Studying Oppression in the Lab

Last registered on May 27, 2022

Pre-Trial

Trial Information

General Information

Title
Studying Oppression in the Lab
RCT ID
AEARCTR-0009160
Initial registration date
April 22, 2022

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
April 28, 2022, 5:57 PM EDT

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

Last updated
May 27, 2022, 10:23 AM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Primary Investigator

Affiliation
Nuffield College

Other Primary Investigator(s)

PI Affiliation
Oxford University
PI Affiliation
Richmond University

Additional Trial Information

Status
On going
Start date
2022-04-22
End date
2022-06-03
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Human history is marred by countless instances of systematic oppression (economic and otherwise) even though oppression is objectively detrimental. Despite our extensive knowledge of the moral, social, and monetary costs of oppression, we know much less about its behavioral determinants, how it sustains, or its second-order effects. This is likely due to the inherent ethical and moral considerations imposed by attempting to create oppression in a controlled setting. To this end, we introduce a novel experimental framework that mimics the behavioral underpinnings of oppression in an ethical, replicable way. We use this new framework to study the relationship between social status, monetary incentives, and the willingness to oppress others. Further, we investigate how the experience of an oppressive regime impacts the trust exhibited by both its perpetrators and victims.
External Link(s)

Registration Citation

Citation
Bacine, Noah, Haley Harwell and Ryan Rholes. 2022. "Studying Oppression in the Lab." AEA RCT Registry. May 27. https://doi.org/10.1257/rct.9160
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2022-04-22
Intervention End Date
2022-06-03

Primary Outcomes

Primary Outcomes (end points)
public good contribution, punishment application, change in amount sent/returned in trust game before and after experience of oppression
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
relationship between behaviour in the experiment and demographic data collected in survey. The impact of experience during the experiment on self-reported self-esteem and strength of group attachment. The relationship between SDO and application of punishment points (oppression) during the experiment
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experiment has 6 different treatments based on the levels of 2 factors: Type Assignment [Random/Random_Reinforced/Earned_Reinforced] x Trust type [generalized/group specific]
Following some initial piloting, we have decided to focus on just two cells of our full factorial design during our first wave of data collection. We are currently collecting data for the Random and Random_Reinforced conditions with group specific trust. After we finish data collection, we intend to revisit the remainder of our proposed to design to reconsider other factors which may be of interest.

The Experiment takes place over 5 parts: 1-Type Assignment, 2-Trust Game, 3-Lopsided Public Good Game, 4-Oppression Game, 5-Trust Game

Experimental Design Details
Stage 1:
The first part of the experiment is adapted from (Ball, Eckel, Zame, 2001). Its purpose is to test the role that entitlement has in encouraging oppressive behaviour in the later stages of the game.

In the ‘Random’ treatment, participants are assigned types (1 or 2) at random and are informed of their type (but not what types signify until later in the experiment).

In the ‘Random_Reinforced’ treatment, participants are assigned types (Star or No Star) at random and are informed of their type (but not what types signify until later in the experiment). After participants are informed of their type assignment, an award ceremony is held in which Star participants are called to the front of the laboratory, given a star sticker to wear for the duration of the session and are congratulated by ‘No Star’ participants for their newfound status via a round of applause. Additionally, researchers give preference to ‘Star’ participants where applicable: Instructions are given to Stars first, experimenters are more attentive to questions from Stars, and instructions reflect a clearer difference in status between groups.

Subjects in the ‘Earned_Reinforced’ treatment are assigned types (Star or No Star) based on the outcome of a real effort task. Participants are told that ‘Top performers’ have been assigned to be ‘Star’ types and that the other participants are assigned to be ‘No star’ types. After participants are informed of their type assignment, an award ceremony is held in which Star participants are called to the front of the laboratory, given a star sticker to wear for the duration of the session and are congratulated by ‘No Star’ participants for their superior scores and resulting status via a round of applause. Additionally, researchers give preference to ‘Star’ participants where applicable: Instructions are given to Stars first, experimenters are more attentive to questions from Stars, and instructions reflect a clearer difference in status between groups.


Stage 2:
The second part of the experiment is adapted from Berg et al. 1995, commonly known as the trust game. A facet of oppression that is worthy of study is its implications for intergroup relations. As a result, we conduct a trust game before and after the ‘oppression’ stage of the game. In the trust game, participants are randomly assigned to a pair. In each pair, one participant will be the sender and the remaining participant will be the receiver. Each participant begins with 20 ECU. The sender chooses how many of their 20 ECU to send to the receiver (in multiples of 4 ECU). Any amount sent to the receiver is multiplied by three by the experimenters. The amount sent is considered a measure of trust. The receiver than gets to choose how much of the tripled amount plus their starting 20 ECU to send back to the sender. This is considered a measure of trustworthiness. To elicit a measure of trust for each participant, we ask participant’s to make decisions in both roles (sender and receiver). This employs the strategy method in which participants are asked to provide a response to every possible scenario they could face. At the end of the experiment, it is randomly determined which participant’s decision in each pair will be used in each role. The treatments vary whether the identity of the each participant’s counterpart (Star/Type 1 vs. No star/Type2) is revealed (specific) or not (generalized) prior to making decisions. In the specific treatment, the identity of participant’s partner in the trust game is held constant across decisions (i.e. sender/receiver) and parts (i.e.2/5). This treatment variation allows us to separate out the generalized and group specific components of changes in trust.

Stage 3: Lopsided public goods game

The purpose of the third stage is to introduce participants to the oppressive environment prior to giving one type the opportunity to oppress the other type. The game is an adaption of the classic VCM mechanism in which players must divide an endowment, e, between their private account which provides a return of 1 for every unit of the endowment invested and the public account which is multiplied by M and split evenly among group members. In our setting, participants are placed into groups of 5 (composed of 2 Star/Type 1 participants and 3 No Star/ Type 2 participants) and remain in the same group for five rounds. Each round, participants must choose whether to keep their 20 ECUs tokens in their private account or to invest them in the public account. Any token placed in the public account is multiplied by 2 (MPCR=0.4). However, unlike the standard VCM the public account is not paid out evenly. Each member of group A obtains R_A% of the public account while group B members only obtain R_B% (R_A= 0.37; R_B= 0.09 in both treatments). The rates were chosen to maximize earnings for No Star/Type 2 participants conditional on removing conditional cooperation as a viable profit-maximizing strategy. After each round, participants get to review the contribution decisions of their entire group and their total earnings before the start of the next round.

Stage 4: Oppression Game

In this stage of this experiment, participants play the Lopsided public goods game with punishment. However, unlike the standard VCM with punishment, only the Star/Type 1 participants have the opportunity to punish and only No Star/Type 2 participants are capable of being punished. This stage is identical to 3 except that after each participant makes their contribution decision, the Star/Type 1 participants get to review each participant’s contribution and decide whether or not to punish any of the No Star/Type 2 participants. To do so, they can choose to assign any number of punishment points to each of the No star/Type 2 members (conditional on not ending with negative earnings). For each punishment point assigned to a participant, the assigner’s income for that round is reduced by 1 ECU while the assignee’s income for the round is reduced by 10%. Of note, although each assigner can only assign a maximum of 10 points to any No Star/Type 2 participant, we allow participants to be assigned more than 10 punishment points in total. In this case (i.e. the sum of punishment points > 10), the assignee still earns 0 ECU for that round but the assigner’s still pay for all of the points they have assigned.

Stage 5

To study how the oppression game affects trust, we repeat the trust game from stage 2. Of note, we maintain the information given to participants about their partner (i.e. if they were not told the group affiliation of their match in stage 2 they will not be told in stage 5) and the group identity of the partner in the group-specific treatment (i.e. participants informed they are matched with a Star/Type 1 participant in Stage 2 will also be matched with a Star/Type 1 participant in Stage 5).

Post-Experimental Survey

After the conclusion of the decision making portion of the experiment, we ask participants to fill out a short survey while payments are being determined. The survey includes:
Demographic questions
Age
Gender
Sexual Orientation
Ethnicity
Income
Immigration to the UK
Political ideology (general, not party specific)
Beliefs about purpose of experiment and feelings over the course of the study (open-ended)
What did you think the purpose of this study was?
Please describe your feelings during the experiment:
Is there anything else you would like to say about your experience during today’s experiment
Type affiliation
Degree of attachment to type
Happiness/comfortability with type
Willingness to interact with types in the future
Rating of appropriateness of actions by different types
Questions about societal structure:
Social Dominance Orientation
GSS Fairness
Randomization Method
All eligible participants are invited to all sessions. Order of treatments across sessions is randomized conditional on preserving an even number of sessions for each combination of factors. Where applicable, randomization within the experiment is done via the programming interface, zTree.
Randomization Unit
Both of our factors: Type assignment & trust type are randomized at the session level. In the Random assignment treatments, participant type is randomized at the individual level.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
Following some initial piloting, Our current funding covers approximately 90 clusters of 5 participants
Sample size: planned number of observations
Each cluster contains 5 participants, for a total of 450 observations. Some aspects of the design involve repeated decision-making implying we will have 2250 observations of behaviour in part 3 and 4500 observations of behaviour in part 4. However, we are primarily interested in the punishment behaviour of Type 1/Star participants which is only observed in Part 4 and only amongst a subset of each cluster. Conditioning on this, there are a total of 2 observations of punishment over 10 rounds in each cluster implying a total of 180 observations based on our current funding.
Sample size (or number of clusters) by treatment arms
Following initial piloting, our goals is to get 64 observations in each of the two cells we are currently running (i.e. a total of 64 clusters & 320 participants). Once we complete data collection for the first two cells, we plan to decide which additional treatments of interest to run prior to applying for additional funding.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We aim to be able to identify a medium effect size (Cohen's d=0.5) with regards to total punishment over 10 rounds. To be able to do this, our power calculation suggests that we will need 64 observations in each treatment (i.e. 32 clusters per treatment, or 160 observations per treatment). Based on our initial piloting, we estimate a pooled SD of 26 implying we would be able to detect a mean difference of 13 penalty points over ten rounds. However, our piloting also suggested a large difference in variance across groups (SD=16,36). Accounting for this, we would only be able to detect a slightly larger mean difference of approximately 14 penalty points over ten rounds.
Supporting Documents and Materials

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

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
Nuffield's Centre for Experimental Social Sciences's Ethics Review Board
IRB Approval Date
2022-03-08
IRB Approval Number
N/A

Post-Trial

Post Trial Information

Study Withdrawal

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

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