Date:         Sun, 19 Jul 2009 13:15:18 -0500
Reply-To:     Joe Matise <snoopy369@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Joe Matise <snoopy369@GMAIL.COM>
Subject:      Re: PROC SUMMARY
Comments: To: OR Stats <stats112@gmail.com>
In-Reply-To:  <6eca73440907191050r3688f3d6r39e4f56f6b20ffb5@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

Sounds like a CLASS situation to me...

proc means data=have;
var y;
class a b c d;
types a b c d;
output out=want mean=;
run;

or whatever.  If you just want the A=1 etc. (not 0) then put a where
condition on your output dataset to require A or B or C or D to be 1.

-Joe

On Sun, Jul 19, 2009 at 12:50 PM, OR Stats <stats112@gmail.com> wrote:

> Hello:
>
> I would like to do a PROC SUMMARY for the means and size of the group
> membership of my data.  The problem is that each record can belong to more
> than one group.  And I would like to calculate the summary statistics for
> the groups in a non-exclusive way.
>
> Group membership is currently captured across binary columns of my data.
>
> record_id   A   B   C   D   y
> 1              1    0    0   0    #
> 2              0    1    0   0    #
> 3              1    1    1   1    #
>
> Group is therefore A, B, C, or D.  In the output summary table, I would
> have
> summarized y by groups A, B, C, and D.And I would like to get the summary
> statistic for y by Group, where A includes records 1, 2 and 3; B is just 2
> and 3; C is just record 3; and D is also just record 3.
>
> Can I do this in one PROC SUMMARY datastep?  Or do I need an intermittent
> dataset to reorganize my table above?
>
> Thank you!!
>