Identifying Team Reasoning

• In Event: Theory and Experimental Evidence of Cooperation and Coordination

Abstract

In the laboratory, we implement an experimental coordination game that allows us to cleanly identify when and whether individuals reason in the self-regarding I-mode or a group-regarding We-mode. In a 4-player Stag Hunt game, we vary the success threshold and whether we induce group identities for the individual subjects (ID vs. no-ID treatments), and elicit beliefs about others' behavior. Assuming that individuals hold a combination of I- and We-concerns and that inducing identities increases the weight on the We-concerns - holding fixed the beliefs about other players, the sensitivity to the threshold change should come at the I-mode reasoning margin, rather than at the We-mode margin, and so be no greater (and possibly smaller) in the ID treatment than in the no-ID treatment (a difference-in-differences test). Our laboratory results show that we can reject the hypothesis of the significant incidence of We-mode reasoning. However, we also find evidence that subjects believe that there are some members of the group who will, indeed, reason in the We-mode. Best-responding to that belief can lead to I-mode behavior that is observationally equivalent to the We-mode behavior. The experiment provides the first well-identified test of the theory of team-reasoning that has been used to explain a number of pro-social behaviors in politics, including voter turnout and private provision of public goods.

Authors

• Dominik Duell, University of Innsbruck

• Catherine Hafer, New York University

• Dimitri Landa, NYU