Hi, I would like to use Chamberlain's approach to estimate an equation. The model I am estimating is a linear model y = bXit + ci (ci is the individual specific effect). My concern is about the specification of ci. I have followed his recommendation for the assumed determination of ci in the Probit model case i.e. that ci is a linear function of Xit. I have specified ci = linear function of the means of the Xits. So basically it is a random effects estimation with the mean of the Xits as additional regressors. The results I get are remarkably similar to the estimates from the usual fixed effects model.
I would like to know from anyone familiar with Chamberlain's approach if it is usual in the linear case to specify a ci as a linear function of the Xits as is usually done in the non-linear case and if there is a better way of specifying ci.
Many thanks
*
* 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/
