Date:         Tue, 22 Nov 2005 15:55:47 -0500
Reply-To:     Jay Weedon <jweedon@EARTHLINK.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Jay Weedon <jweedon@EARTHLINK.NET>
Organization: http://newsguy.com
Subject:      Re: Proc Means with statistical Test
Comments: To: sas-l@uga.edu
Content-Type: text/plain; charset=us-ascii

On Tue, 22 Nov 2005 08:55:00 -0800, annekolean@YAHOO.COM (anne olean)
wrote:

>Hi,
>i have a variable (age) that i want to compare wrt
>means across three groups. I used proc means to do
>that:
>
>proc means data=preds n mean median std;
>  class group;
>  var age;
>run;
>
>this gives a nice table, but is there an option to get
>a test statistic comparing the 3 groups(all i found is
>the "t-test" and that's not what i want). i know i can
>get the statistic from proc glm, but i want to create
>a table as in proc means with the last column being
>the test statistic and p-value....can that be done
>within proc means? if not, how can i get the nice
>looking table that proc means produces in proc glm?

Your most likely statistical tests are:

1. ANOVA (e.g., proc glm)
2. Kruskal-Wallis test (proc npar1way)
3. Exact permutation test (proc npar1way)

AFAIK the only way to get a p-value into the same table as the means
is to get the different bits that you need using various procedures,
and cobble them into a table yourself using proc report or data
_null_. It's unlikely to be worth the effort.

JW