The effects of different types and sources of advice on promoting cooperation intergenerationally

Last registered on December 16, 2024

Pre-Trial

Trial Information

General Information

Title
The effects of different types and sources of advice on promoting cooperation intergenerationally
RCT ID
AEARCTR-0014994
Initial registration date
December 10, 2024

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
December 16, 2024, 2:03 PM EST

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

Locations

Region

Primary Investigator

Affiliation
ZheJiang University

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2024-12-18
End date
2025-01-18
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Sustainability and resource allocation are central concerns in modern economics. Ensuring the well-being of future generations is vital to the survival of families, organizations, nations, and even the global ecosystem. Ensuring that today’s consumption does not deplete resources for future generations is a crucial issue for policymaking. Moreover, the challenge lies in the fact that securing the future often demands sacrifices in the present. Artificial intelligence (AI) is starting to permeate economic and social life. It has demonstrated significant potential in optimizing decision-making processes, offering personalized financial advice, and reshaping how individuals communicate and interact. Given the importance of sustainability and the rise of artificial intelligence in the real world, we ask the research question of whether advice given by artificial intelligence can improve intergenerational cooperation.

Specifically, we employ advice generated by a large language model and compare the effect of AI-generated advice to human advice on promoting cooperation among non-overlapping intergenerational players.
External Link(s)

Registration Citation

Citation
lin, yangfei. 2024. "The effects of different types and sources of advice on promoting cooperation intergenerationally." AEA RCT Registry. December 16. https://doi.org/10.1257/rct.14994-1.0
Experimental Details

Interventions

Intervention(s)
Intervention (Hidden)
This is a between-subject design. Participants will receive advice from either humans or Chatgpt. The first-generation players will be asked to advise the next generation, including a recommended amount of effort contributing to a public good game pool. We will tell Chatgpt to generate advice based on the instructions shown to the first-generation players. The second-generation players the same game and leaves advice to the third-generation.
Intervention Start Date
2024-12-18
Intervention End Date
2025-01-18

Primary Outcomes

Primary Outcomes (end points)
Advice increases contribution to the public good game, and the effects of human advice and AI-generated advice on contribution are different.
Primary Outcomes (explanation)
We will analyse the context of advice in depth using NLP and sentiment analysis to see which kind promotes cooperation more than others.

Secondary Outcomes

Secondary Outcomes (end points)
Participants' contributions will be affected by the externality made in their advice. Participants who are asked to give advice to the next generation will contribute more than those who do not.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
This is a between-subject design. We introduce an intergenerational public good game with non-overlapping participants. After playing the game, participants in the first generation will be asked to give advice (incentive-compatible) to the next generation, and this procedure iterates. We plan to recruit participants for three generations.
Experimental Design Details
In the control group, participants will receive no advice before they play the intergeneration public good game. Regarding treatments, we consider two types of advice, private and public and two sources of advice: human and artificial intelligence. The private advice is only available to the specific player, and the public advice will be given to all players in the group. Participants will be randomly assigned to receive private advice or public advice. For private advice, participants will either receive private human advice left by the previous generation or private AI-generated advice by randomization. This randomization also applies to participants who receive public advice.
Randomization Method
Computer randomization.
Randomization Unit
individual randomization for treatments
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
700
Sample size: planned number of observations
700
Sample size (or number of clusters) by treatment arms
200 no advice, 200 human advice, 200 Chatgpt advice, 100 first generation for generating human advice
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
For one treatment group, we need 96 participants (95\% confidence level, 10\% precision level, and the maximum variation level (p=0.5, q=1-p)). Based on previous studies, we decided to have 200 participants per treatment (2 generations for each treatment).
IRB

Institutional Review Boards (IRBs)

IRB Name
Institutional Review Board of Laboratory of Neuromanagement, Zhejiang University
IRB Approval Date
2024-12-10
IRB Approval Number
N/A

Post-Trial

Post Trial Information

Study Withdrawal

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

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials