Social networks and online communities

Social networks are graphs of individuals (nodes) connected by relationships or interactions (edges) such as friendship, following, message exchange, or content sharing. In the context of misinformation research, social networks are the primary medium through which fake news, rumors, and false claims spread.

Key aspects of social network analysis for misinformation:

Network structure — Homophily (similarity-based clustering), community detection, bridge strength between communities, and degree distribution affect information flow and echo-chamber formation.

User roles — Different users occupy different structural positions: influencers (high follower counts), bridges (connecting communities), isolates, and core/periphery positions. Influence is not uniform—high-follower accounts disproportionately shape exposure.

Propagation dynamics — Information spreads via cascades. Early adopters and reputable sources condition downstream adoption. Time-sensitive viral phases differ from sustained diffusion.

Platform affordances — Twitter (retweets), Facebook (shares), Reddit (upvotes) create different incentive structures. Algorithm recommendations shape visibility. API limitations constrain what researchers can observe.

Key papers

• Rumor Cascades — Large-scale empirical study of rumor cascades on Facebook; analyzes cascade structure and depth
• Tacchini et al. (2017) — Some Like it Hoax — Analyzes Facebook user interaction networks to detect hoaxes; demonstrates that the diffusion pattern of user "likes" encodes reliable signals about post veracity, achieving >99% accuracy without content analysis.
• Soares, Recuero & Zago (2018) — Influencers in Polarized Political Networks on Twitter — Social network analysis using modularity, indegree, and outdegree metrics on Twitter conversations during Brazil's 2016 impeachment; identifies how influencer network positions correlate with polarization and echo-chamber reinforcement.
• Stella, Ferrara & De Domenico (2018) — Bots increase exposure to negative and inflammatory content in online social systems: applies spectral clustering (Fiedler vector) to identify polarized factions; analyzes human-bot network structure to show bots from periphery target human hubs and influencers; demonstrates strategic network positioning for maximum influence on human-to-human interactions.
• Brady et al. (2017) — Emotion shapes the diffusion of moralized content: analyzes retweet networks on Twitter; shows moral contagion effects are stronger within in-group networks than out-group networks across political topics.

Connections

• Information diffusion in social networks — how content spreads through networks
• Political polarization and ideological echo chambers — how networks become ideologically segregated
• Misinformation spread and diffusion — networks are the vehicle for false-claim propagation