--- On Wed, 29/4/09, Allison Milner wrote:
> I am trying to better understand sensitivity analysis
> following my XTGEE estimation for 73 countries. So far, I
> have tried to include alternate control variables, regional
> dummy variables (geographic location), and year dummy
> variables into the analysis. I have also tried to use a
> different correlation structure (eg. ar) in the xtgee and
> also use a FE and RE model. Is there anything other ways in
> which I can conduct sensitivity analysis? Is there something
> I am missing/doing wrong?
You can always add more and different models to a sensitivity
analysis, so the trick is not to try to be exhaustive, because
than you will always fail. Instead you should define a set of
potential problems with your models that you care about and
from that derive a set of models that make reasonable but
different assumptions about these problems. When writing it up
you should make the argument that these models capture a
reasonable range of possible assumptions concerning the problems
you care about.
-- Maarten
-----------------------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
*
* 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/
