Two Variables without Covariance PreviousNext
Mplus Discussion > Missing Data Modeling >
Message/Author
 Fabian Wolff posted on Thursday, March 08, 2018 - 2:50 pm
Hello,
I would like to calculate a SEM with a data set in which two predictors have no covariance, because half of the sample answered to the one item, whereas the other half answered to the other item. There are a lot of other variables in the model with values for all participants. I got a solution, but most of the fit indices are not computed. The warning/error messages are the following:

THE STANDARD ERRORS FOR H1 ESTIMATED SAMPLE STATISTICS COULD NOT BE COMPUTED. THIS MAY BE DUE TO LOW COVARIANCE COVERAGE. THE ROBUST CHI-SQUARE COULD NOT BE COMPUTED.

THE MODEL ESTIMATION TERMINATED NORMALLY

THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES MAY NOT BE TRUSTWORTHY FOR SOME PARAMETERS DUE TO A NON-POSITIVE DEFINITE FIRST-ORDER DERIVATIVE PRODUCT MATRIX. THIS MAY BE DUE TO THE STARTING VALUES BUT MAY ALSO BE AN INDICATION OF MODEL NONIDENTIFICATION. THE CONDITION NUMBER IS -0.436D-17. PROBLEM INVOLVING THE FOLLOWING PARAMETER: Parameter 168, XXX WITH YYY (equality/label)

THE ROBUST CHI-SQUARE COULD NOT BE COMPUTED.

Is there a way to get the fit indices?
Thanks a lot!
 Bengt O. Muthen posted on Thursday, March 08, 2018 - 3:09 pm
Try a 2-group analysis where the groups are formed by the two halves of the sample. Each group has only one of those 2 predictors. That also makes it easy to test if the effects of the 2 versions of the predictor are the same.
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: