Identifying teamwork indicators in an online collaborative problem-solving task: A text-mining approach

Teamwork is an important competency for 21st century learner. However, equipping students with an awareness of their teamwork behaviors is difficult. This paper therefore aims to develop a model that will analyze student dialogue to identify teamwork indicators that will serve as formative feedback for students. Four dimensions of teamwork namely coordination, mutual performance monitoring, constructive conflict and team emotional support are measured. In addition, the paper explores multi-label classification approaches combined with feature engineering techniques to classify student chat data. The results show that by incorporating linguistic features, it is possible to achieve better performance in identifying the teamwork indicators in student dialogue.