Experimental investigations on information transmission and cooperation in an indefinitely repeated dilemma game

Abstract

This thesis experimentally explores how people create reputational information through reporting a partner’s past behaviour –– and whether the various forms of endogenous information transmission help sustain cooperation –– using an indefinitely repeated Prisoner’s Dilemma game. The research is based on two sets of laboratory experiments, one in 2018 and the other in 2021. Both are based on an infinitely repeated Prisoner’s Dilemma game with random matching.

Chapter 1 briefly introduces the stage game and selective literature on reputation, and it also discusses theory that informs our experimental setups in subsequent chapters.

Chapter 2 experimentally discusses the transmission of objective information, such as truthful information about one partner’s past choice in various settings. We consider short-lived and long-lived information with both free and costly natures. The chapter is based on the 2018 experiments at the University of York. We show that subjects rarely use costly reporting, even when there is a public record, but groups can foster cooperation norms by accumulating reported information over time.

Chapter 3 extends the discussion to subjective ratings, as well as free-form word-of-mouth reviews, by considering the long-lived costly case. The chapter is based on 2021 experiments at the University of Durham. The results show that both rating and review treatments lead to higher levels of cooperation than the baseline, with the review treatment interestingly exhibiting significantly greater levels of cooperation, than those of rating-only treatment. Additionally, similar rating habits were observed in laboratory and field experiments.