Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: gologit2


From   Mike Lacy <Michael.Lacy@colostate.edu>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: gologit2
Date   Thu, 17 Apr 2008 13:04:48 -0600

In an earlier response in this thread,
Richard Williams <Richard.A.Williams.5@ND.edu> remarked:

>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.

I would sound in to support the idea that the Brant test commonly detects departures from proportional odds that are so small as to be uninteresting. In fact, I would suggest as a conjecture that, if the sample size
is large enough to trust the asymptotic p-values from the Brant test, then the sample size is large enough that trivial departures from prop. odds will achieve small p-values. I would suggest instead approaching this specification problem by looking at the relative increase in the pseudo-R^2 value associated with moving to a non-proportional odds model. My own experiments on using such measures to address the related problem of variable choice ordinal logit models shows that one measures is about as good as the next. (see my comment in http://www.stata.com/statalist/archive/2008-03/msg00249.html for a brief discussion of this point and a citation.)
Now, I admit that there is a problem in knowing exactly how big a *relative* change in R^2 (10%?) warrants a more complicated model, but I don't think this is worse than to p-values as the sole arbiter.

Regards,

=-=-=-=-=-=-=-=-=-=-=-=-=
Mike Lacy, Assoc. Prof.
Soc. Dept., Colo. State. Univ.
Fort Collins CO 80523 USA
(970)-491-6721

*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index