Human agency and infection rates: implications for social distancing during epidemics

Christopher Bronk Ramsey

In many ways, the best approach is to use large-scale epidemic simulation [2, 1] or to use a complex network model [3] , but it is worth testing these approaches against simpler models which can be used to explore aspects of the problem. In this paper, a totally theoretical probability-based approach is taken, which treats the population as a whole, and which specifically addresses variation in social interaction of individuals, with the aim of understanding how different human agents contribute to the spread. This model is not useful for studying geographic spread and, because it does not have any regional granularity, it is cannot address specific questions of how to deal with households or institutions, though in principle the model could be considered for institutions as approximately independent units, where infection rates are low. The purpose of the model is not to replace fuller simulation or network based studies but to provide a model which is simple enough for nonspecialists to use and which can be used to understand why some strategies might be more effective than others.