on page 681 of the V7 UG. Note that you hold intercepts equal across time for the same item and fix the intercept growth factor mean at zero. Second-order factor loadings should also be invariant over time. Fix the latent intercepts of the first-order factors at zero at all time points. Fix the intercept of the second-order factor at zero and all time points. Hope I said that right.
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: