Department of Computational Social Science Seminar Abstract - Cotla Dissertation Proposal

Thursday, February 14 - 2:00 p.m.
161 Research Hall

DISSERTATION PROPOSAL: Heterogeneous Preferences and the Dynamics of Cooperation in Networked Societies: A Dialogue Between Computational and Experimental Approaches

Chenna Reddy Cotla
CSS PhD Student
George Mason University

Abstract: Understanding the fundamental patterns and determinants of cooperation within social groups remains a challenge across disciplines. It has become an accepted paradigm that social networks are important determinants of group outcomes. Results from agent-based simulations have shown that network structure plays a central role in determining the evolutionary dynamics of cooperation. Similarly, recent behavioral experiments suggest that individuals vary in their degree of cooperativeness. A group with a network structure that promotes cooperation but consists of a majority of uncooperative types could just be as unproductive as a group consisting of mainly cooperative types but with a network structure that is not conducive to the evolution of cooperation. Despite the importance of both the network structure and the distribution of cooperative types that populates it, there is a little research on how these two elements interact to promote cooperation in the absence of costly sanctioning institutions. Literatures in agent-based modeling and experimental economics have documented the importance of networks and heterogeneous preferences in the context of cooperation separately, but their juncture has not been studied. The proposed research aims to address this existing void. Agent-based models will use data from human subject experiments to develop behavioral specifications of agents and then model the dynamics of real world networked systems with high fidelity. The accuracy of such agent-based models will be further analyzed and refined iteratively with a series of human subject experiments that will test the predictions of the model. The first part of the research aims to explore the role of fixed network structures on the cooperative behavior. The agent-based models will simulate cooperative behavior over various network structures to inform which structures and placement configurations of cooperative types on the networks are most successful at sustaining cooperation. Predictions from the simulations are then taken back to the laboratory to see how human subjects actually behave in those settings. The results will provide information for better calibrating agent-based models and inform how people update beliefs and make decisions in networked settings. The second part of the proposed research will investigate the strategic link formation and link dissolution of cooperative types by considering a dynamic social network setting. This component will shed light on the endogenous self-organization of cooperative behavior within dynamic groups. This research broadens the current scientific understanding of cooperation in networked societies via an empirical dialogue between computational and experimental approaches. Methods from computer science, experimental economics, computational statistics, social network analysis and quantitative analysis will be integrated.