This paper is a comparison between two different methodological approaches
to identifying and interpreting market segments in conjoint analysis. One o
f the methods is segmentation by fuzzy clustering accompanied with a logist
ic regression of membership values versus consumer attributes. The other ap
proach is based on a mixed model ANOVA of both conjoint and consumer variab
les simultaneously The two methods are compared on a three attribute conjoi
nt study of margarine. The results from the two methods are very similar fr
om a qualitative point of view. The most important information found is tha
t well educated women separated strongly from men with little education in
their purchase intent for products with a high fat content. (C) 2001 Elsevi
er Science Ltd. All rights reserved.