Date: Sun, 20 May 2001 09:50:44 -0400
Reply-To: Jay Weedon <jweedon@EARTHLINK.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Jay Weedon <jweedon@EARTHLINK.NET>
Subject: Re: Group Mean Comparisons
Content-Type: text/plain; charset=us-ascii
You can't directly compare the two means if you have overlap.
One suggestion might be to discard cases with both DEP=1 & DYS=1.
Or consider THREE groups: (DEP=1,DYS=0), (DEP=0, DYS=1), (DEP=1,
DYS=1). Or even four groups if you have people with DEP=0 DYS=0. Make
a new grouping variable GROUP coded 1,2,3 or whatever, and use code
PROC GLM; CLASS GROUP; MODEL BP=GROUP;
Another possible way of looking at the data if there are four groups
is factorial ANOVA - examine the joint effect of DEP & DYS on BP:
PROC GLM; CLASS DEP DYS; MODEL BP=DEP|DYS;
On 20 May 01 03:47:50 GMT, gongfang@EXCITE.COM (Judy Xu) wrote:
>This should be simple, but I don't know how to do it: I want to take poeple
>who have different types of psychiatric disorders and test mean differences
>between each on the subscales of health outcomes.
>For example, "Depression" (DEP) is a dummy variable with 1=having
>depression disorder and 0=no depression disorder. The same is for
>the "Dysthymia" variable: 1=having dysthymia and 0=no dysthymia. "Bodily
>pain" (BP) is a continuous variable from 1-100. Suppose the mean score of
>bodily pain (BP) for people with depression (DEP) is 72 while the mean of
>BP for people with Dysthymia is 68. How could I find out whether the 2
>mean scores are significantly different from each other? What is the SAS
>code for this? The trouble is that there is overlapping between people who
>have depression and those who have dysthymia. I tried the GLM procedure,
>but it didn't work. Any help will be greatly appreciated!