Towards understanding Intolerance: An Experimental Approach

Our experiment does not have an intervention. There are two experiments. In Experiment 1, we employ a randomized treatment variation that examines whether an individual is presented with conservative or liberal views for the vignettes described below. Experiment 1 was already pre-registered and conducted. The preregistration given below only refers to that of Experiment 2.
In Experiment 2, we use a two by two randomized treatment design. One treatment variation is, as in Experiment 1, about whether an individual is shown conservative views or liberal views for the vignettes; the other treatment variation is about the default number of people who will receive the vignettes if they are spread out.

2025-11-02

2026-04-30

1. Willingness to pay to prevent the spread of a view to either 1 or 100 people.
2. Allocation of points between different reasons for preventing the spread of an incongruent view

We have three sets of secondary outcome variables.
The first set of secondary outcome variables relate to survey-based measures of intolerance.
-Self-reported levels of intolerance based on survey questions
-Staircase-based method of measuring intolerance
The second outcome is the decision to cooperate or defect in a Prisoner's Dilemma game.
The third set of outcome variables are related to:
- Personal Norms
- Decision to cooperate in a prisoner's dilemma game
- Measure of pro-social behavior
- Other survey-based measures of intolerance

We will describe a series of four vignettes in which a decision-maker makes a decision and we will frame these vignettes in the form of social media posts expressing a view associated with each decision. We will evaluate the personal norms associated with each decision. We will ask the participants if they are willing to block the spread of the view expressed in the post, and how much they are willing to pay to block the spread of this view. Importantly, we will recruit individuals who have a conservative or a liberal view (defined in the PAP). We will investigate the reasons why individuals choose to pay to prevent the spread of a view when the treatment is incongruent with one's orientation (conservative vs. liberal) and to what extent this willingness to pay depends on the default number of posts' recipients. One randomized treatment variation comes from whether an individual is exposed to liberal views or conservative views. The other randomized treatment variation comes from whether the post is, by default, spread to a small (1) or a large (100) number of individuals.

Not available

Qualtrics-based individual-level randomization

individual

No

NA

We will use a representative sample from the US, excluding participants who already participated in Experiment 1. This sample will consist of 1000 participants recruited through Academic Prolific. We will categorize a participant as Liberal or Conservative, based on their self-reported measure of how they identify themselves on an 11-point liberal-conservative scale. In this sample, we will not include anyone who reports 5 (i.e., they are at the center) on this scale, or anyone who reports they would vote for neither the Republicans nor the Democrats, or anyone who would express an inconsistent ideological position and voting preference (e.g., a liberal who votes for the Republicans). Such participants will not be allowed to proceed after the initial screening questions.

500 in each treatment variation.
We will pool data from the treatment where the views are being spread to 1 person or 100 people when we compare liberal and conservative treatments.
Likewise, we will pool data from the liberal treatment and the conservative treatment when comparing the outcomes of treatments that spread views to either one person or 100 people.