Date:         Wed, 18 Jun 2008 10:01:40 -0500
Reply-To:     "Paul A. Thompson" <paul@WUBIOS.WUSTL.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Paul A. Thompson" <paul@WUBIOS.WUSTL.EDU>
Subject:      Re: How to do Multidimensional Scaling using SAS on this data
Comments: To: susie.li@BOEHRINGER-INGELHEIM.COM
In-Reply-To:  <91A7A16D54C7C54B92398BFBA953A80002EE6AA6@RDGEXM03.am.boehringer.com>
Content-Type: text/plain; charset="us-ascii"

There are several approaches, depending on how you wish to work with the
data.

1) Construct a dissimilarity or distance matrix from the data.  To do this,
you compute the distance (Euclidean) of each term from the other.  I believe
that SAS now has a distance computation procedure.

2) Use an unfolding approach, in which you directly model the data using an
asymmetric analysis method.

Both are available in SAS using PROC MDS.  You should look at it, since the
sort of problem that you discuss is pretty common, and there might be an
example of just such an analysis.

Paul A. Thompson, Ph.D.
Division of Biostatistics, Washington University School of Medicine
660 S. Euclid, St. Louis, MO 63110-1093
314-747-3793
paul@wubios.wustl.edu

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Susie C
Y Li
Sent: Wednesday, June 18, 2008 9:58 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: How to do Multidimensional Scaling using SAS on this data

Hi all,

I have a set of snack rating data (scales from 1-10).  I am trying to
visually map the similarity/dissimilarity of snacks using PROC MDS.

Can anyone help me with this?




Snack   Salt    Sweet   Nutrition
A       1       5       2
B       2       4       3
C       3       2       3
.
.
.
.

Thanks.

Susie Chung Li