Investigating social phenomena with agent-based models

Abstract

Many macro-level social phenomena are complex: they emerge from repeated interactions among interconnected individuals, and from the mutual feedback between these individuals and their environment. Complex social phenomena tend to be chaotic and highly sensitive to the behaviours of individuals and to their environmental conditions. This makes these phenomena difficult to study, understand and forecast. Furthermore, useful empirical data about complex social phenomena is often insufficient or unavailable. This chapter provides an overview of agent-based modelling (ABM) as a method of social inquiry to theorise and test mechanisms behind complex social phenomena, even when empirical data is scarce. ABM consists of computer simulations where social agents interact in autonomous, adaptive, and interdependent ways. We summarise ABM theory, present its ingredients and the inner workings of such a simulation model, and then explain how this type of model can be analysed meaningfully. We use two peer-reviewed applications examples of ABM for illustration purposes. Finally, we provide some general outlook and guidance for further reading into the topic.