Date: Mon, 22 Nov 2004 08:44:39 -0500
Reply-To: Jim Groeneveld <jim1stat@YAHOO.CO.UK>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Jim Groeneveld <jim1stat@YAHOO.CO.UK>
Subject: Re: How to deal with the missing values in PROC CORR
Hi Fred,
We already had some off-line discussion on this at the end of last week.
Now it is clear to me what you mean by M and N, not the number of
observations in each dataset (which should be equal), as I wrongly
interpreted, but the number of variables; and those do indeed not have to
be equal. But an ID variable is mandatory or you should be very sure that
the observations match one by one in their current order. So you would want
M*N correlation coefficients, right? Missing values are automatically taken
into account of (pairwise i think), because it is impossible to calculate a
correlation coefficient involving missing values.
Regards - Jim.
--
Y. (Jim) Groeneveld, MSc., Biostatistician, Science Team
Vitatron B.V., Meander 1051, 6825 MJ Arnhem
P.O.Box 5227, 6802 EE Arnhem, the Netherlands
Tel: +31/0 26 376 7365, Fax: +31/0 26 376 7305
Jim.Groeneveld_AT_Vitatron.com (replace _AT_ by AT sign)
http://www.vitatron.com, http://home.hccnet.nl/jim.groeneveld
My computer multi-boots OS's, each of them adapting the DST twice a year.
[common disclaimer]
On Fri, 19 Nov 2004 12:19:46 -0500, Fred <ieaggie2002@GMAIL.COM> wrote:
>Hi, I have two data sets for correlation analysis.
>
>The data set 1 is:
>var1 var2 var3 .... var19 var20
>xxx xxx xxx xxx .
>xxx xxx xxx ... . .
>xxx xxx xxx ... . .
>xxx xxx xxx ... xxx xxx
>xxx xxx xxx ... xxx xxx
>.
>.
>.
>
>Dataset 2 is
>var2_1 var2_2 var2_3 .... var2_9 var2_10
>xxx xxx xxx .... . .
>xxx xxx xxx ... xxx .
>xxx xxx xxx ... xxx xxx
>xxx xxx xxx ... xxx xxx
>.
>.
>.
>Acutally, I do not know aprior what rows in each data set will
>contain missing values.
>So in proc corr, how to delete these missing values to
>get more accurate correlation analysis?
>
>Thanks so much for your kine help.
>
>Fred
| 
