Expert-led Debunking of Health Misinformation on TikTok

Paper Title

Expert-led Debunking of Health Misinformation on TikTok

Publication Info

• Topic area: Addressing health misinformation through expert-led interventions on social media.
• Keywords: TikTok, health misinformation, expert-led debunking, digital health interventions, misinformation correction, social media governance, video-against-video, credibility, user perception, debunking effectiveness.

Background and Problem

• Problem / challenge: Existing platform-led and user-led approaches to debunking health misinformation on social media are ineffective, especially for nuanced and complex health-related claims. These methods often lack transparency, credibility, and the ability to address misinformation in real-time.
• Significance: Health misinformation can lead to harmful outcomes for individuals and public health. Developing effective strategies to counter misinformation is critical to reducing its impact.
• Motivation and related work: Prior studies have explored platform-led and user-led debunking but found them inadequate for addressing health-related misinformation. Expert-led debunking, particularly on TikTok, has emerged as a promising alternative, leveraging features like stitching and duetting to directly counter false claims. However, its effectiveness and user perceptions remain underexplored.

Solution

• Proposed approach: Expert-led debunking using TikTok's stitching and duetting features, where certified health professionals directly refute misinformation in a video-against-video format.
• Novelty:
  1. Evaluation of organically created expert-led debunking videos in their natural TikTok format.
  2. Quantitative and qualitative assessment of user perceptions of credibility and effectiveness.
  3. Application of the "Debunking Handbook" framework to analyze the structure of effective debunking.
  4. Exploration of the socio-technical and governance implications of expert-led debunking.
• Procedure and key techniques:
  • Selection of six misinformation claims and corresponding expert-led debunking videos across general health, mental health, and nutrition topics.
  • Mixed-methods study involving a survey (n=420) and interviews (n=20) to assess user perceptions.
  • Quantitative analysis using Mann-Whitney U-tests to compare credibility and accuracy ratings.
  • Thematic analysis of open-ended survey responses and interview data to understand reasoning behind user assessments.

Results

• Concrete findings:
  • Expert debunkers were perceived as more credible than misinformation creators in 5 out of 6 cases.
  • Exposure to expert-led debunking significantly reduced belief in misinformation across all six claims.
  • The "Debunking Handbook" framework was effectively applied in most debunking videos, enhancing their impact.
• Advantage over baselines:
  • Expert-led debunking outperformed misinformation videos in credibility and accuracy ratings.
  • Unlike platform-led and user-led approaches, expert-led debunking provided transparency, contextualization, and a structured refutation method.
• Experiments / evaluation:
  • Participants were randomly assigned to view either misinformation-only or debunking-against-misinformation videos.
  • Videos were evaluated on a 5-point Likert scale for credibility and accuracy.
  • Thematic analysis revealed that expert credibility was linked to their professional presentation, use of evidence, and alignment with the "Debunking Handbook" method.
• Limitations and future work:
  • Limited to six specific claims and English-language videos, which may not generalize across all health misinformation or cultural contexts.
  • Excluded emotionally charged or stigmatizing content to protect participants.
  • Future work could explore the scalability of expert-led debunking across platforms and its effectiveness against emerging misinformation types, including AI-generated content.

Summary

This study evaluates expert-led debunking of health misinformation on TikTok, focusing on its credibility and effectiveness. Certified health professionals used stitching and duetting features to directly counter six health-related misinformation claims. The findings show that expert-led debunking significantly reduces belief in misinformation and is perceived as credible, particularly when following the structured refutation method outlined in the "Debunking Handbook." The approach offers a promising alternative to platform-led and user-led methods, with implications for social media governance and public health interventions. Future research should address its scalability, sustainability, and applicability across diverse contexts.