Source-linked AI summary
Complex Contagions: A Decade in Review
Douglas Guilbeault, Joshua Becker, Damon Centola
TL;DR
Research on complex contagions addresses why some social behaviors do not spread like diseases through weak ties alone. This review synthesizes empirical work across four domains and related theoretical models, finding that diffusion depends on multiple reinforcing contacts, network topology, and threshold variation, and identifies directions involving interacting contagions, context, and diversity. It also highlights ethical concerns surrounding experiments that manipulate existing peer networks without explicit consent.
Problem
Universal contagion models did not explain why networks that spread infectious diseases could fail to spread preventive behaviors requiring contact with multiple activating sources.
Method
The paper reviews empirical studies in health, innovation, social media, and politics alongside theoretical models of network topology, thresholds, and diffusion.
Results
The reviewed literature shows that complex contagion diffusion depends on reinforcing contacts, network topology, individual thresholds, and social relevance across diverse domains.
Takeaways & Limitations
The review directs future research toward ecologies of interacting contagions, context-dependent thresholds, and the roles of diversity and homophily in diffusion.
Takeaways & Limitations
Experiments using existing peer networks can manipulate user behavior without explicit informed consent, creating an unresolved ethical concern.
Abstract
from arXiv · showhide
Since the publication of 'Complex Contagions and the Weakness of Long Ties' in 2007, complex contagions have been studied across an enormous variety of social domains. In reviewing this decade of research, we discuss recent advancements in applied studies of complex contagions, particularly in the domains of health, innovation diffusion, social media, and politics. We also discuss how these empirical studies have spurred complementary advancements in the theoretical modeling of contagions, which concern the effects of network topology on diffusion, as well as the effects of individual-level attributes and thresholds. In synthesizing these developments, we suggest three main directions for future research. The first concerns the study of how multiple contagions interact within the same network and across networks, in what may be called an ecology of contagions. The second concerns the study of how the structure of thresholds and their behavioral consequences can vary by individual and social context. The third area concerns the roles of diversity and homophily in the dynamics of complex contagion, including both diversity of demographic profiles among local peers, and the broader notion of structural diversity within a network. Throughout this discussion, we make an effort to highlight the theoretical and empirical opportunities that lie ahead.
1. Introduction
Social contagion research initially generalized disease-spread models to social and political diffusion, emphasizing the power of weak ties. Complex contagions challenged that view by showing that some behaviors require multiple reinforcing contacts, making clustering and social relevance central to diffusion.
- Social networks provide pathways through which collective behaviors such as norms, innovations, and movements spread.
- Universal contagion models applied epidemiological tools across ideas, information, behaviors, and diseases, emphasizing weak ties and network structures that accelerate diffusion.
- Complex contagions require multiple activating sources because adoption may depend on legitimation, credibility, uncertainty, or complementary value.A fax machine illustrates complementarity: several independent adopters can make adoption appear necessary, whereas one adopter provides little value.
- The simple–complex distinction explains why networks that spread infectious diseases may fail to spread preventive behaviors such as HIV prophylaxis.Simple contagions can transmit from one activated source, while complex contagions require contact with multiple sources.
- Network topology affects the two contagion types differently: shrinking worlds accelerate simple contagions but can hinder complex contagions.For complex contagions, insufficient clustering reduces reinforcing contacts and can prevent cascades entirely.
- The decade after 2007 expanded complex-contagion research across health, innovation, social media, and politics while advancing models of topology and threshold variation.The review identifies future directions involving interacting contagions, context-dependent thresholds, and diversity or homophily.
2. Empirical Advances
Empirical research finds complex contagions across health, innovation, social media, and political domains, while showing that reinforcement depends on peer similarity, structural diversity, content, and platform design.
- Health: Multiple peer exposures influence smoking cessation and exercise adoption, with homophily and diversity amplifying reinforcing signals.Health studies also show clustered networks can spread anti-vaccination norms while increasing vulnerability to measles outbreaks.
- Diffusion of Innovations: 56% greater uptake followed seeding based on complex contagion models than benchmark village-leader selection, while benchmark villages showed no diffusion in 45% of cases after three years.Controlled studies also found farmers’ crop adoption increased with more adopting neighbors and menstrual-cup adoption depended on multiple peers.
- Social Media: Social technologies including Facebook, Twitter, and Skype diffuse through reinforcement, while also shaping how other contagions spread.Facebook adoption increased with requests from multiple friends in separate network components, and peer influence accounted for most observed Twitter adoption patterns.
- Social Media: Across social media, complex diffusion varies by content: political hashtags require repeated exposure, whereas idioms, memes, and news hashtags can behave as simple contagions.Politically controversial hashtags are especially persistent, with repeated exposures retaining unusually large marginal effects on adoption.
3. Theoretical Advances
Theoretical work has expanded from network topology and threshold distributions to node-level attributes, interventions, and interactions among multiple contagions.
- Complex contagions require a critical mass for global cascades, with thresholds and degree distributions interacting sensitively with network topology.
- Node-level models examine how memory capacity, adoption thresholds, and synergistic neighbor effects constrain or shape diffusion.
- Intervention studies use oppositional nodes and seeding heuristics to inhibit harmful cascades, with 20-core selection outperforming highest-degree selection.The 20-core heuristic increases the likelihood that selected nodes are adjacent and can reinforce one another’s influence.
- Recent models study competing and cooperating contagions, showing that contagions can decrease or increase one another’s spreading probabilities.
- Health research documents ecological interactions among contagions, including competition between positive and negative health practices.
4. New Directions
New directions emphasize interacting contagions, context-dependent thresholds, and the distinct roles of identity-based and structural diversity in diffusion.
- Ecologies of Complex Contagions: Future research should examine ecologies of complex contagions operating within and across networks.
- Thresholds and Social Context: Future work should model how threshold structures and behavioral consequences vary across individuals and social contexts.
- Thresholds and Social Context: Social-media interfaces can change adoption costs, while political hashtags may require exposure from 2-5 peers and profile-picture changes up to 8 or more.
- Thresholds and Social Context: Identity, demographic characteristics, and social context jointly shape adoption thresholds and responses to contagions.Political responses can include joining a committed minority or punishing deviant behavior, rather than merely adopting or rejecting a behavior.
- Diversity and Homophily: Identity-based diversity refers to varied local peer profiles, whereas structural diversity refers to peers from separate network components.
- Diversity and Homophily: Homophily improved health-behavior diffusion, producing a 200% increase in overall adoption when similar peers reinforced one another.In the cited study, no obese individuals adopted in diverse networks, while homophilous networks matched the diverse networks’ total overall adopters.
- Diversity and Homophily: Structural diversity can increase credibility and social reinforcement, including when invitations come from separate components of an ego network.
5. Conclusion
Complex contagion research now spans diverse social behaviors and links diffusion to network structure, heterogeneous thresholds, and interacting contagions.
- Complex contagions occur across social domains, and topology plus adoption-threshold distributions can determine whether global saturation is possible.
- Key future questions concern multiple interacting contagions, context-dependent heterogeneous thresholds, and the roles of identity-based and structural diversity.