st: AW: poisson distribution, transformed variables and implications for xtreg fe or re

["Martin Weiss" <martin.weiss1@gmx.de>]

<> 

Austin Nichols had a couple of thoughts on the -poisson- thing in Chiacgo,
see page 22 of http://www.stata.com/meeting/snasug08/nichols_gmm.pdf
For the "best" transform, also see -h ladder-



HTH
Martin


-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von
Allison.Milner@student.griffith.edu.au
Gesendet: Montag, 2. März 2009 11:43
An: statalist@hsphsun2.harvard.edu
Betreff: st: poisson distribution, transformed variables and implications
for xtreg fe or re

Dear statalist,
I have a query about using the poisson distribution and transformed
variables. I have a panel data set with mortality information (DV) from 74
countries from 1980 to 2006. I have transformed the DV using a square-root
transformation. Previously I have been estimating this DV using xtreg fe, re
and xtabond2. However, someone mentioned to me that I should try and use the
poisson distribution in my estimation techniques. I have tried this using
xtgee but this will not run on some of my clusters (due to missing data?).
Also, I am not sure whether I can run a dynamic model such as xtabond2 using
the poisson distrubution.
I have noticed that other research papers have not bothered to use a poisson
distribution and merely used transformed DV scores. Is it neccessary that I
use a poisson distribution?
Any advice would be greatly appreciated
Thanks
Allison

*
*   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/


*
*   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/