Functional and Behavioral Application

This paper presents a model for explaining preference among similar alternatives. Preference is hypothesized to be governed by the relative position of the alternatives on specific dimensions (attributes). More specifically, preference is stated to be related to the distance of an alternative from an ideal alternative. A summary of a test of the model is presented. Next, a probabilistic form of the model is obtained assuming that the measured distance is affected both by individual uncertainty and measurement error. For a particular case, the resulting normality of the distribution is demonstrated. This introduction of a random element into the model is seen to be a major advance, although final judgment must await extensive testing.