Date:         Tue, 30 Apr 1996 14:27:00 PDT
Reply-To:     Fry Chris <chris@EMAIL.DFI.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Fry Chris <chris@EMAIL.DFI.COM>
Subject:      Re: saving corr. coeff. as a variable

I've done this before, but don't have the code handy.  You can use the
output statement in proc corr to put the correlation matrices into a
dataset.  Then, do a summary on this dataset, with a subsetting "if" to make
sure you are only looking at one observation per correlation value, to get
the average corr.
 
Chris Fry
chris@dfi.com
 ----------
From: owner-sas-l
To: Multiple recipients of list SAS-L
Subject: Re: saving corr. coeff. as a variable
Date: Tuesday, April 30, 1996 12:47PM
 
9449673@NESSIE.CC.WWU.EDU writes:
 
>I AM WONDERING IF SAS WILL ALLOW ME TO SAVE A BIVARIATE CORRELATION
>COEFFICIENT AS A NEW VARIABLE.   I KNOW THAT IT CAN SAVE PREDICTED AND
>RESIDUAL SCORES, AND WILL SAVE THE RESULTING CORR. MATRIX.  HOWEVER, WHAT
>I AM DOING IS COMPUTING A LARGE NUMBER OF CORRELATIONS FROM A POPULATION,
>AND TAKING THE AVERAGE OF THEM.  I WOULD LIKE TO BE ABLE TO DO THIS WITHOUT
>HAVING TO COMPUTE THE AVERAGE BY HAND. DOES ANYONE KNOW IF SAS CAN
>ACCOMPLISH THIS?
 
YES.