Re: st: mlogit iterations not concave

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Mon, 22/3/10, Jessica Bishop-Royse wrote:

> Thank you for the iter(20) option.  I did seem to find
> some of my items that were problematic.  Not sure there
> is anything I can do about it, since I feel like I
> should include the variables that were a problem (maybe
> reconsider what I am using as a reference group).  

If there is anything you can do depends on what the 
problem is. Often it is just a set of variable at one
comparison of categories (often sparse categories) that
is the problem. Which means that you'll have to combine
categories. I would start very simple, e.g. take only
three categories (death of cancer, other deaths, alive)
and try if your model converges, than step by step 
expand your model to predict more categories. For how 
to combine categories, see:
http://www.maartenbuis.nl/example_faq/index.html#recode

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/