Improving evidence-based assessment of players using serious games

Abstract

Serious games are highly interactive systems which can therefore capture large amounts of player interaction data. This data can be analyzed to provide a deep insight into the effect of the game on its players. However, traditional techniques to assess players of serious games make little use of interaction data, relying instead on costly external questionnaires. We propose an evidence-based process to improve the assessment of players by using their interaction data. The process first combines player interaction data and traditional questionnaires to derive and refine game learning analytics variables, which can then be used to predict the effects of the game on its players. Once the game is validated, and suitable prediction models have been built, the prediction models can be used in large-scale deployments to assess players solely based on their interactions, without the need for external questionnaires. We briefly describe two case studies where this combination of traditional questionnaires and data mining techniques has been successfully applied. The evidence-based assessment process proposed radically simplifies the deployment and application of serious games in real class settings.