There is also a non-canned solution at
http://www.stata.com/support/faqs/data/panel.html
which you could use after (temporarily, if you like)
-drop-ping all observations with missing values.
. egen nmiss = rmiss(y x1 x2 x3)
. drop if nmiss
. by id: gen run = cond(L.run == ., 1, L.run + 1)
. by id: egen maxrun = max(run)
. drop if maxrun < 4
Nick
[email protected]
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of Nick Cox
> Sent: 10 November 2003 16:57
> To: [email protected]
> Subject: st: RE: Drop observations
>
>
> I assume identifier variable -id-, time variable -t-
> and that you have -tsset- the data.
>
> For regression it's problematic if either the response
> or any of the covariates is missing. Compute
>
> . egen nmiss = rmiss(y x1 x2 x3)
>
> One way of doing it then is to use -tsspell- from
> SSC.
>
> . ssc inst tsspell
> . tsspell, cond(nmiss == 0)
>
> defines spells for which all variables are present,
> and does so automatically within panels.
>
> The longest spell in each panel is then
>
> . egen longest = max(_seq), by(id)
>
> and you can
>
> . drop if longest < 4
>
>
> Nick
> [email protected]
>
> Vidya Mahambare
>
> > For panel data regression, I want to include only those
> > firms which have
> > observations for at least for four consecutive years out of
> > ten. I suppose I
> > have to use drop or keep command, but can't figure out how.
>
> *
> * 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/
*
* 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/
