RE: st: RE: RE: A Question on Selecting a Sample from a Panel Data Set

["Nick Cox" <n.j.cox@durham.ac.uk>]

No, I don't think so. You need to change the code upstream, not
downstream. 

clear

set obs 19
gen int year=_n+1986

//10 firms
expand 10
bys year: gen byte firm =_n
gen byte derivuse=runiform()>.25 if year==2000 | year == 2001 
bysort firm (derivuse) : replace derivuse = derivuse[1] == 1

Nick 
n.j.cox@durham.ac.uk 

Martin Weiss

*************
bysort firm (derivuse):  /* 
*/ replace derivuse = derivuse[1] == 1 |  /* 
*/ derivuse[2] == 1
*************

Hong Nguyen

I have two years of derivative use data (2000 and 2001). I could select 
a sample of firms based on derivative use in either 2000 or 2001 (and 
spread that property to other observations for the same firms). How 
would I modify your previous commands to fit this situation?

Nick Cox wrote:

> Not quite. 
>
> -derivuse- is undefined, namely missing, for years not 2000. You need
to
> spread that property to other observations for the same firms. 
>
> bysort firm (derivuse) : replace derivuse = derivuse[1] == 1 

Martin Weiss

> If your dataset looks like this:
>
> ***********
> clear*
>
> set obs 19
> gen int year=_n+1986
>
> //10 firms
> expand 10
> bys year: gen byte firm =_n
> gen byte derivuse=runiform()>.25 if year==2000
> ***********
>
> you can simply condition on -if derivuse==1- in your analyses.

Hong Nguyen

> I have a panel data set consisting of financial data for 423 firms
over 
> a 19-year period.  For one of the years (2000) I have data on 
> derivatives use (use or non-use) for these 423 firms. I need to
generate
>
> a sample of firms that used derivatives in 2000 and use it to run 
> regressions over the entire period 1987-2005. This is where I'm stuck.

*
*   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/