st: Modeling outcomes with xtgee

[Ignacio Pardo <pardo.nacho@gmail.com>]

Hi everybody... I need some help...

I'm trying to use panel data analysis to model continuos, ordinal and
binary outcomes (it's a before-and-after study, no control group) ...
I read that xtgee would allow me  to do so (through different -family-
and -link- commands), so I' trying to use it (actually, adding gllamm
for the case of ordinal dependent variables)


For GEE, I'm including, in the same model, as predictors:

a) invariant variables (sex, race...)
b) time varying variables
c) a dummy for year ("after"), in order to see the averaged change in
that outcome and
d) the initial outcome value, in order to account for the baseline

Are these decisions right, in your opinion?

Sorry about my english, I hope everyone understands...
Thank you so much,
Ignacio
*
*   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/