st: Adjusting sample to observations used in regression

[Christian Schroetel <christian.schroetel@googlemail.com>]

Dear statalist-users,

I've got a total sample of around 37k observations. Those are
unbalanced, though, meaning not every variable is filled well.
Furthermore, I've got different dynamic panel data models (with lagged
dependent variable) with different omitted variables, so I get models
with different numbers of observations due to listwise deletion. In
order to make regression coefficients comparable, I'd like to
standardize the variables for each model's sample.
Now, when I use drop if !e(sample) I have exactly those observations
left but if I run the regression again, I get less observations in
that regression. That's most likely due to mainly using lagged
explanatory variables in the model.

My question now is: Is there a way to drop observations from the
dataset so that I get the same number of observations in the
regression like before?

I hope, everyone's understanding my problem and thank you in advance.

Best regards

Christian
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/