<>
Dimitris said
I am a new stata user, so excuse me for my ignorance...I am
using panel data and I want to do my descriptive analysis and my
exploratory analysis having (of course) the same number of
observations. I first rum my prbot model and then I make tables with
means using e(sample in my "if" statement and everything goes well.
However when I do crostabs (using ta or table) the number of
observations is different (higher to be exact). I have observed, while
making the means tables, that when I dont include in my descriptives
all the variables used in the probit model then the sample size
changes. How can I keep the sample size the same without deleting the
missings?
sysuse auto
probit foreign rep78 price mpg weight // drops 5 obs due to missings
on rep78
g byte want = e(sample)
tab mpg foreign if want
su displacement turn if want
etc.
Kit Baum | Boston College Economics & DIW Berlin | http://ideas.repec.org/e/pba1.html
An Introduction to Stata Programming
| http://www.stata-press.com/books/isp.html
An Introduction to Modern Econometrics Using Stata | http://www.stata-press.com/books/imeus.html
*
* 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/
