Using agent-based modelling to predict the effects of common classroom scenarios on student learning outcomes

This multi-disciplinary PhD research is focused on the design of a novel agent-based model and simulation that will emulate the autonomous interactions between students and their teacher in typical classroom lessons.

The model is driven by student, teacher and lesson characteristics, guided by unusually detailed, empirical data (collected using a custom-built program) and considers their effect on productive, disruptive and other types of behaviour.

If successful, the model should provide insights into the dynamics of classroom interactions, helping reveal the consequences of some common classroom interventions.

The data-collection tool and simulation will provide new types of data for lesson analytics (as contrasted to learning analytics and teaching analytics).

As an example, a successful model and simulation could be an excellent data-driven, decision-support tool, enabling educators to simulate the consequences of actions such as the use of a different lesson plan, the provision of a teaching assistant, or the rearrangement of seating. This would be particularly valuable in circumstances where an intervention cannot easily be tested in practice.

Keywords: agent-based modelling, agent-based simulation, social simulation, classroom behaviour, lesson analytics