|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
Evaluating trajectories with categori... |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
Daniel posted on Thursday, January 12, 2006 - 7:16 am
|
|
|
Hi Linda and Bengt, I hope you had a nice new year. I am presently writing up the results of an analysis for which I conducted GGMM, with repeated ordered-categorical outcomes. My BIC suggests two trajectories, and I agree from a substantive point of view. However, is it feasible to employ the LMR LRT to evaluate data based on categorical outcomes in GGMM? On a side note, I freely estimated my within trajectory continuous latent varialbe variances (slope and intercept), and employed your recommendations for a more thorough analysis (including starts=100 10, and STITERATIONS=20) and used algorithm=integration. The analysis took nearly five hours to run! I am becoming a very patient person thanks to multivariate modeling, so one might say that multivariate modeling is akin to Yoga in teaching one patience ![:-)](http://www.statmodel.com/discussion/clipart/happy.gif) |
|
|
Yes, LMR LRT can be used with categorical outcomes. Ask for TECH11 in the OUTPUT command. |
|
Daniel posted on Friday, January 13, 2006 - 6:43 am
|
|
|
Thanks Linda |
|
Back to top |
|
|