Communication, Guilt, and Agency Risk with Payoff Externalities

Last registered on August 10, 2023


Trial Information

General Information

Communication, Guilt, and Agency Risk with Payoff Externalities
Initial registration date
August 10, 2023

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
August 10, 2023, 1:43 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.


There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Purdue University

Other Primary Investigator(s)

PI Affiliation
University of Queensland
PI Affiliation
Monash University

Additional Trial Information

In development
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
This lab experiment investigates how different types of communication between principals and agents affect cooperation and efficiency in a setting where actions not only affect payoffs for the pair within the agency relationship, but also the earnings for an external principal-agent pair. Communication is hypothesized to impact beliefs, which in turn affects the psychological costs (e.g., guilt and lying aversion) that principals and agents feel when choosing their actions. A baseline treatment allows no communication opportunities. Two communication treatments introduce (a) a single, private message from the agent to the principal in each pair; and (b) preliminary free-form chat between both agents and principals whose payoffs are affected by actions, in addition to the single, private message in (a). Complete first- and second-order beliefs are elicited from all subjects following action choices. The experiment also elicits subjects’ risk tolerance and social value orientation.
External Link(s)

Registration Citation

Cason, Timothy, Lana Friesen and Lata Gangadharan. 2023. "Communication, Guilt, and Agency Risk with Payoff Externalities." AEA RCT Registry. August 10.
Experimental Details


Communication opportunities will be varied across treatments. This includes a no-communication baseline, and treatments in which the agent in each pair can send a single, private message to their principal before choices are made, as well as rich and free form (chat) communication between both agents and both principals whose payoffs are affected by actions.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Cooperation choices (frequency of IN for principals, and HIGH for agents)
Overall group cooperation rates (both principals choose IN and both agents choose HIGH in the four-person set of principal-agent pairs)
First-order beliefs
Second-order beliefs
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Communication content
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
This is a laboratory experiment, conducted in a standard university experimental economics lab, recruiting students broadly across the university. Subjects will be randomly assigned to treatments, and then randomly assigned to roles (agents or principals) and into pairs to make a choice.
Within each principal-agent pair, the principal chooses IN or OUT and the agent chooses HIGH or LOW (effort). The OUT choice corresponds to an outside option for which the agent’s choice is irrelevant; the IN choice leads the pair’s earnings to depend on the agent’s effort choice. These choices have payoff externalities for one other principal-agent pair, so final earnings depend on the binary choices of four subjects in each group. Efficient cooperation is achieved only if both principals choose IN and both agents choose HIGH.
Experimental Design Details
Not available
Randomization Method
Subjects will be recruited by email using ORSEE. They choose between a list of available sessions, and the session is randomized to a treatment before it is initialized.
Randomization Unit
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
80 groups of 4 subjects
Sample size: planned number of observations
320 individuals
Sample size (or number of clusters) by treatment arms
120 individuals in the no communication baseline
120 individuals in the single message treatment
80 individuals in the group chat treatment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Based on pilot sessions conducted in April 2023, as well as findings from the related treatments of Charness and Dufwenberg (2006). Proportion of cooperative IN/HIGH decisions that differ by 0.25 compared to the baseline and 0.3 difference between communication treatments can be detected with 80 percent power (5-percent significance level two-tailed).

Institutional Review Boards (IRBs)

IRB Name
Purdue University Institutional Review Board
IRB Approval Date
IRB Approval Number