Modelling Agent Interdependency in Group Decision Making: Methodological Approaches to Interactive Agent Choice Experiments

Abstract

The past thirty years has seen a growing interest in the development of statistical methods to model choices made by individual agents. The dominant method to emerge, discrete choice modelling, has been applied to a wide number of applications in the areas such as transportation, marketing, environmental science, health economics and public utility regulation. Yet despite the wide level of acceptance, those who employ discrete choice models have often failed to acknowledge that such models assume independency between decision makers. The assumption of independence has significant implications in terms of which contexts discrete choice models should appropriately be applied to. This paper begins by establishing the rationale behind interactive agency choice experiments (IACE), an extension to the traditional discrete choice method that is designed to model agent interdependence. The paper then proceeds to discuss how to model both independent and interdependent decision making processes using the IACE methodology in order to capture information on preferences for all agents within a decision making group. The empirical case study used to illustrate the IACE method focuses on distributive work practice choices.