Re: st: Reoprob - problem with convergence and superxt

[Maarten Buis <maartenlbuis@gmail.com>]

On Fri, Jun 17, 2011 at 4:13 PM, Andrea Mannberg  wrote:
> I am trying to estimate a random effects ordered probit model of the form
>
> reoprob trust_trade ptreatment treatment  year99 tradeout age sex read write maleadult femaleadult oxen livestock landarea security1 security2 conflict zone , i(hhid) quadrat(12) level(95)
>
> Where the dependent variable has 4 categories. I have tried changing the quadrature, the level, and the tolerance levels (ltol and tol) , nothing helps. Stata is having a problem of getting the "constant only" model to converge.

The constant only model is a very simple model. So, if that does not
converge than that is an indication that there is something seriously
wrong with your model, and it is unlikely that tricks like -superxt-
are going to work in such situations. In those situations you need to
go back to your data and inspect that very carefully and find the
problem. In this case I suspect that your dependent variable has some
very rare categories, which means you'll probably need to merge that
category with an adjacent category.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   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/