|Date: ||Wed, 27 Apr 2005 14:57:19 -0700|
|Reply-To: ||s huynh <senthuynh@YAHOO.COM>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||s huynh <senthuynh@YAHOO.COM>|
|Subject: ||Random Sampling Question in Mainframe or MSExcel|
|Content-Type: ||text/plain; charset=us-ascii|
I have a random sampling question and was hoping you'd
be able to help.
I have a dataset with 14,400 unique patients variables
are MRN, Age, Gender, Provider_Name and several
others. Out of the 14,400 patients I will need to
randomly select 1,000 patients. I then will need to
divide that 1,000 patients into 2 groups of 500
patients. First group of 500, we will send an
outreach letter to. The 2nd group, we will do nothing
with, but only use for comparisons to the 1st group.
Is there a way to randomly select 1,000 patients where
it is evenly distributed among Gender, Age, and
Provider_Name? I don't want my sample size to be more
females than males or too many younger age or majority
are from a handful of providers. I want a somewhat
random, but stratified sample size. Any suggestions
on how to accomplish this either in SAS mainframe 8.2
or MS Excel?
Dataset FINALPOP has the following variables:
Age (age >= 55)
Gender (M, F)
Provider_Name (28 providers/doctors total)
Thank you in advance for your time and suggestions.
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around