Correlated residuals in LCA PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Daniel Stegmueller posted on Tuesday, July 15, 2008 - 1:14 pm
If one is using categorical indicators in a latent class analysis, the 'a with b' statement in the overall part of the model does not work.

Is there any other way to achieve local nonindependence, besides the way described in Ex.7.16 of the user guide?


Thank you!
 Linda K. Muthen posted on Tuesday, July 15, 2008 - 1:45 pm
That is the only way to specify this relationship.
 D. Stegmueller posted on Tuesday, July 15, 2008 - 2:56 pm
Thank you, Linda.

Did I specify my model correctly?
One variable 'a' is correlated with variables 'b' and 'c', so i put :
%overall%
f by a b c;
f@1;
in the Analysis part.

The results suggest model misfit (LR=28(16)=.027) while the results of my colleague (using different software) are satisfactory.


Thanks,
Daniel
 Linda K. Muthen posted on Tuesday, July 15, 2008 - 5:26 pm
You would need two f BY statements for two residual covariances.
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: