Behavioral Insights to Tackle Elite Capture

Last registered on November 01, 2023

Pre-Trial

Trial Information

General Information

Title
Behavioral Insights to Tackle Elite Capture
RCT ID
AEARCTR-0012356
Initial registration date
October 26, 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
November 01, 2023, 4:00 PM EDT

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

Locations

Region

Primary Investigator

Affiliation
World Bank

Other Primary Investigator(s)

PI Affiliation
World Bank
PI Affiliation
World Bank
PI Affiliation
World Bank

Additional Trial Information

Status
In development
Start date
2023-11-01
End date
2024-01-01
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
The objective of this research is to test the effectiveness various interventions to increase cooperation and trust among Malagasy by exposing participants to a series of framings or questions that promote cooperative behavior. Our work aims to develop a chatbot-based game to achieve this objective of boosting cooperation and feelings of trust. In addition, we aim to understand the factors that drive competitive and corrupt behavior. The success of the intervention will be measured by comparing the performance in an economic trust game between the different interventions.
External Link(s)

Registration Citation

Citation
Afif, Zeina et al. 2023. "Behavioral Insights to Tackle Elite Capture." AEA RCT Registry. November 01. https://doi.org/10.1257/rct.12356-1.0
Experimental Details

Interventions

Intervention(s)
There are five study arms including four treatment and pure control.

Pure control: Participants read the basic instructions of the study and

Warm glow/injunctive norm: After reading the instructions of the study, participants are primed with the concept of 'fihavanana' to make salient cultural beliefs of mutual aid and generosity.

Descriptive norm: After reading the instructions of the trust game, participants are provided with information about how many tickets people typically send to other players.

Self-affirmation: Participants are asked to think of three qualities that make a great person, and reflect on a time when they lived up to one of those qualities themselves. The prompt for self-affirmation was modified from those most often used in psychological literature after feedback to make it more culturally appropriate, as it is not typical to think about one's own positive qualities in the same way.

Perspective-getting: Participants are presented with a brief narrative of a poor person in Madagascar who has struggled but exhibits good values.

Intervention (Hidden)
Intervention Start Date
2023-11-01
Intervention End Date
2023-11-22

Primary Outcomes

Primary Outcomes (end points)
Lottery tickets shared in round 1 of a trust game
Lottery tickets shared in round 2 of a trust game
Primary Outcomes (explanation)
Participants are always assigned to the role of "Player 1" and will be endowed with 100 tickets that count toward their chances of winning a 450,000 Ariary gift card for air time. In round 1 of the trust game, participants make the first choice of how many tickets {R1} to send to a "Player 2", which is our first outcome variable. The tickets that participants send will be tripled {R1*3} when received by Player 2 (e.g., if participants choose to send 30 tickets, player 2 will receive 90 tickets).

In round two, "Player 2" is endowed with 100 tickets and sends {R1} tickets to Player 1, which are then tripled {R1*3}. To complete the round Player 1 makes the final decision of how many tickets to share back with Player 2.

Secondary Outcomes

Secondary Outcomes (end points)
Zero-sum mindset
Perceptions of poor (or elite) people of Madagascar
Attitudes toward corruption
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Participants will be recruited through Facebook ads targeting the entire Facebook user population aged 18 and older in Madagascar. Clicking on the ad, participants will be directed to Facebook Messenger to complete the study through a chatbot. The participants will then be asked to provide informed consent.

Before the intervention, participants will complete a measure assessing generalized trust. Following this module, the instructions of the trust game and intervention prompts will be introduced. The trust game consists of two rounds with players switching roles in each round. Each round has two stages: (1) A player (the trustor) is endowed with 100 lottery tickets and has a choice to decide how many tickets to share with a second player. The number of tickets they send to the second player is then tripled. (2) The player receiving these tickets (the trustee) then has a choice of how many tickets to send back to the trustor.

After reading the game instructions, participants are provided an example of how the tickets are tripled. To account for potential anchoring effects, participants are randomly assigned to see an example with a low value of tickets (10 --> 30 tickets when tripled), mid-range value (50-->150 tickets when tripled), or high (90-->270 tickets when tripled). Participants then complete a series of comprehension questions to ensure they understood the instructions fully.

After completing the two rounds of the trust game, participants will be asked to complete the remaining survey questions measuring: zero-sum attitudes, perceptions of economic fairness, perceptions of the poor, social dominance orientation, and attitudes toward and perceived norms of corruptions.
Experimental Design Details
Participants will be recruited through Facebook ads targeting the entire Facebook user population aged 18 and older in Madagascar. Clicking on the ad, participants will be directed to Facebook Messenger to complete the study through a chatbot. The participants will then be asked to provide informed consent.

Before the intervention, participants will complete a measure assessing generalized trust. Following this module, the instructions of the trust game and intervention prompts will be introduced. The trust game consists of two rounds with players switching roles in each round. Each round has two stages: (1) A player (the trustor) is endowed with 100 lottery tickets and has a choice to decide how many tickets to share with a second player. The number of tickets they send to the second player is then tripled. (2) The player receiving these tickets (the trustee) then has a choice of how many tickets to send back to the trustor. In our paradigm, the participant is always assigned to the role of trustor in Round 1 and the trustee in Round 2.

After reading the game instructions, participants are provided an example of how the tickets are tripled. To account for potential anchoring effects, participants are randomly assigned to see an example with a low value of tickets (10 --> 30 tickets when tripled), mid-range value (50-->150 tickets when tripled), or high (90-->270 tickets when tripled). Participants then complete a series of comprehension questions to ensure they understood the instructions fully.

We use deception when describing "Player 2"--while Player 2's responses are programed as a bot, we tell participants they are connecting with another Facebook user from Madagascar. In addition, we inform participants that players are from the Southern region of the country, which is an area of high vulnerability and poverty, as we are particularly interested in understanding prosocial behavior toward poorer populations.

After completing the two rounds of the trust game, participants will be asked to complete the remaining survey questions measuring: zero-sum attitudes, perceptions of economic fairness, perceptions of the poor, social dominance orientation, and attitudes toward and perceived norms of corruptions.
Randomization Method
Randomization will be conducted automatically through the TextIt software used to implement the chatbot-based intervention. The software allows for random splits of users to different treatment arms.
Randomization Unit
Individual
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
N/A
Sample size: planned number of observations
1,500 individuals
Sample size (or number of clusters) by treatment arms
300 individuals per treatment arm
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
The minimum detectable effect with this design is 0.09 standard deviations difference in the number of lottery tickets shared.
IRB

Institutional Review Boards (IRBs)

IRB Name
Health Media Lab IRB
IRB Approval Date
2023-09-28
IRB Approval Number
2336

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