Date:         Fri, 25 Jun 1999 11:11:20 GMT0BST
Reply-To:     PAUL MCGEOGHAN <Mcgeoghan@CARDIFF.AC.UK>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         PAUL MCGEOGHAN <Mcgeoghan@CARDIFF.AC.UK>
Subject:      cluster analysis

Hi,

I have data as follows: 5 different levels or variables and 10
chemicals.
The 10 chemicals are at the 5 levels A, B,C, D and E.
For each of the 10 chemicals, runs are carried out which result in a
condition being present or absent at each of the 5 levels.

So I have data as follows:
For each chemical, the number of runs varies:
So for chemical 1, I might have 3 runs:
A B C D E
1 0   1 0  0
1 1   1 1  0
1 1   1 1  1

Chemical 2: 5 runs
A B C D E
1 0   1 0  0
1 1   1 1  0
1 1   1 1  1
0 1   0  0 1

I want to do a cluster analysis for each chemical and compare those
chemicals which have similar clusters for A to E.
So for Chemical 1, the clusters might be (1) A and C, (2) B, D and E.
Chemical 2: the clusters might be A and C again; B and E; D by
itself.

I am comparing similar trends of 1 and 0 above across variables.
To do the above, I have a grouping variable called CHEM.
I split the file by CHEM and compare groups.
Then do a hierarchical cluster analysis using Simple matching
Binary method.

Is this the best approach to take to compare the trends or clusters
for the different chemicals?

Thanks for any replies in advance,
Paul
-----------------------------------------------
Paul McGeoghan
(Statistics and databases support)
Information Services
University of Wales College of Cardiff
Tel (01222) 874875 Ext 5035