Preference Scaling via a Modification of Shepard's Proximity Analysis Method 1

Abstract

A method of scaling based on an iterative computer program is described. The program accepts as input any set of paired-comparison preference data and produces as output a one-dimensional scale of preference. The method is illustrated by application to actual data and the solution compared to that obtained by Thurstone's method on the same data. Finally, some remaining problems and present limitations of the method are described.