I have a large panel of student-level data over multiple years. I want to
estimate a model that will include both student-level fixed effects and
school-level fixed effects, where there are 800,000+ students and 3,000+
schools. Areg will only allow one level of fixed effects. I thought about
using xtdata to compute differences from student means and then using areg
with school as the "absorb" variable. The manual indicates that running
regress after using xtdata produces the correct coefficients, but not the
correct standard errors. I presume the same would be true for areg. Has
anyone written an .ado file that will compute the correct standard errors
when using areg (or regress) after mean-differencing the data with xtdata?
Tim
Tim R. Sass
Professor Voice: (850)644-7087
Department of Economics Fax: (850)644-4535
Florida State University E-mail: [email protected]
Tallahassee, FL 32306-2180 Internet: http://garnet.acns.fsu.edu/~tsass
