I don't think you meant what you wrote.
sort prov time
by prov: generate t = _n
tsset prov t
or
bysort prov (time) : gen t = _n
tsset prov t
is more likely to be what you meant.
But while this may make a command
happy to play with sone data, it ignores
much of the information in those data,
and creates time equivalences that may not
match the original (e.g. observations with
t = 1 may well have different underlying
-time-s).
Otherwise put, this is not just a trick to
get a command to play. It needs justification
in terms of the underlying science (or whatever).
Nick
[email protected]
Mansour Farahani
> You can create another variable for time like
>
> sort prov
> by prov: generate t=[_n]
> tsset prov t
>
> where your id is the identification number for each entity in
> the panel. In this way there should not be any gap then in
> the data. then you and you can run xtabond.
>
*
* 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/
