Re: st: ICC in crossed effects models

[rgutierrez@stata.com (Roberto G. Gutierrez, StataCorp)]

Elan Cohen <cohened@upmc.edu> writes:

> I just read the following thread regarding crossed effects models:

> http://www.stata.com/statalist/archive/2009-04/msg00376.html

> I'm linking to that thread since my data & model are almost identical.

> My question is regarding intraclass correlations for this model. Am I right
> in thinking that the random-effect under _all:identity, var(R.year) is the
> variance at the _year_ level, and the next random-effects parameter,
> NAME:Identity, var(_cons) is the variance at the _NAME_ level?

> If so, is it correct to calculate the ICC at the year level as:

> var(R.year)/(var(R.year)+var(_cons)+var(Residual) ?

> Does it even make sense to think of ICCs in a crossed model?

It does makes sense and your calculation is correct, provided that the
two observations are in the same year but do not have the same "NAME".
If they did also share the same value of NAME, then you would add
var(_cons) to the numerator.

--Bobby
rgutierrez@stata.com
*
*   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/