It doesn't sound to me as if -rowmean()- is what
you want.
You want to know over what
fraction of a period each person was married.
That is
egen fr_married = mean(married), by(nr)
Then you can combine that with whatever else
you want.
Nick
[email protected]
Richard Goldstein
> try using egen with the rowmean function
[email protected]
> > I have a data set which is sorted by nr (person
> identification number) and year(1980-1987).
> > I want to create a variable for the time average of a
> "married" which is a
> > time-varying dummy variable(taking a value of either zero or one).
> > I have tried for instance:
> > sort nr, year
> > tsset nr, year
> > gen x=.
> > by nr year: replace
> >
> x=(married*d80+married*d81+married*d82+married*d83+married*d84
> +married*d85+m
> > arried*d86+married*d87)/8
> >
> > The point is that I want to have the resulting average
> value NOT to be time
> > variant for each individual e.g: if say marital status is
> taking a value of
> > one for 4 years and zero for the remaining 4 years for a
> given individual I
> > want the resulting variable to take a value of 1/2 for each
> year for the
> > given individual.
> > I have tried many alternative specifications and although
> this seems fairly
> > easy in practice it turns out to be rather tricky.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
