 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: gologit2
At 11:54 AM 4/14/2008, Verkuilen, Jay wrote:
Maarten buis wrote:
>>If you are unsure, than go through the logic of testing: formulate the
null-hypothesis. <snip> <<
The one addendum I would add is this: If the formal test says reject the
null but the resulting violation is "small", you may want to think twice
about tossing out the proportional odds assumption. Such violations are
often found by capitalizing on chance and wouldn't replicate (instead
you'll find some other violation elsewhere). It may be worth it to
assess these kinds of assumptions on a calibration sample and have a
randomly selected holdout sample for later validation of your model.
My experience is that it is rare to have a model where the
proportional odds assumption isn't violated! Often, though, the
violation only involves a small subset of the variables, in which
case gologit2 can be useful. You might also want to consider more
stringent alpha levels (e.g. .01, .001) to reduce the possibility of
capitalizing on chance. You can also try to assess the practical
significance of violations, e.g. do my conclusions and/or predicted
probabilities really change that much if I stick with the model whose
assumptions are violated as opposed to a (possibly much harder to
understand and interpret) model whose assumptions are not violated.
Finally, while I happen to like gologit2, there are a lot of other
categorical and ordinal models out there that might be worth a look
depending on the problem.
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: [email protected]
WWW: http://www.nd.edu/~rwilliam
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/