Preferences for Rights: Mechanisms Experiment

The main task is an “inalienability” task, which we first conducted in AEA #0012065. Unlike AEA #0012065, the entire task in this follow-up is hypothetical. In this task, we tell people that one of two “recipients” (facing some legal issue, e.g., eviction) was allocated a lawyer in a lottery. The lawyer, legal issue, or recipient has certain features that we experimentally vary. These features are held constant across both recipients in the lottery, but we randomly vary the features of the lottery across participants.

The participant can choose to preserve the outcome of a lottery, or rerun the lottery. We tell the participant that rerunning the lottery would save money for the nonprofit. Both recipients are ex ante equal in the eyes of the participant. Thus, rerunning the lottery is more efficient, in the sense that it would save money for the nonprofit. However, rerunning the lottery risks taking the lawyer from one recipient and giving it to another. We tell the participant that the recipients will only know the final allocation, so they aren’t told about the initial results from the first lottery which might be rerun. Nevertheless, some participants might not like to rerun the lottery, for instance because it feels morally wrong.

There are three experimental manipulations, involving six arms in total. In our primary specification, we compare arms within each comparison. Secondary specifications pool arms 1 and 3 and compare with arms 2 and 4.
Comparison 1: Fine Magnitude. In this comparison, we manipulate whether the participant sees that the lawyer can address a speeding ticket fine of small or large magnitude ($50 fine [arm 1] or $2,000 fine [arm 2]).

Comparison 2: Lawyer Efficacy. In this comparison, we manipulate whether the participant sees that the lawyer is “ineffective and almost always loses their cases” (arm 3) and “effective and almost always wins their cases” (arm 4).

Comparison 3: Recipient Identity. In this comparison, we manipulate whether the participant sees that the recipient of the lawyer has income of $20,000 (arm 5) or $80,000 (arm 6).

The rationale for each intervention is as follows. First, comparisons 1 and 2 vary the "expected value” of providing a lawyer. Our goal is to see how moral behaviors (that is, not rerunning the lottery) change when the expected value, or stakes, of the decision change. Second, comparison 3 varies the identity of the recipients. Our goal is to see how moral behaviors change when the people affected by the choice change.

2024-08-20

2024-09-10

We focus on three outcomes:
1. Whether the person chooses to rerun the lottery (binary).
2. The amount that the person pays to rerun the lottery (continuous).
3. Whether the person chooses to pay the maximum amount to rerun the lottery (binary).

In our primary analysis, we study differences in each outcome separately for each comparison group. For instance, we will look at outcomes 1–3, among only the subjects exposed to arms 1 and 2 in comparison 1, and similarly for comparisons 2 and 3. That is, we will test for differences in outcomes 1–3, comparing arm 1 vs. arm 2; arm 3 vs. arm 4; and arm 5 vs. arm 6.

We focus on the following secondary heterogeneity, treatment effect on outcomes by:
Political party
Education

As a secondary test, we pool the tests in comparison 1–2 (that is, we compare arms 1 and 3 together, vs arms 2 and 4 together).

Our experiment proceeds in several parts. First, we randomize subjects to one of the six arms. They conduct the task where they choose whether to save (hypothetical) money, based on the conditions they see. Next, we ask subjects some demographic questions.

We recruit participants on Prolific.

Explanation of each arm and hypotheses. Comparisons 1 and 2 change the expected value of the decision. In arms 1 and 3, the expected value of receiving the lawyer is low. In arms 2 and 4, the expected value of receiving the lawyer is high. We want to see how moral decisions change when the expected value of receiving a lawyer changes. Do people become more or less likely to make non-welfarist decisions (that is, forgoing money not to rerun the lottery) when the “welfarist stakes” grow?

Comparison 3 changes the identity of people affected. In arm 5, the people are more vulnerable than in arm 6. Do people become more or less likely to make non-welfarist decisions when the identity of the recipient grows?

Given the results from AEA #0012065, we want to highlight that there is genuine uncertainty about the results from the Comparison 3 study. That is, we do not necessarily expect that having the lottery affect more vulnerable people will change whether people rerun the lottery. In AEA #0012065, we found that people made more universal (less targeted) allocations of lawyers and health care (what we called “rights goods”) than bus tickets or YMCA passes (“comparison goods”) when recipients differ in need. That might suggest that arm 5 and 6 will not induce significant differences in moral behaviors — moral reasoning may reduce the role of identity or redistributive motives, as suggested by the experiment on targeting behaviors. On the other hand, the differences in identity might change the welfarist stakes, and thus induce different propensity to engage in moral reasoning. It is an empirical question which force will dominate.

Randomization done on Qualtrics.

Individual

No

6,000 people

6,000 people

Each arm: 1,000 people