Dear Stata Listers,
I am using STATA/SE 9.2 and running a simple random
effect logit model with a combination of discrete and
continuous independent variables. The dependent and
discrete independent variables are all dichotomous.
When I run it, after several iterations it says "not
concave" and keeps iterate with no results. Can
anybody provide any suggestions to make it work.
Thanks.
Pallab
Sometimes adding the -difficult- option will solve convergence
problems. Alternatively, you may need to try modifying your model a
bit, perhaps dropping one or more vars. See -help maximize- for more
details and other options.