Re: st: Expanding single observation data to create unbalanced panel data set-up

[Geoffrey Wallace <gw46@cornell.edu>]

Thanks Jeph, this works great for creating the overall data set.

I'm still working through some issues with keeping the relevant 
observation for countries experiencing multiple wars, but will post 
again if I can't resolve this.

Best, Geoff


Jeph Herrin wrote:
> Assume you have four variables  -country-, -warid-, -start-, and -end-
> which contain the country, an identifier for the war, the start year
> and the end year.
>
> Then if I understand your data and problem correctly
>
>  gen length=2004-start+1
>  expand length
>  bys country warid : gen year=start+_n-1
>  bys country year  : egen war=max(inrange(year,start,end))
>  bys country year  : keep if _n==1
>
> should do it.
>
> hope this helps,
> Jeph
>
>
> Geoffrey Wallace wrote:
> > I have a civil war data set that contains a single observation for 
> > each war, including start year, end year, and other conflict-level 
> > variables for all wars that were ongoing or began between 1988 and 
> > 2004. My objective is to expand this data set into a panel format with 
> > separate annual observations for both the war and post-war periods, 
> > while all years before the war started are excluded. So for a war 
> > lasting 1990-1992, this means there would be no observations for 
> > 1988-1989, but then observations for the war years 1990-1992 as well 
> > as postwar years 1993-2004.
> >
> > A complicating factor is that some countries experience several 
> > separate wars during this period, each of which has a unique id. I 
> > want to make sure I don’t create duplicate observations for those 
> > cases, but rather have a simple indicator variable with “1” when a 
> > country is experiencing war and “0” for the post-war period.
> >
> > My question is whether anyone can suggest a relatively efficient way 
> > for producing such a data set.
> >
> > I’ve thought of manually going through the data and including a 
> > variable indicating the number of years for the relevant war and 
> > post-war period, then using the –expand– command to create the larger 
> > data set, and then somehow reformatting it afterward for indicate each 
> > relevant year for the war and post-war periods. However, I was 
> > wondering if there was a more straightforward way to do this.
> >
> > Thanks for your consideration.
> >
> > Geoff
> >
> > -
> > Geoffrey Wallace
> > PhD Candidate
> > Department of Government
> > 214 White Hall
> > Cornell University
> > Ithaca, NY 14853
> > *
> > *   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/
> *
> *   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/

--
Geoffrey Wallace
PhD Candidate
Department of Government
214 White Hall
Cornell University
Ithaca, NY 14853
t.202.302.4209
f.607.255.4530
http://www.people.cornell.edu/pages/gw46/
*
*   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/