Re: st: brant test

["Michael I. Lichter" <mlichter@buffalo.edu>]

Katrin,

I agree with Richard that you're probably spreading your data too thin. 
I suggest you approach the problem by figuring out why you got the 
original error message, "not all independent variables can be retained 
in all binary logits".

If some variables cannot be retained, it is probably because there is 
redundancy. You should be able to identify the problem by performing 
each of the binary logits. E.g., if your dependent variable Y takes on 
the values 1, 2, and 3, do something like this:

y12 = y == 2 if inlist(y, 1, 2)
y23 = y == 3 if inlist(y, 2, 3)
logit y12 x1 x2 x3
logit y23 x1 x2 x3

Once you've identified the problem variables, you will have to make some 
decisions.

Michael
*
*   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/