Network Interventions to Limit Misinformation Spread on Reddit
Published:
Social media has become the primary news source in the United States, yet user awareness alone cannot prevent misinformation spread. This study investigates network-based interventions to limit misinformation propagation on Reddit using the FACTOID dataset of users and their interactions. We evaluate seed selection strategies using centrality metrics and implement a competitive cascade model where factual information competes against misinformation. Our key finding addresses the superspreader dilemma: while removing high- influence accounts is topologically optimal, it incurs substantial social costs. We compare two intervention strategies under a super-linear cost model: removing few influential users versus removing many smaller users. Results show that (1) out-degree centrality outperforms other metrics for seed selection, (2) misinformation spreads more efficiently than factual content, and (3) removing many smaller users achieves greater impact than removing few superspreaders under the same budget constraint. These findings suggest that targeting numerous low- influence accounts provides a more cost-effective approach to neutralizing misinformation spread than banning high-profile users.