On Sat, 02 Oct 2004 22:24:15 +0000, Dana Shills <[email protected]> wrote:
> How do you obtain the residual variance in a model? Im trying to see what
> percentage of total variance is explained by the variables in the model. So
> for instance, when I run:
>
> gllamm y, i(size country) link(ologit) adapt
>
> I obtain
>
> Variances and covariances of random effects
> ------------------------------------------------------------------------------
> ***level 2 (size)
>
> var(1): .05399151 (.04772325)
>
> ***level 3 (country)
>
> var(1): 0.0116962 (0.0113)
> ------------------------------------------------------------------------------
>
> So is residual variance(ie unexplained variance)
> just=1-(0.05399151)-(0.0116962)?
No. There are no natural concepts of total or explained variances
outside of linear regression model. The primary use of the reported
variance is to test the hyporhesis that the variance is equal to zero,
and there is no effect at this level. For your results, this seems to
be the case: neither the country nor the size seem to matter. This is
a good news: you don't have to complicate things and -gllamm- them;
the built-in -ologit- will do.
Besides, those variances add to each other, so the total variance is
_pi^2/6 + 0.054 + 0.012.
By the way, what does the size mean in your model?
--
Stas Kolenikov
http://stas.kolenikov.name
*
* 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/
