Multilevel med. model w. FIML not con... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Gerben posted on Thursday, March 02, 2017 - 7:22 am
Hi,

I am estimating a multilevel mediational model with a fair amount of missing data.

Using the default FIML method to handle the missing data (with the required monte carlo numerical integration) the model will not converge, regardless of number of iterations, integration points, etc.

This is the error message:

THE ESTIMATED BETWEEN COVARIANCE MATRIX COULD NOT BE INVERTED. COMPUTATION COULD NOT BE COMPLETED IN ITERATION 2. CHANGE YOUR MODEL AND/OR STARTING VALUES.

The model converges if I use listwise deletion and also if I run a single level model where the hierarchy is ignored.

Do you have any advice regarding this error or alternative methods to handle the missing data in Mplus (apart from listwise deletion)?
 Bengt O. Muthen posted on Friday, March 03, 2017 - 5:09 pm
Perhaps you don't have enough integration points. Check if TECH8 shows any negative ABS changes which would suggest more points.

As an alternative you can try Estimator = Bayes.
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: