Statalist The Stata Listserver


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

RE: st: RE: Pseudo R2 for xtlogit


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: Pseudo R2 for xtlogit
Date   Wed, 9 Aug 2006 02:01:59 +0100

This may well be what is wanted in this case, 
but I don't think it helps to agree and disagree 
at the same time on the general point. The easiest 
valid statement in my view remains that in a large 
and multidisciplinary literature there just
is not anything like a consensus. 

My samplings of the literature indicate that many 
(groups of) analysts prefer Paul's interpretation,  
but at the same time many others don't. For example, 
many others regard percent explained or, better, 
correlation squared as the primary meaning. So, 
it would be dangerous to infer that a particular 
sense is implied unless you knew that a particular
tribe mostly follow that. 

Nick 
n.j.cox@durham.ac.uk 

Paul Millar
 
> As Nick indicates, there are many versions of the pseudo-R2.
> However, if mentioned without qualification (like in the standard
> output from logit models), it refers
> to the improvement in the log-likelihood accomplished by the 
> model and can be
> displayed after any model by typing:
> 
> display "pseudo R2=" (e(ll_0)-e(ll))/e(ll_0)
> 
> - Paul
> 
> At 03:52 PM 08/08/2006, you wrote:
> >Two answers, one negative and one positive.
> >
> >There is no such thing as "the" pseudo-R^2.
> >
> >But you can concoct one of your own designing.
> >
> >In an up-to-date Stata,
> >
> >. search r-square
> >
> >for information.
> >
> >Incidentally, Carter Rees suggested
> >
> >. search xtlogit postestimation
> >
> >which is good advice in general, but,
> >it seems, not relevant for this particular
> >question (please correct me otherwise).
> >
> >Nick
> >n.j.cox@durham.ac.uk
> >
> >Bégin Karine
> >
> > > I estimated a logit model with random effects (xtlogit, re)
> > > with Stata 9.  I would now like to know the value for the
> > > Pseudo-R2, but I can't find it anywhere in the output
> > > produced.  I tried to look at the saved estimation results
> > > using the ereturn command, but there are as well I find no
> > > sign of a Pseudo-R2 statistic.  Is there a way to get it in Stata?

*
*   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