st: hausman test after xtlogit fixed and random effects?

[Leiv Marsteintredet <leiv.marsteintredet@isp.uib.no>]

Hi, all.

I am using panel data and my dependent variable is dichtomous. I am  
relatively new to stata and panel data, and have not found any  
solution to the following questions:

Is it possible to use Hausman to test for the use of fixed vs. random  
effects?

I have looked through several sources and find that some use hausman  
with no caveats, while others (some on statalist) argue that since  
fixed effects with xtlogit is conditional fixed effects, hausman  
cannot be used.

If Hausman cannot be used to test the approriateness of a random  
effects xtlogit vs. fixed effects, are there any other ways to do  
this for random effects xtlogit? (or any ways of testing the  
assumptions in xtprobit?)

I'd like to be able to use random effects since some of my groups  
have all negative outcomes on the dependent variable and fall out  
using fixed effects.

Thanks,

Leiv



*
*   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/