Date:         Thu, 19 Jun 2003 17:08:18 GMT
Reply-To:     Mark AC Emmett <markacemmett@AOL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Mark AC Emmett <markacemmett@AOL.COM>
Organization: AOL, http://www.aol.co.uk
Subject:      [Q] Calculating the Optimal Sample Size?

Hello all,

I want to keep a sample of my overall group as a control group to measure back
against.  My control group is about 10% of the overall number of observations.


This is fine if my number of overall observations is large, say 5,000 (thereby
having a control group of 500).  But what if the number of my observations was
only 10?  A control group of 1 is hardly going to be statistically robust to
test against.

I am wondering if there is any formulaic way (rather than just saying x%) of
determining what your control group should be based on the total number of
observations available to you?

Thank you in advance
Mark