Cluster Analysis: A Brief Discussion of Some of the Problems

Abstract

In the last two years psychiatrists, amongst others, have become aware of the technique of cluster analysis, and having obviously decided that this technique may be useful have applied it to sets of data (e.g. Pilowsky et al., 1969; Paykel, 1970), and in some cases drawn conclusions about the results. This paper will attempt to point out some of the problems of this type of analysis, in the hope that in the future casual users of the technique may be more aware of its possible pitfalls.