Dear all,
I want to run a kind of "partial" hausman test to compare two estimates. The point is that to have unbiased coefficient I have to include dummies, but I think that theses dummies make the hausman test rejecting the null hypothesis ("no systematic difference between the coefficients").
More precisely, I have 10 explanatory variables + 5 dummies. Then, what I want to do is to run a hausman test that would only compare the 10 coefficients associated to the 10 variables (but after an estimate including the 10 variables+the 5 dummies), without taking into account any correlation with any of the 5 dummies in the variance-covariance matrix. Do you know any mean to do this ?
Best regards
Christopher
*
* 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/
