Abstract

The rapidly increasing use in marketing of the techniques of quantitative analysis and the concepts of the behavioral sciences has led many to conclude that marketing is becoming a science. It appears that marketing is now more like the professional areas of engineering and medicine were in their early stages of development than like the social and behavioral sciences. Three approaches to the estimation of distributions of consumer preferences have been briefly discussed. The Benson method treats the fitting of distributions as a "practical research problem" and fits curves directly to data with simple procedures. The more complex Kuehn method fits a distribution without prior assumptions about its shape and also provides an estimate of the basic ability of consumers to recognize differences in the product attribute being studied. A method of fitting distributions and estimating discrimination ability by computer simulation was described and illustrated with the results of two large field preference experiments. It is simpler to apply than the Kuehn method and is proposed for those cases where the researcher's prior assumptions are supported by preliminary analysis of product test data.