How Exploitation Matters

Last registered on September 28, 2026

Pre-Trial

Trial Information

General Information

Title
How Exploitation Matters
RCT ID
AEARCTR-0019683
Initial registration date
September 28, 2026

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
September 28, 2026, 10:03 AM EDT

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

Locations

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

Request Information

Primary Investigator

Affiliation
ESMT Berlin

Other Primary Investigator(s)

PI Affiliation
Middlebury College
PI Affiliation
University of Warwick
PI Affiliation
Queen Mary University of London

Additional Trial Information

Status
In development
Start date
2026-09-28
End date
2026-10-19
Secondary IDs
Prior work
This trial is based on or builds upon one or more prior RCTs.
Abstract
We aim to elicit judgments of various scenarios that describe the interaction of two people and vary in how much exploitation there is. Those judgments are on those people's wellbeing, the wrongfulness of the behavior in the scenario, and norms around that wrongfulness. We will recruit two samples: a professional academic philosopher sample, and an online sample through Prolific.
External Link(s)

Registration Citation

Citation
Ferguson, Benjamin et al. 2026. "How Exploitation Matters." AEA RCT Registry. September 28. https://doi.org/10.1257/rct.19683-1.0
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2026-09-28
Intervention End Date
2026-10-19

Primary Outcomes

Primary Outcomes (end points)
1. Judgments on three categorical assessments of wellbeing (Satisfaction, Happiness, Flourishing)
2. Wrongfulness of behavior in the scenarios
3. Norm of wrongfulness of behavior in the scenarios
Primary Outcomes (explanation)
For each scenario presented:

1. Participants answer the question "To what extent does this scenario impact person A?" by responding via a 5-point Likert matrix where rows are Satisfaction, Happiness, Flourishing (shown in a randomized order) with options: falls significantly, falls somewhat, no impact, rises somewhat, rises significantly. They then respond similarly for person B.

2. Participants answer "How wrong is A’s behavior towards B?" by responding via a 5-point Likert scale with options: not at all, a little, moderately, significantly, extremely.

3. Participants answer "Which do you think is the most popular answer to the last question (How wrong is A’s behavior towards B?) among all other academic philosopher respondents?" where the words "academic philosopher" are replaced with "Prolific" if a participant is in the public sample. They do so by responding to the same Likert scale as in point 2 above. The question is incentivized (if they are right, they enter a prize draw).

Secondary Outcomes

Secondary Outcomes (end points)
Comparisons between samples on the primary and other secondary measures. Philosophical and political views; demographics.
Secondary Outcomes (explanation)
After the scenarios, professional philosopher participants answer philosopher-specific specialization, professional-position and philosophical-view questions, followed by demographics. Participants in the public sample report their political views and demographic questions. The exact questions asked can be viewed in the transcripts attached to this registry entry.

Experimental Design

Experimental Design
*Overview*
This is a vignette experiment in which participants – in this case, professional philosophers and Prolific participants – see a subset of three scenarios, with philosopher participants able to choose three additional scenarios, drawn from a pool of 12 scenarios in the way described in the Randomization Unit section below.

All the scenarios concern the hypothetical purchase, sale and production of mugs. They differ along various dimensions: the explicit exercise of power, the existence of inequality, background luck, the satisfaction of basic needs, an unjust source of power and disrespect. We introduced these scenarios in Ferguson et al. (2026), where participants reported how much exploitation they saw in them on a 0-100 scale. Based on the data in that study, we selected 12 vignettes for the current study to form three categories in terms of the level of exploitation participants saw in them: four low-exploitation, four medium-exploitation, and four high-exploitation scenarios. Using the coding in the Supplementary Materials of that paper (Section B available here: https://onlinelibrary.wiley.com/action/downloadSupplement?doi=10.1111%2Fajps.70058&file=ajps70058-supp-0001-Appendices.pdf), those are:

Low = {R3B, R3L, R1D, R3D}
Medium = {R3N, R2L, R4B, R3U}
High = {R2N, R4D, R4N, R4U}

We will provide summary statistics on the response data from the earlier study for these scenarios in the paper or its appendix.

For each scenario:
(i) Participants assess how the well-being of the two people changes as a result of the scenario for each of three categories of subjective well-being (Satisfaction, Happiness, Flourishing, which appear in a random order) on a five-point Likert scale (falls significantly, falls somewhat, no impact, rises somewhat, rises significantly).
(ii) They next assess how morally wrong they find A's behavior on a five-point Likert scale (not at all, a little, moderately, significantly, extremely).
(iii) They next report what they think the most frequent response to the wrongfulness question (of point ii) is.

The full transcripts (of the initial email confirmation page for the professional philosopher sample and the main survey for each sample) are attached to this registry entry.

An outline of our analysis plan is attached to this registry entry.

*Payments to participants*
Individual choices on subjective well-being and wrongfulness judgments are not incentivized. The norm-elicitation question is incentivized for both samples. The philosopher draw awards ten Amazon vouchers worth USD 100 (or the local-currency equivalent) each to eligible participants so long as they have not opted out; the Prolific draw awards ten bonuses worth USD 100 each. These are two independent draws.

Participation is compensated by a fixed payment of 4 GBP for the Prolific sample via the platform. For the professional philosopher sample, participation is compensated via a prize draw for Amazon vouchers worth 100 USD which will be sent to ten randomly selected participants from the set of respondents who complete the survey (so long as they have not opted out of the prize draw). This third draw is also independent from the others.

*Participants: philosophers*
We will recruit philosophers (academic philosophers and philosophy graduate students) online, through two well-known philosophy blogs (detailed just below); the Philos-L listserv (hosted by the University of Liverpool); and a direct email campaign. Participants responding via the blogs and the Philos-L listserv are required first to enter their institutional email address in an initial one-page form, which then emails them a personalized link to the main survey. This helps us verify their association with a recognized academic institution.

The blogs are Leiter Reports: A Philosophy Blog (https://leiterreports.com) and Crooked Timber (https://crookedtimber.org). Leiter Reports is the preeminent blog in professional philosophy, reporting “news and views about philosophy, the academic profession, academic freedom, intellectual culture, and other topics.” It is edited by Prof. Brian Leiter (Chicago), receives over 300,000 monthly visits, and is ranked 233,978 on SimilarWeb.com (as of 12 September 2026). Crooked Timber is a popular blog focusing on moral and political issues, with significant contributions from 15 core editors, many of whom are philosophers. It receives around 100,000 monthly visits and is ranked 580,800 on SimilarWeb.com (as of 12 September 2026).

A database of ~2100 professional philosophers’ emails was compiled from the addresses available on the websites of the world’s top-rated philosophy departments, as ranked by the Philosophical Gourmet Report (https://www.philosophicalgourmet.com).

After the scenarios, we ask professional philosophers about their areas of specialization, broad philosophical and political views, and demographic characteristics.

Professional philosophers may of course be from anywhere in the world. The locations we selected in the Trial Information section of this entry reflect the location of our Prolific sample, which we detail next.

We expect to receive approximately 550 responses after exclusions, as in Ferguson et al. (2026), but we will include all responses passing our exclusion criteria received in the specified time frame (see the Planned Number of Observations section).

*Participants: public*
We will recruit participants from the public via the online subject-pool operated by Prolific. We will restrict participation using Prolific's filters such that they are resident in one of the countries listed in the Trial Information (US or UK) and that English is their Primary Language, First Language, Fluent Language, and Earliest Language in Life. We will recruit 300 participants from each location (with the expectation that up to 50 are excluded by the criteria below in total over the two locations). This follows the roughly 50/50 composition of the public sample across these two locations in Ferguson et al. (2026). Following the primary questions, Prolific participants answer basic demographic questions (and are not asked opinion and professional questions relevant to philosophers).

*Exclusion criteria*
From either sample, we will exclude a participant's data if:
(a) their response time on any of the main question pages (covering points (i)-(iii) above) is less than 2 seconds;
(b) their response is flagged by Qualtrics' or Prolific's built-in bot detector;
(c) their IP address falls within a listed subnet of known VPNs and data centers (available here: https://github.com/X4BNet/lists_vpn, we will add an amendment to this registry with a permalink to the version that we apply);
(d) their response is incomplete.

For participants on Prolific, we exclude their data if:
(e) they fail the CAPTCHA (in which they are unable to complete the study);
(f) they fail two attention checks (in which case they are returned to Prolific).

From the philosopher sample, we will exclude a participant's data if:
(g) they provide a non-institutional email domain (we intend to collect data from academic philosophers and philosophy graduate students only);
(h) they declare themselves to be undergraduate students.

Data from participants meeting any of the criteria (a)-(h) will not be analyzed or presented. We will report how many exclusions there are, and of what type.

*References*
Ferguson, B., Matthews, P. H., Ronayne, D., & Veneziani, R. (2026). What exploitation is. American Journal of Political Science. https://doi.org/10.1111/ajps.70058
Experimental Design Details
Not available
Randomization Method
Implemented by the employed survey software, Qualtrics.
Randomization Unit
Participants in either sample initially answer the subjective well-being, wrongfulness, and norm-elicitation questions for three scenarios. One scenario is drawn at random from each of the low-, medium-, and high-exploitation categories, and the three are presented in random order. Participants in the philosopher sample may then choose to evaluate three additional scenarios. Those scenarios are drawn randomly from the remaining bank of 9 scenarios, such that they face an additional one from each of the three categories (low, medium, high), in a random order.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
No clustering
Sample size: planned number of observations
For our public sample, we will collect data from 600 Prolific participants (300 from each location) and expect up to 50 to be excluded with our criteria listed in the Experimental Design section. For our professional philosopher sample we expect to recruit several hundred participants; a similar recruitment strategy yielded 551 participants in Ferguson et al. (2026). As in that study, we do not restrict collection to a predetermined sample size due to the nature of the data collection process. Instead, we fix the amount of time the survey is live: the initial link will be posted on the blogs no earlier than September 28, 2026 (the exact date and time the blog owner posts the link is out of our control). We will send the link to our database of email addresses on September 28, 2026. We will close the study to new responses and close any unfinished responses at 09:00 UTC on October 19, 2026.
Sample size (or number of clusters) by treatment arms
n/a
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

Documents

Document Name
philosopher sample email confirmation (PDF)
Document Type
survey_instrument
Document Description
File
philosopher sample email confirmation (PDF)

MD5: 8995fce8608bf6abdf97743398e83e03

SHA1: 455bf6a6ec800fbe39563cad1519fd6c6856b35c

Uploaded At: September 28, 2026

Document Name
philosopher sample email confirmation (QSF)
Document Type
survey_instrument
Document Description
File
philosopher sample email confirmation (QSF)

MD5: 3fdc5da7461f592bc7d91baa8dd5a3bb

SHA1: fac87e83984a5d9169be193eba669d6be7ef397d

Uploaded At: September 28, 2026

Document Name
philosopher sample study (PDF)
Document Type
survey_instrument
Document Description
File
philosopher sample study (PDF)

MD5: 532dde12ef605012b6f43397a7d98fa3

SHA1: e0bd931bd64353eba5d17ac31efd5773c5df62f3

Uploaded At: September 28, 2026

Document Name
philosopher sample study (QSF)
Document Type
survey_instrument
Document Description
File
philosopher sample study (QSF)

MD5: 968213d969e9f7dd5fec90f3fcc28be0

SHA1: 1775b1717e96910789b25b6de2d4a5db2982d42d

Uploaded At: September 28, 2026

Document Name
public sample study (PDF)
Document Type
survey_instrument
Document Description
File
public sample study (PDF)

MD5: 170378b48962b9f087ff0f303351741d

SHA1: ae224185358b957cc272f4c1da0d552ef1c24e60

Uploaded At: September 28, 2026

Document Name
public sample study (QSF)
Document Type
survey_instrument
Document Description
File
public sample study (QSF)

MD5: 1ce2a20416a5157b8f6757bba75354ff

SHA1: 3256ca0d55f4e2ddef6df4a34b6f7e87e208be5b

Uploaded At: September 28, 2026

IRB

Institutional Review Boards (IRBs)

IRB Name
ESMT Ethics Committee
IRB Approval Date
2026-07-14
IRB Approval Number
26-001-E
Analysis Plan

Analysis Plan Documents

Analysis plan

MD5: edbba66d3c194140be9b718d3892f709

SHA1: 5cd00afa32eed3e2119ec00314e8871ee808e75d

Uploaded At: September 28, 2026