Contagion, fast and low: Modeling social influence in socio-ecological systems

Abstract

The concept of contagion defines how attitudes, emotions, or behaviors spread within a specific group. Traditional models of social contagion account for network characteristics that define the spread of simple or complex ideas based on the degree of exposure of susceptible agents to their affected “contact neighborhood”. However, these models understate two principal characteristics of real-world dynamics: (i) the competition among ideas of varying degrees of complexity (as observed in voting models) and (ii) the feedback-driven evolution of contagion complexity. Certain network features (e.g., network size, topology, neighborhood size) might favor the diffusion of competing simple or complex ideas, depending on their distinct characteristics. Furthermore, the dominant behavior disseminated may alter the agent’s operational environment, thereby modifying the cost of influencing additional agents. For instance, in co-evolutionary games such as the tragedy of the commons, encouraging agents to collaborate in the context of abundant and untapped resources requires a costly and complex contagion; conversely, in situations of resource depletion, prompting agents to reduce excessive exploitation is less costly—in line with the pathway of simple contagion. In this report, we develop an agent-based model that incorporates all of these aspects.