Source-credibility information and social norms improve truth discernment and reduce engagement with misinformation online
Abstract
Misinformation on social media is a pervasive challenge. In this study (*N* = 415) a social-media simulation was used to test two potential interventions for countering misinformation: a credibility badge and a social norm. The credibility badge was implemented by associating accounts, including participants', with a credibility score. Participants' credibility score was dynamically updated depending on their engagement with true and false posts. To implement the social-norm intervention, participants were provided with both a descriptive norm (i.e., most people do not share misinformation) and an injunctive norm (i.e., sharing misinformation is the wrong thing to do). Both interventions were effective. The social-norm intervention led to reduced belief in false claims and improved discrimination between true and false claims. It also had some positive impact on social-media engagement, although some effects were not robust to alternative analysis specifications. The presence of credibility badges led to greater belief in true claims, lower belief in false claims, and improved discrimination. The credibility-badge intervention also had robust positive impacts on social-media engagement, leading to increased flagging and decreased liking and sharing of false posts. Cumulatively, the results suggest that both interventions have potential to combat misinformation and improve the social-media information landscape.
Study specs
- Discipline
- Social Science
- Year
- 2023
- Human Data Platform
- Prolific
- Source
- View Source DOI Google Scholar
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.