A lot easier...

proc means clm ; var x ;

*** CLM stands for Confidence Limits for Means ; *** The ALPHA= option can also be used to set the confidence level ; *** By default you get a 2-sided CI with alpha=5% ; *** Available since Release 6.07 ; *** Works with proc summary also ; *** Output statement supports LCLM & UCLM keywords ;

Francois

---- Francois Derderian - Consultant SAS Institute (Canada) Inc.

tel: (514) 395-8922 (240) fax: (514) 395-8962

Internet: canfdd@mtl.sas.com

-----Original Message----- Date: Tue, 10 Feb 1998 17:59:44 -0500 From: "Abelson, Robert" <RobertA@MOCR.OAPI.COM> Subject: Confidence Intervals for means - correction

Sorry!

To get a 95% confidence interval, the code should be as follows:

proc sql; select n, xbar, s, xbar - (tinv(.975,(n-1))*(s/sqrt(n))) as ci_lower, xbar + (tinv(.975,(n-1))*(s/sqrt(n))) as ci_upper from (select count(x) as n, avg(x) as xbar, std(x) as s from a); quit;

Bob Abelson Otsuka America Pharmaceuticals

-----Original Message----- From: Edward S. Suarez [SMTP:usbscb63@IBMMAIL.COM] Sent: Tuesday, February 10, 1998 5:03 PM Subject: Confidence Intervals for means

Hello SAS gurus, I've been trying to find a simple way to obtain confidence intervals for the mean of a group of numbers. I thought that there would be an option in proc means to do that, or in another proc, but I don't seem to find it. Can anybody help me with this? I of course know that it can be done using a proc means and a data step with the TINV() function but i can't help thinking there must be an easier way. Thanks in advance, Edward

Edward S. Suarez Engineer - Technology Department BETHLEHEM STEEL CORPORATION, Burns Harbor Division (219)787-3116 Fax:(219)787-3052 e-mail:usbscb63@ibmmail.com

Bob Abelson