Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Still gllamm


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu, narazani@econ.unito.it
Subject   Re: st: Still gllamm
Date   Mon, 20 Sep 2004 08:43:14 -0400

On Mon, 20 Sep 2004 14:22:25 +0200, Edlira Narazani
<narazani@econ.unito.it> wrote:
> 1. I'm using gllamm as follows and at the end of iterations I dont get t-values
> or standard errors. Therefore I cant use it without estimation statisctics.
> 2. when I want to include some random effect (33 firms in 23 years follow some
> choices which take value 0, 1,2)I have done in this way. At the end of
> iterations I get the note: Hessian matrix is not symmetric.maximization did not
> converge.

Both of those are the evidence of difficulties with maximization. Try
-difficult- -- I think -gllamm- supports it without documenting. I
also thought of tightening the convergence criteria with -gtol()-
option, but it is not supported by -gllamm-.

Otherwise, you would have to admit that your model is misspecified and
not supported by data, so you would have to rethink it.

There is still a possibility that there are some minor bugs in
-gllamm-, in which case you should address Sophia Rabe-Hesketh and ask
if that is something that -gllamm- is responsible for (as opposed to
-ml-).

-- 
Stas Kolenikov
http://stas.kolenikov.name
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index