Statalist The Stata Listserver


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

st: RE: relative importance of covariates in an xtmixed or mle model


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: relative importance of covariates in an xtmixed or mle model
Date   Thu, 19 Oct 2006 10:40:04 +0100

Relative importance of covariates is in 
general highly elusive. Someone who 
worked in your institution until his 
recent death -- Fred Mosteller -- warned
that covariates come as a bunch, not 
as a set of individual variables, in his 
book with John Tukey, "Data analysis and 
regression" (1977). And that's just one 
of several accounts repeating the same basic
idea. 

I know that doesn't answer your question, 
except insofar as you asked for "any advice". 

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

Hillel Alpert
 
> Can someone please advise:
> 
> How can we determine the proportion of variance explained per 
> covariate in a model that is
> based on maximum likelihood estimation?
> 
> The model in hand is generated with xtmixed, and  I wish to 
> determine how much of the model is explained by one of the
> independent variables. The beta coefficient for the model 
> with only that
> variable is 0.02.  The beta coefficient for this variable 
> when the model others variables is about
> 0.01. Would the difference between these coefficients be indicative?
> 
> Any advice is much appreciated.

*
*   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