The effects of networks on cooperative actions in human groups in repeated prisoner’s dilemma games: which is better to promote cooperation between networks and communication?
How to cite
Yim, Kyubin; Hoffman, Elizabeth. 2021. "The effects of networks on cooperative actions in human groups in repeated prisoner’s dilemma games: which is better to promote cooperation between networks and communication?." AEA RCT Registry. AEARCTR-0007441. https://doi.org/10.1257/rct.7441-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
We are connected with others in the world. By using connections with others, we can strengthen cooperation and our community resilience. Without connections, it is not easy to cooperate with others. In this study, we focus on measuring the effect of networks on cooperative actions in human groups in the repeated prisoner's dilemma. In particular, we compare the cooperation performances of communication and networks in the repeated games. Also, we analyze the relationship between network robustness and cooperative actions in human groups in the repeated prisoner's dilemma. We design online experiments on Amazon Mechanical Turk.
Experimental design
We will do three experiments: the simple repeated prisoner's dilemma, the repeated prisoner's dilemma on a network, and the repeated prisoner's dilemma with cheap talk. Each experiment consists of 20 rounds with forty human subjects. All games are simultaneous games.
In the simple repeated prisoner's dilemma, two players are randomly matched, and they play the prisoner's dilemma game in each round. In the simple repeated prisoner's dilemma, all players do not know all information about others except on payoff structure.
In the repeated prisoner's dilemma on a network, the simple repeated prisoner's dilemma game is repeated from the first to the fifth round. From the sixth round, all players send proposals to others with whom they want to play the game before playing the game in each round. If the proposal is accepted, the proposer and the acceptor will play the game. All players use other players' reputation scores and the history of players' strategies during the past five rounds. The reputation score is calculated by the average times of a player's cooperative actions during the past rounds. In addition, they see the network in the previous round. Using the network information, they can see who was connected with more players and fewer players in the previous round.
In the repeated prisoner's dilemma with cheap talk, the simple repeated prisoner's dilemma game is repeated from the first to the fifth round. From the sixth round, players do cheap talk with free-form messaging with a partner randomly selected in each round. However, all players do not know all information about others except on the payoff structure.
Design details
Cited by
No citation data available for this registration yet.