st: Hausman test for -xtlogit- using -suest-

[Andreas Schiffelholz <Andreas.Schiffelholz@gmx.de>]

Hi everyone,

I try to do the Hausman test to check if I can use a random effects 
xtlogit model or if I need to estimate my model using a fixed effects 
model. The problem is, that  receive the following result:

   Test:  Ho:  difference in coefficients not systematic
                 chi2(17) = (b-B)'[(V_b-V_B)^(-1)](b-B)
                          =      174.43
                Prob>chi2 =      0.0000
                (V_b-V_B is not positive definite)

-help hausman- recommends in this case to calculate a generalized 
Hausman test by using the suest command. I tried this using the 
following command:
xtlogit varlist, re
estimates store re
xtlogit varlist, fe
estimates store fe
suest     fe re

If I run this command I receive the following result:

suest    fe re
unable to generate scores for model re
suest requires that predict allow the score option
r(322);

Is there a mistake from my side or is the -suest- command not able to 
handle -xtlogit-?
Is there an alternative method to do the Hausman test?


Thanks a lot!
Andreas
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/