Modification Indices and nonlinear co... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
 Judith Biesen posted on Monday, December 07, 2015 - 4:27 pm
Hi Drs. Muthen,
I'm using Kenny's STARTS model to analyze my longitudinal data, which requires me to impose some non-linear constraints. However, when I use the MODINDICES command, I get the error message that this command cannot be used in conjunction with non-linear constraints.

Are there any other way to obtain modification indices for my data?

Thank you!
 Bengt O. Muthen posted on Monday, December 07, 2015 - 6:03 pm
No, all you can get is 1st-order derivatives but they are not "standardized" as with modindices (they are the numerator of the modindices) so they are hard to interpret. You can also do modindices "by hand", by including an extra parameter at a time and seeing if it is significant and how chi-square (or other fit indices) improve.
 Judith Biesen posted on Wednesday, December 16, 2015 - 12:24 pm
Thank you! I obtained the derivatives by requesting the TECH2 output, however, I'm unsure which set of derivatives to look at as there are several sets of "derivatives with respect to [NU,LAMBDA,ALPHA, THETA,BETA, PSI]" and there doesn't seem to be anything in the Mplus manual on this topic.

Any advice where I can find out what these different Greek letters stand for?

 Bengt O. Muthen posted on Wednesday, December 16, 2015 - 5:44 pm
They are described in the UG chapter 18 under the Output command.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message