| Date: | Sat, 4 Jan 1997 01:24:37 GMT |
| Reply-To: | Casmrl <casmrl@AOL.COM> |
| Sender: | "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU> |
| From: | Casmrl <casmrl@AOL.COM> |
| Organization: | AOL http://www.aol.com |
| Subject: | CHOOSING A SAMPLE WITH CONSTRAINTS |
|---|
Hi !
I need help with the following situation:
consider the following situation: YEAR = 1, 2 (2 LEVELS)
COUNTRY = 1, 2, 3
(3 Levels)
PNB = 1, 2, 3, 4 (4
levels)
To be more elaborate the following program selects a random sample of
size 5 without replacement from the a finite pop. of say 100 obs. The
program works well.
How could you modify it in a such a way that it will select a certain
number of levels for YEAR, (for example 3 levels of 1's and 2 levels of
2's) and for country
a certain fixed number of 2's and 3's and probably not the level 1 and so
forth.
This is almost selecting samples and discarding them untill you obtain
the one you want. Could we have looked at each variable as strata and
then select from each the required number of levels ? How one would
have implemented that idea.
data econ;
RETAIN K 10 N 101;
DROP N K;
INFILE ECONDAT;
INPUT IDNUM YEAR COUNTRY PNB ;
N = N - 1;
IF RANUNI(0) , K/N;
K=K-1;
PROC PRINT;
An acceptable sample with the requirements specified above could be.
Year country PNB
1 2 3
1 2 3
1 2 4
2 1 4
2 3 1
Any help, suggestion greatly appreciated.
Casimir B.
|