Leadership and cooperation in a sequential Prisoners’ Dilemma

How to cite

Lippert, Steffen; Feess, Eberhard; Tremewan, James; Martini-Tibbs, Jamie. 2022. "Leadership and cooperation in a sequential Prisoners’ Dilemma." AEA RCT Registry. AEARCTR-0010663. https://doi.org/10.1257/rct.10663-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Steffen Lippert* , Eberhard Feess , James Tremewan , Jamie Martini-Tibbs

* Primary investigator

Registry ID
AEARCTR-0010663
DOI
https://doi.org/10.1257/rct.10663-1.0
First registered
Last updated
Status
In development
Trial dates
2022-12-21 – 2023-07-31
Countries
Private
JEL codes
C91, C92, D02, D63, D90
Registry record
http://www.socialscienceregistry.org/trials/10663

Abstract

We analyze theoretically and with an online experiment on Prolific the impact of leadership in a sequential Prisoners' Dilemma. In the exogenous treatment E, one of the two subjects is randomly assigned to the role of the first mover (FM), and the other two the role of second mover (SM). In the transparent treatment T, one of the two subjects (the principal) is asked to decide between the role as FM and SM. The other subject (the agent) is informed about the principal's decision. In the opaque treatment O, the principal is also asked to decide between the role as FM and SM, but the agent is not informed that the order of moves depends on the principal's choice. We have developed a behavioral game theoretical model where subjects have private information on the weight they put on the payoff of their partner in case they cooperate. Our model yields predictions for the impact of the treatments on (i) the principal's choice (for treatments T and O), the frequency of cooperating FMs, and the frequency of SMs who cooperate after the FM has cooperated.

Experimental design

We have three treatments:
In the exogenous treatment E, one of the two subjects is randomly assigned to the role of the first mover (FM), and the other two the role of second mover (SM).
In the transparent treatment T, one of the two subjects (the principal) is asked to decide between the role as FM and SM. The other subject (the agent) is informed about the principal’s decision.
In the opaque treatment O, the principal is also asked to decide between the role as FM and SM, but the agent is not informed that the order of moves depends on the principal’s choice.
In all treatments, we apply the strategy method for SMs, that is, they are asked for their choices conditional on whether the FM cooperated or not. Principals as SM in treatments T and O are asked for this only after they have self selected to the role as FM or SM.
Each subject plays only one role, that is, we apply a between-subject design.

Design details

Randomization method
Participants will be randomly assigned to treatments and roles (computerized).
Randomization unit
Individual level randomization.
Clusters
No Clusters
Observations
We plan to collect # obervations
Arms
Treatment E: 300 Treatment T: 150 principals, 150 agents Treatment O: 150 principals, 150 agents
Primary outcomes
Our primary outcomes are the pairwise comparisons for the behavior of subjects between two treatments. With these pairwise comparisons, we test the hypotheses derived from our behavioral game theoretical model. These comparisons concern the (i) self selection of principals, the frequency of cooperating FMs, and the frequency of SMs who cooperate after the FM has cooperated. For the principal’s self selection, we have only one hypothesis: Hypothesis 1 (Self selection): The frequency of principals choosing FM is higher in treatment T than in treatment O. For FMs, our model yields the following order for the frequency of cooperative choices: The frequency is highest for principals who have self selected to FM in treatments T and O, followed by subjects in treatment E and agents in treatment O, followed by agents is treatment T. The pairwise comparison of these frequencies yields 10 Hypotheses. For SMs, our model yields the following order for the frequency of cooperative choices after a cooperative choice by the FM: The frequency is highest for agents in treatments T, followed by agents in treatment and subjects in treatment E, followed by principals in treatment O, followed by principals is treatment T. The pairwise comparison of these frequencies yields 10 Hypotheses.
Secondary outcomes
Our subjects are asked questions relaed to the dark triad (see #). Our first secondary outcome is the impact of the # on self selection, and on the choice as FM and SM. Our secondary outcome is the impact of demographics (age and gender).

Cited by

No citation data available for this registration yet.

Keywords