Different SEs for SAS vs Mplus PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Jennifer Rose posted on Tuesday, July 22, 2008 - 2:09 pm
Hi,

I'm running a random effects ZIP regression model to test the effect of an intervention (2 levels) on number of days in treatment. I ran the analysis using SAS Proc NLMIXED and then using Mplus. I get the exact same regression coefficients, but the standard errors differ considerably between SAS and Mplus. Do you have any suggestions for why I may be getting different standard errors?

Thanks,
Jen Rose
 Bengt O. Muthen posted on Tuesday, July 22, 2008 - 2:39 pm
One reason may be if Mplus uses MLR SEs and SAS uses ML SEs.
 Jennifer Rose posted on Tuesday, July 22, 2008 - 4:50 pm
That's it - thanks! The question I have now is which estimation method is preferable for a random effects ZIP model?

Jen
 Bengt O. Muthen posted on Wednesday, July 23, 2008 - 3:14 pm
We use MLR as the default so that would be our general choice. But only a Monte Carlo simulation study could answer the question of which is best for a given setting.
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: