Andrea Molinari wrote:
> This is probably a very basic question, but I couldn't find the answer (or
> the correct syntax) for it.
> I am trying to run an ANOVA command on my regression, to see how much of
> the
> variation in the dependent variable do my fixed effects explain.
> Any suggestions will be more than welcomed!
One 'back of the envelope' technique that you could use to achieve this
was proposed by the late Michigan sociologist Donald E. Stokes (1974),
which I hope I've adapted correctly in the Stata context:
(1) EITHER run your panel regression using -xtreg- and then standardise all
of your coefficients in the usual way OR run a pooled OLS regression
with the -beta- option;
(2) sum up the beta-weights;
(3) divide the sum of fixed-effect beta-weights by the total sum of beta-
weights. This is the proportion you want.
I hope this helps.
CLIVE NICHOLAS |t: 0(044)191 222 5969
Politics |e: [email protected]
Newcastle University |http://www.ncl.ac.uk/geps
Stokes DE (1974) "Compound Paths: An Expository Note" AMER J POL SCI
18(1): 191-214.
*
* 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/
