Parameter estimates in TYPE=COMPLEX PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Xu, Man posted on Wednesday, October 29, 2008 - 11:36 pm
Hello, I found some small differences between the model parameter estimates from a typical SEM and a SEM model with TYPE=COMPLEX. I thought the only difference should be between standard errors. Is this normal? or could there be something wrong with the models I specified? Thank you!
 Xu, Man posted on Thursday, October 30, 2008 - 3:18 am
Sorry, I was comparing results from Amos and results from the same model with type=complex in Mplus. I re-run the analysis in Mplus without type=complex and the parameter estimates are exactly the same. I guess it's normal that the result from Mplus is a little different from Amos, is it?
 Linda K. Muthen posted on Thursday, October 30, 2008 - 8:18 am
There may be slight differences in parameter estimates due to different convergence criteria. If you send the two full outputs and your license number to support@statmodel.com, I can take a look at them.
 Xu, Man posted on Wednesday, November 05, 2008 - 8:44 am
Thank you. I am not too worried though. The data in Amos has missing data in it, whereas the data I used for Mplus are multiply imputated datasets. So it should be normal to have some small discrepencies...
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: