Statalist


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

st: Regressions with many fixed effects


From   Alice <aapdm_999@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   st: Regressions with many fixed effects
Date   Tue, 7 Apr 2009 16:24:44 +0000 (GMT)

Dear all,

I am trying to estimate a regression with 4 different sets of fixed effects. The problem is that if I were to compute all fixed effects, each set would consist of about 10.000 dummies which would require me to include about 40.000 dummies in the regression - which is obviously not possible.

I know that I can "absorb" one set of fixed effects using either xtreg, areg, or felsdvreg. However I would still need to include the other 3 sets of fixed effects, which is still a problem.

An alternative would be to demean my variables around the 4 different means. However, as I want to cluster my standard errors, I do not know how I should adjust my variance covariance matrix in this case.

Is anyone aware of any command that would enable me to absorb more than one set of fixed effects? Any other suggestions?

Thanks. Alice.


      

*
*   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   |   What's new   |   Site index