LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (November 2005, week 2)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Thu, 10 Nov 2005 13:41:11 -0800
Reply-To:     David L Cassell <davidlcassell@MSN.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         David L Cassell <davidlcassell@MSN.COM>
Subject:      Re: proc freq to find variable most correlated to x
In-Reply-To:  <200511100015.jA9MwTVG002426@malibu.cc.uga.edu>
Content-Type: text/plain; format=flowed

davefickbohm@YAHOO.COM wrote: >I want to use proc freq to find among variable X what other variable, A,B >or C is most correlated to Y. >I think that the following will do this. > >proc freq data = xxx; > tables Y * (a,b,c) / measures; >where x = 1; >run; > >Am I correct ?

I'm not sure. It seems to me that by fixing the level of X (with that WHERE statement) you may be missing a lot of the features of the relationships with Y.

Are you saying that you would like to look at the correlations while controlling for the effect of X? In that case, I would go with PROC CORR instead. I would use the PARTIAL statement to get the partial correlations of Y with {abc} while adjusting for X. This is really like regressing all the variables on X, then looking at the correlation of the residuals for Y on the residuals for {a b c}. The code would look like this:

proc corr data=YourData; var Y; with a b c; partial X; run;

And if you're running on SAS 9.1, you can use the SAS ODS Statistical Graphics to get plots of the partial residuals, to see what the data are *really* doing. The code would look *something* like this:

ods html; ods graphics on;

proc corr data=YourData plots=scatter(alpha=.10 .30); var Y; with a b c; partial X; run;

ods graphics off; ods html close;

The PLOTS=SCATTER above will give you 70% and 90% prediction ellipses superimposed on your plots, along with the points.

Oh, by the way, if you want to use PROC FREQ as above, you need to replace the commas in (a,b,c) with spaces:

tables Y * (a b c) / measures;

HTH, David -- David L. Cassell mathematical statistician Design Pathways 3115 NW Norwood Pl. Corvallis OR 97330

_________________________________________________________________ Don’t just search. Find. Check out the new MSN Search! http://search.msn.click-url.com/go/onm00200636ave/direct/01/


Back to: Top of message | Previous page | Main SAS-L page