# Re: st: RE: question on keeping observations - selecting substitute instituions for sampled institutions

 From Gáti Annamária <[email protected]> To [email protected] Subject Re: st: RE: question on keeping observations - selecting substitute instituions for sampled institutions Date Wed, 06 Jan 2010 13:45:47 +0100

Dear Martin,

The following lines of your syntax solved my problem completely:

gen int school=_n
gen byte substitute=minta[_n+1] | minta[_n-1]

I already had the sample so I only needed to mark the substitutes somehow.
I see the overlapping substitutes, it is a further step to decide how to handle them.
Thank you very much for your kind and fast help,

annamaria

**
**
**

Martin Weiss írta:
<>

You can -generate- a dummy for "being a substitute". What do you want to happen for adjacent sampled schools? Their substitutes must overlap, such as 324 and 325 in my example:

*******
vers 11

clear*
set obs 4000
set seed 43532

gen int school=_n

gen randomnumber=runiform()
sort randomnumber
gen sampled=_n<=300

sort school

gen byte substitute=sampled[_n+1] | sampled[_n-1]

//problem: sometimes sampled schools adjacent to each other
cou if substitute
l in 322/326

*******

HTH
Martin

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Gáti Annamária
Sent: Mittwoch, 6. Januar 2010 12:56
To: [email protected]
Subject: st: question on keeping observations - selecting substitute instituions for sampled institutions

Dear List users,

I am trying to collect a sample out of schools. I have successfully sampled 300 institutions out of 4000. I have now a dataset containing 4000 observations and a variable named sample indicating whether a given institution is part of my sample (code=1) or not (code=0).

As I also need 2 substitutes for each institutions I would like to chose the ones that are right above and under the sampled school. Is there a simple way to do this with stata? I was browsing through the Keep observations commands but didn’t manage to find out how to do this.

Annamaria
*
*   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/

*
*   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/

• Follow-Ups:
• References: