Teacher regulation of CSCL: Exploring the complexity of teacher regulation and the supporting role of learning analytics

During computer-supported collaborative learning (CSCL), students collaboratively solve tasks while being supported by computers. CSCL is beneficial for collaboration because it offers students a platform for communication and a joint working space for solving tasks. All student activities are automatically logged and are available for review to teachers, offering them the opportunity to follow and diagnose the progress of all groups synchronously. Furthermore, they can send messages to the whole class or specifically intervene in one or more collaborating groups as they see fit, for example to give feedback or explanations. CSCL settings therefore offer unique affordances for teacher regulation, but also makes teacher regulation a challenging task because of all the available information. This thesis aimed to answer the questions how teachers regulate CSCL in terms of diagnosing and intervening strategies, and whether automated analyses and visualizations of student activities (learning analytics) can help teachers in this task of regulation.

The studies in Part I of the thesis illustrated how teachers diagnose student activities and how they explain the choice for particular interventions. Central concepts during this investigation were synchronicity and adaptivity, which means that teachers during CSCL try to tailor their support to the characteristics and needs of multiple collaborating groups that all engage in multiple types of activities. Teachers continuously monitor and diagnose the students’ activities, focusing their attention sometimes on the group level and other times on the class level again to make an announcement or to see whether any group needs additional help. The affordances of CSCL can help the teacher during this process. However, these affordances must be seen in light of the challenges that teachers face as well. In particular, the opportunity to monitor the collaborative process in real-time also means there is a challenge to maintain an overview of all the available information. With too much information available, adaptive teaching is hindered and teachers have less time to make informed decisions about interventions and about whether or not to intervene.

Learning analytics (LA) may support teachers because they analyze and summarize information about students. The studies in Part II of the thesis examined the effects of LA that either visualized students’ social activities (such as the occurrence of disagreement) or students’ cognitive activities (such as task progress). In the two studies, different effects of the LA were found, for example more frequent identification of the groups that experienced problems and in another study LA tools led to significantly more teacher interventions. Three mechanisms were proposed for these findings, namely that LA can aggregate information to a manageable level and thereby provide teachers with a quick overview of the situation, LA steer the teachers’ attention and make them aware of the shown information, and LA tools provide additional evidence to enhance the teacher’s diagnosis of the situation. When providing teachers with LA, it is important to consider that teacher beliefs influence the way the LA are interpreted and that a certain amount of training might be needed to achieve the intended effects.