Precedent transfer in experimental conflict-of-interest games

In this paper, I use simple coordination games to examine the process under which precedents (or past equilibrium experiences) transfer across games. I study a 3-player conflict-of-interest game that possesses two focal equilibria: an equal payoff equilibrium and a coalitional equilibrium. The social convention of equal sharing is not sufficiently powerful to select the equal-payoff equilibrium. However, after subjects develop homegrown experience playing the equal-payoff equilibrium in an analogous 2-player game, they are significantly more likely to select the action associated with the equal-payoff equilibrium. However, surprisingly, they are also much more likely to select the action associated with the coalitional equilibrium. Using expectations data also collected in the experiment, I speculate on the reasons for this latter effect.