An Algorithmic Aid for Intensity Measures of Consumer Preference
K. Pasumarty, D. F. Karney, T. D. Morley
Operations Research · 1987 · Volume 35 · Issue 3
Abstract
Recently, J. R. Hauser and S. M. Shugan, and K. Pasumarty, have used constant-sum paired comparisons to estimate the intensity of consumer preferences for various products. Both studies use the consumer's responses to determine whether his/her preference function is ratio, interval or ordinal in nature. This determination involves performing a statistical analysis based upon the experimental design matrix, M , which identifies the various transitivity tests designed into the experiment, and upon a companion matrix, W . Unfortunately, both studies generated W by means of exhaustive enumeration, a costly computational technique. In this paper, we propose an algorithm that generates the matrix W cheaply and efficiently. The algorithm is based in part upon modular arithmetic and a technique developed by R. J. Duffin and T. D. Morley for meshing electrical networks. When implemented on a Zenith 150 personal computer, the proposed algorithm found W , for the examples given in both Hauser and Shugan and Pasumarty, in a fraction of the time (at most 0.032) required by exhaustive enumeration. We also show that even when W is not unique, the statistical measurements that utilize W are.
Keywords
411 consumer preference, 437 marketing application
Enable JavaScript for citation tools and additional paper interactions.