An Alternative Preference Elicitation Procedure in Predicting Choice

Abstract

This research proposes the development and implementation of a better method (MAGL: Multi Attribute Gain Loss) for multi-attribute decision making (MADM) under certainty. MAGL is based upon the tenets of norm theory, prospect theory, Kauffman’s complexity theory and context dependent choice theories. Since the proposed method is a compositional approach using the frameworks of Multi-Criteria Decision Analysis (MCDA), its effectiveness is tested against Analytic Hierarchy Process (AHP) on two MADM problems for predicting the choice behavior of the DM. With less number of preferences taken from each individual in MAGL (order of n) in comparison to AHP (order of n2), the solution prescribed by the former resembles more closely with the actual choice behavior of the individuals. The analytical model provided by MAGL is consistent and strengthening the arguments of academicians and practitioners that the overall satisfaction function might not be linear and/or symmetric. The results of two studies conducted during the course of this research conclude that the alternative valuations on different attributes is a nonlinear- reference point dependent function of the associated objective value of the alternative on the same attribute. The integration rule which best describes how evaluations are integrated into overall valuations follow a non-linear, non-compensatory context dependent relation, overweighing of negative information. MAGL can help marketing/product managers analyze the dominance of attribute levels in the selection/rejection of an alternative. Since the context or the choice set plays an active role in the process of choice and MAGL is able to model the context dependent choice behavior of the consumers, the marketing/product managers can design new products and analyze which combination of these attribute levels will perform better with the choice set already available in the market. With the preferences on product attributes from a sample population, marketing/product managers can also analyze the changes in the market share with the introduction (recall) of products in (from) the market of the similar consumer population.