I was trying to run a multilevel model in R using lmer and my model failed to converge with the default optimizer. Then I used the bobyqa instead of default optimizer and my model converged. Now I need to run the (same) model in STATA and I need some help. Anybody knows how to modify the "optimizer" (not sure if it is called so) in xtmixed in STATA? I read STATA manual and tried all the options but it did not work; my model failed to converge with the default setting (similar to R). Any thoughts or suggestions would be very much appreciated. Thank you!
Asked
Active
Viewed 36 times
1
-
Are you sure it is implemented in Stata? There is no reason to believe that all the software has exactly the same features implemented. Also if you already got the results in R, why duplicate the work in Stata? Problems with optimization often mean problems with model definition, so I'd double-check if your model makes sense for the data. – Tim May 16 '23 at 20:34