Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: pseudo R-squared
From
Richard Williams <[email protected]>
To
Gao Liu <[email protected]>, [email protected]
Subject
Re: st: pseudo R-squared
Date
Mon, 04 Apr 2011 00:59:15 -0500
At 10:46 PM 4/3/2011, Gao Liu wrote:
Thanks, Richard, for your detail explanation. That's very helpful. One
more question: how to calculate the count R-squared (#correct/total
obs.) of partial proportional ordered logistic regression. I know how
to calculate the count R-squared of logistic or ordered logistic
regressions, but the calculation does not seem to work for partial
proportional ordered logistic regression.
Thank you very much.
I am not sure how to do it for an ordered logit, so if you know how
then you are one step ahead of me. I would think the gologit2
procedure would be the same though. With gologit2, you can compute
the predicted probabilities of each outcome for each case. I suppose
you would then have some criteria for saying the prediction was
"correct", e.g. if outcome 3 had a predicted probability of .38 and
that was the highest predicted probability and the observed value
actually was 3, then you would say it was classified correctly. So in
short, my guess is that you compute the probabilities for each case,
then generate another var indicating the prediction was correct or
incorrect for that case, and then compute your stat. But again, that
is just speculation on my part since I don't know the actual
formula/procedure you are talking about.
-------------------------------------------
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/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/