Dear allstat,
I have a data set where N people answer X questions about P behaviours.
The aim is to cluster the behaviours based on the questions. Since I am
not really interested in individuals differences I have decided to
conduct a cluster analysis on a data matrix where each cell contains the
summary (e.g. mean ) of all the individual responses on that question.
However, Everitt, Landau and Leese (2001 - Cluster analysis, 4th Ed)
suggest that it might be more appropriate to use a summary measure which
takes into consideration within-group variation. They suggest using
Mahalanobis's distance. Could anybody tell me how to calculate this
figure in SPSS? Specifically, I am using the aggregate command to create
my 'mean' matrix, and was wondering if there was anyway that I could get
SPSS to compute the matrix using Mahalanobis distance instead of mean
values.
Many thanks in advance for your help.
Rosie
Rosie McEachan
Research Student
Institute of Psychological Sciences
University of Leeds
|