Modeling contagion in policy systems

Scholars of the policy process offer compelling explanations for patterns in the aggregate-level attention of policymakers. Yet, we have little systematic understanding of the day-to-day behavior of these individuals. Why does a given policymaker, on a given day, decide to focus on one pressing issue while ignoring many others? I approach this question from a cognitive systems perspective and argue that policymakers are highly interdependent actors who are subject to cognitive limits and have incentives to closely monitor the political environment. These tendencies contribute to the emergence of widespread herd behavior in their individual attention to policy issues, a phenomenon I conceptualize as 'issue contagion.' I then utilize the methods of computational social science to build an agent-based simulation model of policymakers' issue attention over time. I also outline three empirical expectations regarding the density of communication ties between actors, the presence of segmented groups (e.g. political parties and coalitions), and the rate at which actors take cues from one another. Through a series of sensitivity tests, I document the internal validity of the model and show that incremental changes in network density, segmentation, and cue-taking all generate clear and visible trends in the frequency of issue contagion events.