st: nbvargr for count data

[Laurie Molina <molinalaurie@gmail.com>]

Dear statalist,

I am using the command nbvargr (Philip B. Ender UCLA, Office of
Academic Computing) as in
(http://www.ats.ucla.edu/stat/stata/library/count.htm ). It graphs the
distribution of the count data in a data set with the correspondig
negative binomial distribution and poisson distribution (with the
parameters resulting from the data set) to see which distribution fits
the data better.

My question is : Why is it useful to do such comparision if the
relevant distribution to get efficient estimators  when running a
poisson or negative binomial regression is the conditional
distribution and not the univariate distribution of the dependent
variable ?

Thank you all as always!

LM
*
*   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/