Scaling up Instance-Based Learning Theory to Account for Social Interactions

Authors


  • This research was supported in part by a grant from the Defense Threat Reduction Agency (DTRA) grant number: HDTRA1-09-1-0053 to Cleotilde Gonzalez and Christian Lebiere. We thank Hau-yu Wong for editing this manuscript and Christian Lebiere and Ion Juvina for discussions that contributed to ideas in this article.

Cleotilde Gonzalez, Dynamic Decision Making Laboratory, Carnegie Mellon University, Pittsburgh, 15213 PA, U.S.A.; e-mail: coty@cmu.edu.

Abstract

Traditional economic theory often describes real-world social dilemmas as abstract games where an individual’s goal is to maximize economic benefit by cooperating or competing with others. Despite extensive empirical work, descriptive models of human behavior in social dilemmas are lacking in both cognitive realism and predictive power. This article addresses a central challenge arising from the success of modeling individuals making decisions from experience: our ability to scale these models up to explain social interactions. We propose that models based on the instance-based learning theory (IBLT) will help us to understand how conflictual social interactions are influenced by prior experiences of involved individuals and by information available to them during the course of interaction. We present mechanisms by which IBLT might capture the effects of social interaction with different levels of information without assuming predefined interaction strategies, but rather by assuming learning from experience.

Ancillary