Agent-based social simulation provides modellers with critical new features, including the simulation of interactive human behaviour. This constitutes the ability to explicitly model human decision making processes, including perception, cognition, communication, and learning. This enables modellers to operate outside the narrow rational choice theory. These modelling features allow modellers to define systems at the grassroots level and to observe emerging macro phenomena. While this is a critical methodological advantage stimulating a rapid uptake of agent-based modelling in many disciplines, it also introduces a latent flaw for empirical applications. In empirical agent-based modelling real-world characteristics have to be translated into agent properties and behavioural changes have to be specified by numerical values regarding what triggers different behaviours and behavioural specifics. This session aims to present what techniques are currently in use to characterise and parameterise behavioural traits of human agents. The papers will present methodological details on how methods such as surveys, interviews, participant observation, and role-playing games, were used to characterise and parameterise empirical agent-based models.