The rise of cooperation in correlated matching prisoners dilemma: An experiment
通过多轮囚徒困境实验，研究了两种关联匹配程序（基于行动历史匹配）对合作行为的影响，发现相比随机匹配，关联匹配环境下合作更易持续。
Abstract Recently, there has been a Renaissance for multi-level selection models to explain the persistence of unselfish behavior in social dilemmas, in which assortative/correlated matching plays an important role. In the current study of a multi-round prisoners’ dilemma experiment, we introduce two correlated matching procedures that match subjects with similar action histories together. We discover significant treatment effects, compared to the control procedure of random matching. Particularly with the weighted history matching procedure we find bifurcations regarding group outcomes. Some groups converge to the all-defection equilibrium even more pronouncedly than the control groups do, while other groups generate much higher rate of cooperation, which is also associated with higher relative reward for a typical cooperative action. All in all, the data show that cooperation does have a much better chance to persist in a correlated/assortative-matching environment, as predicted in the literature.