Not positive definite error message PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Metin Ozdemir posted on Saturday, January 24, 2009 - 11:02 pm
I am running a measurement model with six latent variables, two of which are second order factors.

I have been receiving "PSI matrix not positive definite" error. I have checked residual variances, none is negative. The latent variable correlations are not over one. I checked the data file, the is no problem with reading the data file.

What may be other reasons, and how can I solve this problem?

Thanks,

Metin Ozdemir
 Bengt O. Muthen posted on Saturday, January 24, 2009 - 11:50 pm
The factor covariance matrix may be not pos def due to the combination of correlations among several latent variables - none has to be greater than 1. This suggests a misspecified model. I would break down the model and analyze its smaller parts to see where the problem occurs.
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: