Re: st: RE: help needed for fixed effects Negative Binominal regression model

[Scott Cunningham <scunning@gmail.com>]

> > Second, I have no experience in STATA. If anyone has experience using
> > fixed
> > effects Poisson or negative binominal model, can you share some of  
> > the
> > syntax
> > with me? I checked the manual and it looks it's not too hard, but  
> > any help
> > will
> > be appreciated.
I've been estimating a Poisson with fixed effects, and one thing you  
may want to know is that the -xtpoisson- does not have a way of  
correcting the standard errors for overdispersion.  What you can do  
is estimate empirical standard errors through bootstrapping in - 
xtpoisson- by:

. xi: xtpoisson depvar indepvar i.year, fe i(id) vce(boot)

And this will run -xtpoisson- on the data, then run a bootstrapping  
to create the estimated robust standard errors (clustered on id).   
But, I later discovered, through someone else on the listserv, how to  
estimate the robust standard errors the way Cameron and Trivedi (2004  
or 2005 I forget which - MICROECONOMICS book) show.  You have to use  
the -poisson- rather than -xtpoisson- like so:

xi:poisson depvar indepvar i.year i.id, robust

This will estimate poisson with fixed effects, and correct for  
overdispersion.

sc

*
*   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/