Multiplex Networks Provide Structural Pathways for Social Contagion in Rural Social Networks

2025-10-21

AI Summary

Traditional single-layer social network models overlook relational heterogeneity, limiting understanding of how distinct relationship types jointly drive health behavior diffusion in rural settings. Method: We construct a sociocentric multilayer network—spanning kinship, neighborhood, and close-friend ties—using 11 nomination procedures; introduce “network torque” as a novel metric quantifying each layer’s non-redundant contribution to critical transmission paths; and integrate exogenous intervention experiments with an innovative statistical framework to assess diffusion of newborn feeding knowledge, paternal postnatal care participation, and related health practices. Contribution/Results: The close-friend layer provides structurally unique transmission pathways that significantly increase knowledge accuracy (+23%) and enhance attitude change. These findings empirically validate functional differentiation among relationship types within multilayer architectures and demonstrate their essential structural role in collective health behavior change.

Technology Category

• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Application Domains: Social Networks
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Social Networks and Social Media: Social media analysis through the lenses of networks
• Web Mining and Content Analysis: Content-based information diffusion
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs

Abstract

Human social networks are inherently multiplex, comprising overlapping layers of relationships. Different layers may have distinct structural properties and interpersonal dynamics, but also may interact to form complex interdependent pathways for social contagion. This poses a fundamental problem in understanding behavioral diffusion and in devising effective network-based interventions. Here, we introduce a new conceptualization of how much each network layer contributes to critical contagion pathways and quantify it using a novel metric, network torque. We exploit data regarding sociocentric maps of 110 rural Honduran communities using a battery of 11 name generators and an experiment involving an exogenous intervention. Using a novel statistical framework, we assess the extent to which specific network layers alter global connectivity and support the spread of three experimentally introduced health practices. The results show that specific relationship types - such as close friendships - particularly enable non-overlapping diffusion pathways, amplifying behavioral change at the village level. For instance, non-redundant pathways enabled by closest friends can increase the adoption of correct knowledge about feeding newborns inappropriate chupones and enhance attitudes regarding fathers' involvement in postpartum care. Non-overlapping multiplex social ties are relevant to social contagion and social coherence in traditionally organized social systems.

Problem

• Understanding multiplex networks' role in social contagion
• Quantifying network layers' contribution to diffusion pathways
• Assessing relationship types' impact on behavioral change

Innovation

• Network torque metric quantifies layer contributions to contagion
• Statistical framework assesses network layers' global connectivity impact
• Identifies non-redundant diffusion pathways through multiplex social ties