The Promise and Pitfalls of Pro-Social Preferences and Empathy : Reducing Civilian Abuse through Perspective Taking and Incentives

Last registered on March 21, 2022


Trial Information

General Information

The Promise and Pitfalls of Pro-Social Preferences and Empathy : Reducing Civilian Abuse through Perspective Taking and Incentives
Initial registration date
March 21, 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
March 21, 2022, 1:15 PM EDT

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


Primary Investigator


Other Primary Investigator(s)

PI Affiliation
University of California, Berkeley
PI Affiliation
Kivu Agro Pastorale

Additional Trial Information

On going
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
What is the optimal amount of pro-social preferences and empathy in a violent organization? Whatever their mission may be, militias put their combatants and officials in a position where they can use threats of violence on civilians for private gain. In this study, we examine the role that empathy and self-interest play when active militia combatants consider whether to steal from and abuse civilians. Based on a relationship with a large militia, we answer this question by randomly assigning new combatants, shortly after they join the group, to one of three possible treatments. These psychological and economic interventions have been designed to train empathy, correct misperceptions about the private costs of abusing civilians, or to reduce the economic benefits of civilian abuse.

At the same time, combatants are officials of a nascent state organization, the militia itself. They collect taxes and enforce the law. It is when fulfilling these duties that they may abuse their position and take advantage of civilians - for instance, by threatening civilians in exchange for payment. Although promoting empathy may help reduce abuses of position, it may by the same mechanism interfere with the simple compliance of combatants with the rules of their position, which have been designed to fulfill the militia's mission of defending territorial sovereignty. For instance, feeling the pain of civilians when collecting taxes or enforcing the law may lead to greater leniency. Leniency would in turn diminish the organization's ability to finance its operations and to maintain social order. That is, empathy could be the basis of corruption. There is evidence for this concern – operationally, the militia deploys combatants away from their family out of concern that combatants who act as officials of the “militia state" near their communities will be more easily corrupted. This is the same reason why states agencies with staff in far locations regularly relocate officials – out of the fear that they may start developing relationships and thus their incentives become misaligned with the organizational mission. That is, the optimal amount of empathy that an organization must promote depends on the effect of empathy on the officials' tendency to engage in corruption, and on the organizational and social costs of low empathy.

The experiment is thus designed to analyze the tradeoff between fulfilling the mission (which requires taking actions that may be painful to take) vs. empathizing with civilians (which may come at the cost of sacrificing the mission, and thus corruption, but have the benefit of diminishing the risk of civilian abuse, another form of corruption).

External Link(s)

Registration Citation

Dunia Butinda, Lewis, Raul sanchez de la sierra and Hilary Yu. 2022. "The Promise and Pitfalls of Pro-Social Preferences and Empathy : Reducing Civilian Abuse through Perspective Taking and Incentives." AEA RCT Registry. March 21.
Sponsors & Partners

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

Request Information
Experimental Details


Empathy treatments:

1. A Perspective-Taking intervention, based on recent developments in moral psychology and psychology of empathy, and designed to promote combatant empathy. The intervention involves a 6 months-long weekly exercise with scenarios of 1 hour, inducing the combatant to imagine themselves in the position of other individuals of the community. The scenarios were developed by a local community leader with the guidelines provided by our team and local social workers, represented by a local visual artist, and edited by the research team.

Self-interest treatments:

2. Information about the private cost of civilian abuse. A training designed to improve the combatant understanding of the self-interested consequences of abusing civilians. The training is directly derived from the Geneva Convention and involves learning to international humanitarian law concerning specific actions with civilians. A random subset of the combatants assigned to this training will also receive training designed to share experiences from ex-combatants concerning the difficulty to re-integrate civilian life after militia participation

3. Manipulation of the economic benefits of civilian abuse. A livelihoods intervention involving the transfer of assets to the household of the selected combatants. The intervention is implemented by a local non-profit agricultural organization, and is designed to alter the combatant's marginal utility of household income, by creating income generating opportunities for their household. A major driver of civilian abuse, based on our qualitative work, was that combatants use their weapon to steal from civilians because they are worried about their household well being. The livelihoods intervention will be delivered 10 months after the rollout of the other two interventions.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Combatants' tendency to:
- steal from civilians
- use violence against civilians
- use excessive force against civilians in operations

Combatants' supervisor and peer performance evaluations

Primary Outcomes (explanation)
Combatants' behavior will be measured using:
- self-reports
- peer reports
- supervisor reports
- civilian reports in location in which combatants are stationed

These reports will be complemented with publicly available data which we will scrape from online and link to the combatant's deployment history

Secondary Outcomes

Secondary Outcomes (end points)
Combatants' promotions in the militia/career trajectory
Secondary Outcomes (explanation)
Different performance could translate into different career trajectories. Furthermore, the specific of the job may affect opportunities for making certain choices, hence it will be important to analyze career effects to examine whether it is possible to rule out that the effect of the interventions on performance and behavior arises from their effect on job type

Experimental Design

Experimental Design
Population: we developed a system based on the local civilian research coordinator that enables us to know the location of each recruit before they are on boarded and after they are deployed. Specifically:

-- We have trained the 160 HR commanders of the militia so that, whenever a new member is enrolled, they call our study coordinator. The coordinator records the information about the location of the new member and deploys a team of surveyors to that location. The surveyors conduct a baseline survey, and implement the treatment messaging as assigned by the randomization. This allows us to target the recruits when they are most impressionable, and to track them for more than one year once in the group. Our administrative data since 2015 indicates that, on average, the militia recruits 38 people every month. Thus, over a ten months enrollment period, we expect that our population will be 400 new combatants.

-- We have developed a system whereby research team members regularly visit the new combatant over the period to reinforce the relevant treatments, and to record their location in order to conduct endline measurements at the end of the intervention.

The treatments are assigned as follows.

Each individual who is joining the militia is part of the randomization to perspective taking and to training, to be given at the time they join:

-If the individual is assigned to the training, the surveyors (who have received prior training) provide a 30 minute long training session.

-If the individual is assigned to perspective taking, the surveyor gives the perspective taking booklet to the combatant, and trains the combatant how to complete one scenario each week for each of the following 6 months. The exercise takes 1 hour per scenario, and the completion will be verified regularly by a research team member. Successful completion entitles the household of the new member to a 100 USD equivalent in agricultural assets.

In addition, 10 months after they joined, individuals are randomly assigned for their household to receive an agricultural livelihoods support intervention. To implement the intervention, a team of members of the local non profit organization in charge of delivering the treatment visits the household of the new combatant (who joined already 10 months earlier). The intervention consists in a transfer of agricultural assets and a management plan to ensure that they translate into a steady stream of income. If the household consents, our team also informs the new combatant that their household received the intervention.

Outcomes will be measured through two channels:

- survey based outcomes will be measured at two intervals: after 10 months of joining (prior to the agricultural support intervention), and after 12 months

- publicly available data will be recorded in real time and linked to the location of the new combatants based on their reployment history which we will record through the surveys

Experimental Design Details
Each new combatant, at the time of joining the group, is assigned to receiving the training, the perspective taking, both, or none, totaling four treatment arms (40 join each month).

Then, after 10 months of joining, and of being assigned, we collect measurement of outcomes for each combatant. This allows us to isolate the effect of these two treatments prior to introducing the household agricultural support intervention, thus increasing power.

Then, we randomly assign the new combatants (now in their second year) to one of two groups: control, or agricultural livelihoods. Anticipating that the first two treatments could have lasting effects, we will stratify randomization of the agricultural livelihoods intervention with the groups defined by the treatments of the first year, thus obtaining 8 groups.

Overall, we expect to enroll up to 400 new combatants for these treatments.

Randomization Method
Randomization done in office by a computer (using Stata)
Randomization Unit
Individual new combatant
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
400 new combatants, enrolled over a period of 10 months
Sample size: planned number of observations
400 new combatants, enrolled over a period of 10 months
Sample size (or number of clusters) by treatment arms
Year 1:

100 new combatants control
100 new combatants information training only
100 new combatants perspective taking
100 new combatants perspective taking and information training

Year 2:

With agricultural livelihoods household intervention
50 new combatants control
50 new combatants information training only
50 new combatants perspective taking
50 new combatants perspective taking and information training

Without agricultural livelihoods household intervention
50 new combatants control
50 new combatants information training only
50 new combatants perspective taking
50 new combatants perspective taking and information training

Note: The information training and perspective taking are assigned only in the first year of the study. The agricultural intervention is assigned in the second year in the study, after we collected the first measurement of outcomes.

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Institutional Review Boards (IRBs)

IRB Name
Univeristy of Chicago's SBS-IRB
IRB Approval Date
IRB Approval Number


Post Trial Information

Study Withdrawal

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

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials