JOB MARKET PAPER
Who Shares More Fake News? Moral Judgment Explains Partisan Asymmetries in the Spread of Misinformation (Job market paper, with Alessandro del Ponte and N. Lim)
Abstract: This study investigates how political partisanship shapes the sharing of misinformation within a polarized media environment. Using a preregistered, incentive-compatible experiment with a nationally representative quota sample of Americans (N = 3,065), we exposed participants to political news headlines varying in veracity and party benefit in a 2 (veracity: true or false) by 2 (party benefit: help or hurt) between-subject design. Participants assigned to the role of Senders decided whether to share these headlines with Receivers. Subsequently, Receivers played a Dictator Game with Senders where they chose whether to send a reward. All participants expressed their moral judgments about sending the news and answered questions about their partisanship, ideology, and demographics. We find that Republicans were significantly more likely than Democrats to share fake news, regardless of whether it would help or hurt their party. A deeper analysis reveals that strong partisans were the most likely to share fake and helpful news, whereas weak partisans were the most likely to share true and hurtful news. Critically, we find that these behavioral differences are underpinned by distinct moral judgments. Democrats consistently expressed stronger moral condemnation of misinformation, while Republicans exhibited heightened moral condemnation of true information that challenges their in-group. This finding suggests that moral judgment is a key mechanism behind partisan asymmetries in information sharing, with partisan strength amplifying these differences. These findings enhance our understanding of how political identity and moral concerns influence the spread of misinformation and its impact on civic discourse.
WORKING PAPERS
Perceived Correlations in Risk Attitudes (with Fu Jingcheng and Zhong Songfa)
Abstract: This study explores the perception of correlations in risky choices within and across domains with different probabilities and outcomes. In an experimental setting, participants make risky choices and subsequently estimate the correlation between these choices among their peers. We find that participants generally perceive within-domain correlations as positive, albeit significantly overestimated compared to actual correlations, and cross-domain correlations to be closer to reality but exhibit considerable variability. Utilizing these empirical observations alongside a theoretical framework, we show that individuals tend to underestimate the size of noise in within-domain correlations and inadequately account for it in cross-domain evaluations. Our results support the notion that individuals use simplified mental models that overlook the role of noise in predicting choice behavior.
Loss Aversion: Loss Looms Larger or Gain Looms Smaller? (with Zhong Songfa)
Abstract: Loss aversion is a critical contribution of Prospect Theory to the study of decision-making. However, previous research on the construct of loss aversion in decisions under risk unveils mixed results. Some studies provide evidence for loss aversion, whereas others show a tendency for gains to be the same impactful as losses. We employ two representations, high-outcome-first framed and low-outcome-first framed questions, to explore this discrepancy. We find that the reflections are consistent with the fourfold pattern of risk attitudes in both treatment groups. That is, the outcome framing does not affect the persistence of the fourfold pattern. We further note that the extent of loss aversion is indeed greater when negative outcomes are presented first in the mixed prospects as opposed to when positive outcomes take precedence, though to a modest extent. Our study contributes valuable insights to the ongoing discourse on whether loss aversion is contingent upon contextual factors.