ºÚÁÏÕýÄÜÁ¿

How Fairness, Effectiveness, and Intrusiveness Shape Public Support for Misinformation Interventions

Keywords: misinformation interventions, media literacy, social corrections, user behavior

 When it comes to fighting misinformation on social media, public support is far from guaranteed. New research finds that how fair an intervention feels matters more to Americans than whether they think it will work.

What factors affect user support for interventions against misinformation?

Recent research on combatting misinformation mainly assesses the effectiveness and popularity of interventions, but less work has looked into the factors that influence public support. Since social media platforms are unlikely to voluntarily implement measures that drive away users, understanding why people support or oppose these interventions is essential.

In this study, we surveyed over 1,000 active American social media users and asked for their opinions on 10 commonly studied misinformation interventions. Half the participants were told social media companies would implement the measures, while the other half were told the government would.

Finding 1: Fairness was the most important factor, followed by effectiveness and then intrusiveness

In our regression model predicting support, belief in fairness was the strongest predictor, followed by perceived effectiveness and then intrusiveness. Fairness was the most influential factor regardless of whether the interventions were implemented by social media companies or the government, and it was especially important among Republicans and men.

This emphasis on fairness is consistent across different policy areas, such as health care, the environment, and transportation, where perceptions of fairness and effectiveness often strongly influence support.

Finding 2: Transparent interventions that supported user agency were the most popular

Measures that increased information access and transparency, like labeling content or fact-checking ads, were more popular than those that imposed restrictions, such as removing misinformation or suspending accounts.

Table 1 shows the 10 specific interventions surveyed, including their type (e.g., whether they influence how accounts or content are moderated, distributed, or labeled) and their overall level of support. Measures with high support received over two-thirds of respondents' approval. Those with medium support were backed by between one-third and two-thirds of participants, while those with low support were backed by fewer than one-third.

Table 1: Overall support levels of all 10 interventions. 

Finding 3: Women and Democrats supported interventions more

When comparing support across various demographics, gender and partisanship showed the most significant differences. Figure 1 shows the average ratings for all 10 interventions by demographic group, where 1 indicates “Strongly Oppose” and 5 indicates “Strongly Support.” Similarly, a score of 5 for perceived fairness, effectiveness, and intrusiveness corresponds to “Very Fair,” “Very Effective,” and “Very Intrusive.”

     Figure 1: Overall average Likert ratings and 95% confidence intervals of all 10 interventions by party and gender.

Democrats and women tend to support interventions more and rate them as fairer, more effective, and less intrusive than Republicans and men do.

Takeaways

This research highlights the importance of understanding how everyday people feel about the misinformation they see on their social media platforms, and the measures they’d like to see taken against it. Overall, we found that:

1. Most misinformation interventions are widely popular:

• Five of the 10 interventions received average support ratings of around 4 out of 5 on the Likert scale. Over two-thirds of all participants, including more than half of Republicans, supported these interventions.
• Another four of the 10 interventions were rated as having a "Medium" level of support, but most had overall support levels around 60%.
• Only one intervention was rated as unpopular (temporarily preventing users from posting unviewed content). Interestingly, this measure is currently used on multiple platforms.

2. Fairness concerns drive lower support among men and Republicans:
   • Demographic differences were significant only for gender and partisanship, while factors such as age, income, and education were less relevant.
   • Men and Republicans are more likely to see interventions as unfair, possibly because they feel their groups are more often targeted.

Implications for Platforms and Policy-Makers

These findings point to several future directions that platforms and governments should consider when designing and implementing misinformation interventions:

• Design and test fair measures – Mitigating any disproportionate effects on specific groups is essential, since fairness is a crucial factor in gaining broader support.
• Promote user agency – Deploy interventions that are informational and give users choices. For example, warning users about contested content or prominently placing a fact-check nearby rather than removing the content without explanation.
• Build trust among users – Some user groups may initially doubt the fairness of interventions. Increasing transparency and providing clear access to an appeals process will help improve perceptions.
• Communicate effectiveness – Platforms can maintain a public dashboard, share data and reports with researchers and the public, and create educational tools. Previous research shows that increases in perceived effectiveness are directly linked to greater support.

Successful misinformation interventions will likely require collaboration between institutions. However, improving communication, education, and transparency could help garner broader support for these platform-based measures and encourage their more consistent application.