Good. My advice was even better...
Nick
[email protected]
Scott Cunningham
> Yes, what I proposed worked.
>
> I am once again impressed by how much easier my life is with -
> reshape-. Very simple solution, if you can even call it that, but I
> was mapping out much more convoluted solutions and so I am glad for
> that reshape command.
>
>
> On Nov 24, 2006, at 10:33 AM, Nick Cox wrote:
>
> > Maarten Buis recommended a book for you to buy
> > and I can endorse that. In fact, it is so good
> > you should have one for home and one for work.
> > Perhaps even three, because someone will walk
> > away with one and you'll never see it again.
> >
> > In the mean time, here is one approach to
> > what I understand to be your problem.
> >
> > 1. Add a pseudo-county to the end of your data
> > with the missing years. Use an -fip- already
> > used. You can do this in -edit-.
> >
> > county fip year
> >
> > 9999 1 1983
> > ...
> > 999 1 2004
> >
> > 2. Now fill in using
> >
> > fillin county year
> >
> > 3. Now -drop- the observations for the pseudo-county.
> > It never existed anyway.
> >
> > Nick
> > [email protected]
> >
> > Scott Cunningham
> >
> >> I have the following dataset on county-level data. VBL is a dummy
> >> indicating whether an object is present in the county. COUNTY is a
> >> number greater than 0 indicating which United States
> county is being
> >> described, and FIP is indicating which US state.
> >>
> >>
> >> +-----------------------------+
> >> | county fip year vbl |
> >> |-----------------------------|
> >> 1. | 1 1 1980 0 |
> >> 2. | 1 1 1981 0 |
> >> 3. | 1 1 1982 0 |
> >> 4. | 1 1 1984 0 |
> >> 5. | 1 1 1985 0 |
> >> |-----------------------------|
> >> 6. | 1 1 1987 1 |
> >> 7. | 1 1 1988 1 |
> >> 8. | 1 1 1991 1 |
> >> 9. | 1 1 1992 1 |
> >> 10. | 1 1 1995 1 |
> >> |-----------------------------|
> >> 11. | 1 1 1996 1 |
> >> 12. | 1 1 1999 1 |
> >> 13. | 1 1 2000 1 |
> >> +-----------------------------+
> >>
> >> As one can see, I am missing years 1983, 1986, 1989, 1990, 1993,
> >> 1994, 1997, 1998, and 2001-2004. For each county/fip, I want to
> >> create 12 new observations which I can then use to fill in the
> >> missing years. I can do the filling in, but I cannot find
> >> information for simply expanding this dataset by n observations. I
> >> have searched for the answer, but -expand- takes me to various
> >> places, like -fillin-, neither of which seem appropriate. Any help
> >> is appreciated.
*
* 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/