Heterogeneity and Social Norms in Collective-Risk Social Dilemma

How to cite

Li, Tongzhe; Xie, Huan. 2026. "Heterogeneity and Social Norms in Collective-Risk Social Dilemma." AEA RCT Registry. AEARCTR-0017765. https://doi.org/10.1257/rct.17765-1.0

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

Tongzhe Li* , Huan Xie

* Primary investigator

Registry ID
AEARCTR-0017765
DOI
https://doi.org/10.1257/rct.17765-1.0
First registered
Last updated
Status
In development
Trial dates
2026-01-29 – 2026-05-31
Countries
Private
JEL codes
C91, C92, H41
Registry record
http://www.socialscienceregistry.org/trials/17765

Abstract

We propose a lab experiment to study public goods provision and the effect of social norm elicitation in a threshold public goods game with heterogeneous agents. The game features a collective target and the loss of a portion of individual endowment when the collective target is not reached, mimicking social dilemma situations involving collective risk. Our experimental design introduces heterogeneity across agents in both endowments and loss ratios. The experiment examines 1) how public goods provision is affected by the correlation between heterogeneous endowments and loss ratios, and 2) the impact of social norm elicitation in this setting.

Experimental design

We propose a 2x3 between-subject experimental design, with 6 treatments in total. In one dimension, we vary whether heterogenous endowments and loss ratios across subjects are positively or negatively correlated. In the other dimension, we vary the procedure of eliciting the endogenous social norm.

In the social norm elicitation, subjects are asked to rate the appropriateness of different contribution rules and are incentivized to match their ratings with the modal appropriateness in the session.

Each session consists of multiple rounds of a one-shot threshold public goods game.

Treatment 1: Baseline under Positive Correlation
The game is based on a parameterization that introduces a positive correlation between heterogenous endowments and loss ratios across subjects. The appropriateness of different contribution rules is elicited once at the end of the session as a control.

Treatment 2: Baseline under Negative Correlation
The game is based on a parameterization that introduces a negative correlation between heterogenous endowments and loss ratios across subjects. The appropriateness of different contribution rules is elicited once at the end of the session as a control.

Treatment 3: Social Norm with Feedback under Positive Correlation
The game is based on a parameterization that introduces a positive correlation between heterogenous endowments and loss ratios across subjects. The appropriateness of different contribution rules is elicited multiple times across different rounds during the session, and the results of the social norm elicitation are provided to subjects after each elicitation.

Treatment 4: Social Norm with Feedback under Negative Correlation
The game is based on a parameterization that introduces a negative correlation between heterogenous endowments and loss ratios across subjects. The appropriateness of different contribution rules is elicited multiple times across different rounds during the session, and the results of the social norm elicitation are provided to subjects after each elicitation.

Treatment 5: Social Norm without Feedback under Positive Correlation
The game is based on a parameterization that introduces a positive correlation between heterogenous endowments and loss ratios across subjects. The appropriateness of different contribution rules is elicited multiple times across different rounds during the session, and the results of the social norm elicitation are not provided to subjects after each elicitation.

Treatment 6: Social Norm without Feedback under Negative Correlation
The game is based on a parameterization that introduces a negative correlation between heterogenous endowments and loss ratios across subjects. The appropriateness of different contribution rules is elicited multiple times across different rounds during the session, and the results of the social norm elicitation are not provided to subjects after each elicitation.

Design details

Randomization method
Randomization done by computer.
Randomization unit
Treatments are randomized at the session level. Each participant's player type remains the same throughout the session. Group members are rematched and randomly drawn each round.
Clusters
N/A
Observations
Approximately 500 individuals.
Arms
Approximately 80 per treatment arm.
Primary outcomes
Public goods provision, efficiency, equilibrium selection

Cited by

No citation data available for this registration yet.

Keywords