Rumor propagation on social networks

Rumor propagation studies how unverified claims and uncertain information spread through social networks. Unlike factual news or deliberate disinformation, rumors often originate from genuine uncertainty and spread through natural information-sharing mechanisms. Research examines the structural properties of rumor cascades, the role of veracity and verification, and how fact-checking interventions affect spread.

Key distinctions from misinformation: Rumors are not necessarily false—they can be true, false, or unknown. The defining characteristic is uncertainty rather than falsity. This distinguishes rumor research from fake news detection, which assumes binary veracity.

Cascade dynamics: Rumors propagate through resharing, retweeting, forwarding, and copy-paste replication. The depth and breadth of rumor cascades often differ from typical viral content; rumor cascades tend to be deeper in network trees, suggesting greater contagiousness even from peripheral sharers.

Veracity effects: Empirical work shows surprising patterns—true rumors often achieve higher virality than false ones, despite false rumors being more frequently uploaded. This contradicts initial intuitions about misinformation contagiousness.

Fact-checking effects: External fact-checking links (e.g., Snopes.com) increase deletion likelihood for false rumors but have minimal long-term dampening effects, as many reshares occur after fact-check comments are visible.

Rumor evolution: Rumors mutate over time as they are copied, modified, and recirculated. Different variants compete for dominance, and network-level "voting" through propagation can select for more accurate versions.

Key papers

• Rumor Cascades — Large-scale empirical study of 16.7K rumor cascades on Facebook; shows rumor cascades run deeper than typical content; finds true rumors more viral than false despite false rumors dominating uploads; demonstrates Snopes comments increase deletion 4.4× for false rumors but have minimal long-term effect
• Detection and Resolution of Rumours in Social Media: A Survey — Authoritative survey covering rumor definition, detection approaches, stance classification, and veracity prediction; reviews PHEME and RumourEval datasets
• Stance Classification in Rumours as a Sequential Task Exploiting the Tree Structure of Social Media Conversations — Models rumor conversation threads as tree structures using Linear/Tree CRF for stance classification
• Tree LSTMs with Convolution Units to Predict Stance and Rumor Veracity in Social Media Conversations — Tree LSTM with convolution for joint stance and veracity prediction

Connections

• Cascade dynamics and prediction — structural and temporal properties of information cascades
• Information diffusion in social networks — broader study of how information spreads
• Fact-checking and corrections — interventions to reduce rumor spread
• Misinformation spread and diffusion — false claims propagation
• social context effects — how social ties influence rumor spread