Cooperation in repeated Prisoner's dilemma experiments
How to cite
Gaechter, Simon; Sefton, Martin; Seidmann, Daniel; Fang, Tong. 2023. "Cooperation in repeated Prisoner's dilemma experiments." AEA RCT Registry. AEARCTR-0011387. https://doi.org/10.1257/rct.11387-1.0
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 study behavior in finitely repeated prisoner's dilemma experiments, where fixed pairs of players will interact for 100 rounds. The research question is how payoff parameters affect cooperation. Specifically, we vary the efficiency of cooperation and the temptation to defect in an orthogonal 2x2 design. The null hypothesis is that cooperation does not differ between treatments. The alternative hypothesis is that cooperation should be higher and more stable in "easy" games, where efficiency is high and temptation to defect is low, than in "hard" games, where efficiency is low and temptation is high. We expect cooperation in the other two treatments (high efficiency, high temptation; and low efficiency, low temptation) to be between the hard and easy games. We plan to recruit 40 pairs for each treatment, that is, a total of 320 volunteer participants registered with the CeDEx lab at the University of Nottingham. The experiment will be computerized with the experimental software zTree and conducted in the CeDEx lab.
Intervention
Participants will play 100 rounds of a two-person symmetric Prisoner's dilemma with a randomly allocated fixed partner. The prisoner's dilemma is characterised by four parameters: R(eward), T(emptation), P(unishment), S(ucker), where the following order holds: T > R > P > S and 2R > T + S. in each round, participants will simultaneously choose either C(ooperate) or D(efect), which results in four possible outcomes:
1) C, C, which results in payoffs (R, R)
2) D, C, which results in payoffs (T, S)
3) C, D, which results in payoffs (S, T)
4) D, D, which results in payoffs (P, P)
Our intervention is to run four treatments by orthogonally varying parameters (T, R, P, S) denoted in Experimental Currency Units (ECU) and by allocating pairs of participants at random to treatment. The treatments are described in the Experimental Design section below.
Our research question is how the payoff structure that we vary orthogonally affects cooperation rates.
Experimental design
We will run four treatments which orthogonally vary the prisoner's dilemma parameters (T, R, P, S) denoted in Experimental Currency Units (ECU) (see also the section on Interventions):
1) (600, 500, 200, 50)
2) (800, 500, 400, 100)
3) (800, 500, 200, 50)
4) (600, 500, 400, 100)
Participants will be randomly allocated to a partner and play the 100 rounds with the same partner. Pairs are randomly allocated to one of the four treatments, that is, we use a between-subjects design. Participants will play only one set of 100 rounds, after which the experiment is finished and followed by a short demographic exit questionnaire.
Participants will be paid according to all their decisions during the experiment. Accumulated ECUs are exchanged into British Pounds at an exchange rate of £0.01 per 25 ECU.
The experiments will take place in the dedicated CeDEx research labs at the University of Nottingham. Participants will be recruited from a database of registered volunteer participants. Participation is voluntary and subject to informed consent.
Design details
Cited by
No citation data available for this registration yet.