[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: gologit2
Richard Williams <Richard.A.Williams.5@ND.edu>
RE: st: gologit2
Wed, 16 Apr 2008 10:24:50 -0500
At 04:01 AM 4/16/2008, Maarten buis wrote:
Thanks much for doing this Maarten. I'll note that there are at
least two other ways to test the proportional odds assumption in
Stata. First, you can use the -omodel- command, available from
SSC. To test this, I believe you drop Maarten's ologit, brant and
return commands and replace them with
I don't know of any such simulation in the literature, but you can use
-simulate- to do it yourself. In the example below I draw at random
from a population in which the proportional odds assumption holds, and
than use the Brant test to test that assumption. For this I use the
-brant- command which is part of the -spost- package (see:
-findit spost-). I than repeat this 10,000 times. If the Brant test
works then it should reject the null hypothesis in 5% of the draws.
omodel logit y `x'
return scalar p = chi2tail($S_2, $S_1)
I'm running that right now and, assuming I've done it right, I'll report the results tonight or tomorrow.
The other approach is to use gologit2, which is also interesting because the Likelihood Ratio test that follows is (as far as I can tell) identical to the test SPSS PLUM reports for parallel lines. I think the code is
gologit2 y `x', pl store(m1)
gologit2 y `x', npl store(m2)
lrtest m1 m2
return scalar p = r(p)
gologit2 is no speed demon so this may take a few months to run!
I have no particular reason for believing that any of these approaches is clearly superior or inferior to another, but it would be interesting to find out if there are clear differences.
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
* For searches and help try: