Experimental and Behavioural Economics Seminar Series - James Bland (University of Toledo)

Title: Meta-study in the shadow of the future: the effect of continuation probability on strategies

Abstract: This paper develops methodology for estimating and aggregating causal effects for latent outcome variables. Using data from forty treatments in eleven economic experiments, I estimate and aggregate causal effects of the continuation probability on players’ strategy choice and first-round cooperation in the indefinitely repeated Prisoner’s Dilemma. Increasing the continuation probability significantly and substantially increases the probability that a participant will cooperate in the first round of a supergame, and use the “Grim trigger”, “Tit-for-tat”, and “Tit-for-two-tats” strategies. Increasing the continuation probability decreases the probability that a participant will use the “Always defect” and “Suspicious tit-for-tat” strategies.