Isn't that similar to a logistic regression problem, where your dependent variable is either the event (cases - with the outcome) or non-event (controls - without the outcome)?

Maybe you want to add a segment variable (the upper quartile, lower quartile that people belong to) in a logistic model form /multiple regression model form specified to SAS PROC POWER for your power calculation.

Susie Li TV Guide 1211 Avenue of the Americas New York, NY 10036 Tel 212.852.7453 Email susie.li@tvguide.com

-----Original Message----- From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of wraong Sent: Monday, May 08, 2006 12:39 PM To: SAS-L@LISTSERV.UGA.EDU Subject: Re: help on power calculation

Thanks for your kind suggestion.

*Case-control study* is a retrospective observational study in which investigators identify a group of patients with a specified outcome (cases) and a group of patients without the specified outcome (controls). Investigators then compare the cases and the controls to determine possible factors which can be linked to the outcome.

On 5/8/06, Jiann-Shiun Huang <Jiann-Shiun.Huang@amerus.com> wrote: > > Wrong: > > Is it possible we can describe in detail about case-control study so > that people like me without knowledge in the study could give a shot at > your problem? > > J S Huang > 1-515-557-3987 > fax 1-515-557-2422 > > >>> Wrong Wang <wrong.wang@GMAIL.COM> 05/08/06 11:16 AM >>> > Hi, all > > I am asked to do power calculation for a case-control study, the > sample > sizes are equal in the two groups. I will compare the upper quartile > group > to the lower quartile group. > > I know I can use proc power in SAS to calculate power for a normal > case- > control study, which given prevalence of exposure in control group. But > in > the above condition, I don't think it's correct to just set prevalence > of > exposure in control group as 0.25. > > So I need your help on how to solve such problem. > > Thanks. >