Heterogeneity in probability perception and its effect on insurance decisions: How does partisanship affect our daily choices?

Study Procedure
Participants begin by reviewing and consenting to an informed consent form required to proceed with the research. On the next page, they complete a short "well-being survey," consisting of three questions about their current mood and general well-being (including whether they tend to be optimistic or pessimistic). Afterward, they encounter a welcome note, followed by an overview of the study and a detailed explanation of the rewards. Participants will also be briefed on the insurance terminology in the survey. Next, they are asked to complete an “example task”, earning $3 dedicated for use in the subsequent tasks. Upon completion, they move on to Stage 1 of the experiment.

Stage 1: Initial Exposure
Participants are introduced to a fictional neighborhood called "Maplewood Heights" in "Harborview City." Participants then watch three short videos, each approximately 50 seconds long, styled similarly to YouTube "shorts". These videos describe various aspects of "Maplewood Heights" and are labeled as content from Fox News, CNN News, or NeighborhoodSnoop. The order in which the videos are shown is randomized to prevent any bias based on viewing order. While all videos present similar data about the neighborhood, the crime rate might appear depicted with an upward or downward bias in one of three videos (always CNN or FOX). Participants are divided into five groups based on this bias: Republican Biased Upward, Democrat Biased Upward, Republican Biased Downward, Democrat Biased Downward, and Unbiased. After watching the videos, participants estimate the perceived crime rate in "Maplewood Heights" using a slider scale ranging from 0% to 10%. They then select the amount of money they would prefer with certainty to avoid a lottery scenario. Finally, participants choose an optimal insurance deductible based on the perceived crime rate. The responses from the biased groups are compared to those from the unbiased group.

Stage 2: Second Exposure with Bias Disclosure
In this stage, only participants from the treatment groups (all except the unbiased group) watch a second set of videos depicting the neighborhood a year later. Similarly, as in Stage 1, they view three short videos, which maintain the bias from the first stage but also include a disclaimer informing participants about the bias. After watching the videos, participants again estimate the crime rate, evaluate the insurance lottery, and choose an insurance deductible. These responses are compared to their responses from Stage 1.

Stage 3: Final Exposure without Bias Disclosure
In the final stage, only participants from the treatment groups (all except the unbiased group) view a third set of videos, once again describing the neighborhood a year later. These videos continue to reflect the bias from previous stages but without the disclaimer, the same as in Stage 1. Participants repeat the tasks of estimating the crime rate, valuing the insurance lottery, and selecting an insurance deductible. The responses in this stage are compared to those from Stage 2.

2024-12-05

2026-08-31

1. Effect of the Biased Partisan News on Crime Perception and Insurance Decisions
2. Effect of the Correction Signal on Crime Perception and Insurance Decision
3. Effect of the Correction Signal Retention on Crime Perception and Insurance Decision

Effect of the Biased Partisan News on Crime Perception and Insurance Decisions:
Participants will report their perception of the crime rate in the hypothetical neighborhood after exposure to partisan news sources. They will then select deductible amounts in a hypothetical insurance task, allowing the study to assess how perceived crime rates influence their insurance decisions.
Effect of the Correction Signal on Crime Perception and Insurance Decision:
The study investigates whether providing participants with information about the bias in the news sources improves the accuracy of their perceived crime rates and encourages more optimal deductible choices.
Effect of the Correction Signal Retention on Crime Perception and Insurance Decision:
The study also explores whether the effects of the correction signal fade over time when information about the bias is no longer reinforced, leading to a potential return to biased perceptions and suboptimal decision-making.

Participants are recruited on prolific and are asked to respond to a survey. The survey has three stages. In Stage 1, participants watch three short videos about specific neighborhood with varying crime rate biases, estimate the crime rate, evaluate an insurance lottery, and select a deductible. In Stage 2, biased groups watch updated videos with a disclaimer revealing the bias, repeating the same tasks to compare with Stage 1. In Stage 3, biased groups view videos without the disclaimer, repeating the tasks to assess the lasting impact of the correction.

Participants begin by consenting to an informed consent form and completing a brief well-being survey, followed by a welcome note, study overview, and explanation of insurance terminology. After an example task, they proceed to the experiment, which consists of three stages:
Stage 1: Initial Exposure
Participants watch three short, randomized videos about the fictional neighborhood "Maplewood Heights," styled like YouTube Shorts. Videos are labeled as Fox News, CNN News, or NeighborhoodSnoop and may show upward or downward crime biases based on group assignments. Participants estimate the crime rate, evaluate an insurance lottery, and select a deductible.
Stage 2: Second Exposure with Bias Disclosure
Treatment groups watch updated videos, including a disclaimer revealing the bias. They repeat the crime rate estimation, lottery evaluation, and deductible selection tasks, comparing results to Stage 1.
Stage 3: Final Exposure without Bias Disclosure
Treatment groups view another set of videos with the bias but without the disclaimer, repeating the same tasks for comparison with Stage 2.
Rationale:
Stage 1 compares biased groups (Republican/Democrat with upward or downward biases) to an unbiased group. In Stages 2 and 3, within-group comparisons assess the impact of bias disclosure and its subsequent removal. Only biased groups proceed through all stages, while the unbiased group participates in Stage 1.

The randomization method will be done by the computer, specifically in qualtrics. Each participant will be randomly assigned to a different treatment group.

The randomization unit is an individual participant.

No

The upper limit for participants is 150.

The upper limit for participants is 150.

The upper limit for participants is 150. That would make at most 50 participants per each group.