Social norm dynamics in a behavioral epidemic model

Christos Charalambous

Keywords: Agent-based modeling, Behavioral epidemics, Social norms, Coevolving dynamics, Vaccination dilemma

Abstract:

Understanding how social norms shape individual decision-making is central to many collective-action problems, from public health to environmental cooperation. We present an agent-based model that integrates epidemic dynamics with endogenously evolving social norms to study vaccination behaviour across repeated seasonal outbreaks. Agents interact on a multiplex network, where disease spreads on a physical contact layer and behavioural adaptation unfolds on an overlapping social layer. Vaccination decisions combine experience-weighted learning from realized and forgone payoffs with psychologically grounded norm dynamics.

Crucially, the model distinguishes between descriptive norms (beliefs about what others do) and injunctive norms (beliefs about what others think should be done), alongside personal attitudes. Norms coevolve through mechanisms of cognitive dissonance, social projection, and logical consistency, allowing beliefs, expectations, and actions to reinforce or destabilize one another over time. The relative influence of material payoffs and normative considerations adapts endogenously to perceived epidemic risk and local consensus.

Simulation results show that incorporating norm dynamics fundamentally alters epidemic outcomes compared to payoff-driven learning alone. Injunctive norms exert stronger and more persistent effects on vaccination uptake than descriptive norms, sustaining non-trivial coverage even at low infection risk. Moreover, interventions targeting injunctive expectations are robust, while interventions acting solely on descriptive norms can be weak or counterproductive.

By explicitly modelling the coevolution of behaviour, beliefs, and social expectations, this work highlights how different normative channels generate asymmetric and nonlinear collective outcomes. Beyond vaccination, the framework provides a generalizable approach for studying norm-driven dynamics in collective-risk and cooperation problems using agent-based modelling.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 8