Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Re: psedual R2 of mim:gllamm fitted models


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Re: psedual R2 of mim:gllamm fitted models
Date   Fri, 7 May 2010 10:46:05 -0500

My two cents: pseudo-R^2 from logistic regression is flaky (does the
value of 1 indicate perfect prediction? Wait, perfect prediction means
that logistic regression estimator does not converge!). Pseudo-R^2
from a multilevel logistic regression is double-flaky (variance at
what level does it try to explain?). Pseudo-R^2 from a multilevel
model with imputed data is triple-flaky (does it incorporate
imputation variance?).

To compare performance of competing models, you can -gllapred- your
outcomes (preferably in a separate validation sample) and see which
model generates more accurate predictions.

On Fri, May 7, 2010 at 10:35 AM,  <jl591164@albany.edu> wrote:
> Hi all,
> I need to compare which of two sets of variables explain more of the
> variance of the response variable. I use mim:gllamm to fit two multilevel
> logistic regression models containing the two different sets of variables.
>
> Would comparing the pseudo R2 of the two models answer my question of
> which set of varaibles explain more variance? If so, how I can get the
> pseudo R2 of mim:gllamm model. The -trace- option of gllamm can obtain the
> R2 but will not give the R2 in mim:gllamm output. Thanks a lot.


-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
*   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/


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