st: regressing a dummy variable: equivalence between -xtreg and -xtlogit

[Francesco <cariboupad@gmx.fr>]

Dear Statalist,

I have a large panel dataset and I would like to know the effect of a
dummy variable D on the probability of occurrence of another dummy
variable Y.

Lets assume I xtset my dataset first : -xtset ID and I perform the
following regressions :

-xtreg Y D, fe vce(cluster id2)

-xtlogit Y D, fe vce(robust)

My question is the following : in the fixed effect regression (which,
I know, is not the best approach for regressing a dummy variable) the
value of the parameter associated to D (lets say 0.3) tells me that
the proportion Y increases by 30 percentage points when the dummy
variable D is 1.
How can I obtain the same intuitive interpretation for the results of
the fixed effect logictic regression ? I was thinking of the command :
 margins , dydx(D) predict(pu0) atmeans ? Is it correct ?

Many thanks for your remarks,
Best Regards,
*
*   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/