Communication and Cooperation: A Methodological Study

We study the effect of communication in an experimental game where cooperation is consistent with equilibrium play. We examine two methodological questions which affect many studies of communication in games, particularly those studying the relationship between communication and cooperation in supergames. We show that making it easier for subjects to reach an agreement (by allowing more rounds of communication) does not increase cooperation with a limited message space treatment. Thus, limited message space treatments are not a good substitute for the use of chat. Making the game longer, thereby making the connection between the first period and the continuation game more obvious, also has little impact on cooperation. For both the two-and three-period versions of our game we find substantial amounts of contingent play. This implies that use of a finitely repeated game, which has a number of methodological advantages, maintains the main properties of the underlying supergame.