st: Re: restricting the sample

[Christopher F Baum <baum@bc.edu>]

On Oct 7, 2004, at 2:33, Michael wrote:

> 	When using -tsset- data (time series or panel), I have often estimated
> regressions on a sub-sample of time observations by using:
>
> 		reg y x1 x2 if tin(date1,date2)
>
> and even once or twice:
>
> 		reg y x1 x2 if ~twithin(date1,date2)
>
> I can see why those working with cross-sectional data would welcome
> similar convenience.
I fail to see why those working with cross-sectional data would find

reg y x1 x2 if iwant

to be more onerous than the examples above. Define a dummy variable 
that points to the cases that you want (or don't want!), and use it!

In response to Clive's proposed solution, anything that _depends_ on 
observation numbers is IMHO a Very Bad Idea. All too easy to change 
those without realising it. Better to create

gen byte zap = inlist(_n,10,100,250)
reg whatever if !zap

(for inlist(), Yet Another Useful Function that one uses with generate, 
whelp progfun)

Kit Baum, Boston College Economics      baum@bc.edu
http://ideas.repec.org/e/pba1.html

*
*   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/