Abstract

This paper discusses partial results of an ongoing project focused on analysing the current usefulness and implications of developing research on agent-based social simulation models beyond academic, hobbyist or educational purposes. Design, development and testing phases of such modelling are discussed along with common issues evidence-driven modellers often face whilst collecting, analysing and modelling quantitative and qualitative data into social simulations. It also includes a discussion on the evidence gathered in published literature and structured interviews with researchers that have lead mid- to long-term (3–5 years) projects on social simulation in Europe and the United States. Finally, good-practice recommendations are put forward by presenting a development methodology fully guided by evidence.