Cooperation in multiplayer dilemmas

We analyze infinitely repeated multiplayer prisoner's dilemmas in continuous-time experiments. As the number of players changes, our design keeps the payoffs of the all-defection, all-cooperation, and unilateral- defection and -cooperation outcomes constant, thus controlling for the minimum discount factor required for cooperation to be an equilibrium. For all group sizes, we study three different variants of the prisoner's dilemma. In further treatments, we allow actions to be chosen from a continuous set. We find that cooperation rates decrease with the number of players, a result that we can attribute to the increased strategic uncertainty in larger groups. The different payoff matrices also affect cooperation. For the payoff matrices with lower levels of cooperation, the group-size effect is weaker. The availability of a continuous action set strongly reduces cooperation rates.