st: gllamm error "something went wrong in comprob3"

[Anje Van Berckelaer <anje@mail.med.upenn.edu>]

Hello all,
I'm using the gllamm function to do a multilevel logistic regression 
with survey weights.

I am getting an error message and am not sure what could be going wrong 
- it tells me that "can't get correct log-likelihood: -241633.32 should 
be -241586.39
something went wrong in comprob3"
Below is the output.  Anyone have any thoughts as to what's going on and 
how to fix it?


. gllamm NOTREFS chc findex, i(fipsnum statenum) link (logit) family 
(binom) pweight (weight) nip (5) adapt

Running adaptive quadrature
Iteration 0:    log likelihood = -251744.82
Iteration 1:    log likelihood = -246950.97
Iteration 2:    log likelihood =  -244570.1
Iteration 3:    log likelihood = -244042.65
Iteration 4:    log likelihood = -242877.62
Iteration 5:    log likelihood = -242649.64
Iteration 6:    log likelihood = -242271.38
Iteration 7:    log likelihood = -241960.33
Iteration 8:    log likelihood = -241838.02
Iteration 9:    log likelihood = -241770.01
Iteration 10:    log likelihood = -241742.87
Iteration 11:    log likelihood = -241737.78
Iteration 12:    log likelihood = -241734.43
Iteration 13:    log likelihood = -241734.13
Iteration 14:    log likelihood = -241733.85
Iteration 15:    log likelihood =  -241733.4
Iteration 16:    log likelihood = -241733.63


Adaptive quadrature has converged, running Newton-Raphson
Iteration 0:   log likelihood = -241607.19  (not concave)
Iteration 1:   log likelihood = -241605.88
Iteration 2:   log likelihood = -241602.91
Iteration 3:   log likelihood = -241591.86
Iteration 4:   log likelihood = -241586.58
Iteration 5:   log likelihood = -241586.39
Iteration 6:   log likelihood = -241586.39
can't get correct log-likelihood: -241633.32 should be -241586.39
something went wrong in comprob3

--
Anje Van Berckelaer, MD, MSHP
Robert Wood Johnson Foundation Clinical Scholars Program
University of Pennsylvania
office 215 573 2583
fax    215 573 2742
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/