Preferences for Rights

Last registered on September 15, 2023


Trial Information

General Information

Preferences for Rights
Initial registration date
September 11, 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
September 15, 2023, 8:56 AM EDT

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


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

Request Information

Primary Investigator


Other Primary Investigator(s)

PI Affiliation
PI Affiliation

Additional Trial Information

In development
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Many political and philosophical debates over in-kind provision emphasize “rights,” a form of non-welfarist preferences — for instance, “Right to Counsel” for indigent legal defense and “Right to Health Care.” We conduct online experiments that test for and quantify the presence of non-welfarist preferences over allocating legal counsel and health care to tenants facing eviction.
External Link(s)

Registration Citation

Caspi, Aviv, Julia Gilman and Charlie Rafkin. 2023. "Preferences for Rights." AEA RCT Registry. September 15.
Experimental Details


We conduct four primary and two secondary elicitations, using spectator designs. Spectators are participants recruited using an online survey platform (Prolific or similar). Spectators are asked to allocate up to five separate goods to anonymous low-income people (“Recipients”). The primary interventions typically randomize the good itself, and we test for different behavior with the “treatment” goods (i.e., those over which we propose participants may have non-welfarist preferences) vs. “placebo” goods (i.e., those that have value to low-income people but for which spectators are unlikely to have non-welfarist preferences.)

The treatment goods are:
Good 1: Provision of health care.
Good 2: Provision of an attorney to a tenant facing eviction.

The placebo goods are:
Good 3: Bus passes
Good 4: YMCA memberships
Good 5: Cash (only in some experiments)

Primary elicitations
We list here, for each of the elicitations, what we randomize. See Experiment Details section for explanation of each elicitation.
Elicitation 1: Targeting. We randomize across Goods 1–5.
Elicitation 2: Inalienability. We randomize across Goods 1–4.
Elicitation 3: Optional value. We randomize across Goods 1–4. We randomize the value of cash between $200 and $300.
Elicitation 4: Universalism. We randomize across Goods 1–4. We randomize whether the number of people who have already received the good is 1, 5, or 9.

Secondary elicitations
Elicitation 1a: Valuation. We randomize across Goods 1–4.
Elicitation 1b: Information. We randomize across Goods 1–2. We randomize whether information provided is High or Low (see experiment details).
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Elicitation 1: Targeting
Indicator for choosing to allocate the good to all 10 Recipients
Number of Recipients whom the Spectator chooses to receive the good (indifference point)

Elicitation 2: Inalienability
Indicator for choosing the maximum price to avoid switching the lottery
Willingness to pay to avoid switching the lottery (indifference point)

Elicitation 3: Option value
Indicator for having a positive willingness to pay to ensure a choice
Willingness to pay to ensure a choice (indifference point)
Primary heterogeneity: willingness to pay to ensure a choice, conditional on having a high posterior belief that the recipient would choose cash (see experiment details)

Elicitation 4: Universalism
Amount willing to forgo from outside option to provide the good (indifference point)
Primary Outcomes (explanation)
Elicitation 4:
Note that our primary analysis will employ a “difference in differences” specification (see experiment details). Thus, the indifference point specified as a primary outcome is conditional on the number of recipients who have been allocated the good.

Secondary Outcomes

Secondary Outcomes (end points)
Correlations between Elicitation 1 outcomes and Elicitation 2, 3, and 4 outcomes (pairwise and joint)

Tagging people as non-welfarist. We will identify people as non-welfarist if they engage in non-welfarist behaviors in any, some, or all of Eliciations 2, 3, and 4. We will report these numbers.

Incentives. In Elicitation 1a, we employ a secondary treatment to test whether incentives affect behaviors. In 1a, we randomize whether we incentivize lawyers and placebos. The purpose of this randomization is to test the extent to which incentives matter (insofar as it could generate differences between lawyers and health care).

In a second test for the effect of incentives, we leverage the following source of variation. Because the presentation for the health care and lawyers varies slightly (including discussion of incentives), we randomize placebo goods into seeing the exact framing as lawyers versus health care, so some placebo goods are not incentivized throughout (see Experiment Details). This variation jointly tests the effects of presentation and incentives. Since we expect the presentation alone is small, it is also a test of the effects of incentives.

Elicitation 1a: Valuation
Willingness to pay for the good (indifference point)

Elicitation 1b: Information
Share of people who revise choice about how to allocate good
Magnitude of revision in choice about how to allocate good
Suggestive: instrumental-variables specification that instruments for the effect of beliefs on Elicitation 1-outcomes.

Secondary heterogeneity:
The relationship between political party and exhibiting non-welfarist preferences (behaviors in Elicitations 1–4)
The relationship between personal experience (with facing legal problems without a lawyer or not seeking health care due to cost) and exhibiting non-welfarist preferences (behaviors in Elicitations 1–4)
The relationship between support for Right to Counsel and health care and exhibiting non-welfarist preferences (behaviors in Elicitations 1–4)
The relationship between income and exhibiting non-welfarist preferences (behaviors in Elicitation 1–4). This heterogeneity is useful to explore because non-welfarist preferences may only be prevalent among the rich and/or those with high levels of education, in which case welfarist social welfare functions that aggregate up such preferences and place significant weight on “rights” would be regressive.
Secondary Outcomes (explanation)
Correlations. We seek to quantify how much of the targeting we see can be explained by behavior in the non-welfarist preferences elicitations.

Experimental Design

Experimental Design
Experimental Design Details
Not available
Randomization Method
Randomization done in Qualtrics.
Randomization Unit
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
1,800 individuals.
Sample size: planned number of observations
1,800 individuals.
Sample size (or number of clusters) by treatment arms
Unless otherwise stated, participants are randomized throughout the entire experiment into the following goods: N = 600 into health care, N = 600 into lawyers, N = 600 into placebo (300 bus passes, 300 YMCA). Placebo goods are additionally cross-randomized into seeing the same presentation as health care (p = 0.5).

Elicitation 1: Goods are randomized as follows. N = 600 health care, N = 600 lawyers, N = 150 bus passes, N = 150 YMCA, N = 300 cash.
Elicitation 3: We additionally cross-randomize 900 into cash = 200 and 900 into cash = 300
Elicitation 4: We cross-randomize the number Y out of 10 who are receiving the good. Y is cross-randomized with the good. 900 randomized into Y = 9. 450 randomized into Y = 1. 450 randomized into Y = 5.
Elicitation 1b: We cross-randomize whether information provided is High or Low. 600 randomized into High. 600 randomized into Low.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We are powered to detect MDEs of 10 percentage points for binary outcomes. We are powered to detect MDEs of about 100 dollars (of the outside option) for universalism difference-in-differences.

Institutional Review Boards (IRBs)

IRB Name
Massachusetts Institute of Technology Committee on the Use of Humans as Experimental Subjects
IRB Approval Date
IRB Approval Number