When minds intertwine. Collaborative knowledge construction and group-level regulation in computer-supported collaborative learning

Collaborative learning has become widely recognized as a powerful way to support students’ understanding and problem-solving. It relies on students' ability to negotiate and solve disagreements to construct new or update existing knowledge together through social interaction. However, working together is cognitively, emotionally, and motivationally demanding. Productive collaboration does not emerge automatically; it requires both well-designed tasks and group-level regulation of learning, through which students jointly manage challenges, coordinate their thinking, and sustain engagement as a group.

This dissertation examines how collaborative knowledge construction and group-level regulation of learning are intertwined during collaborative learning. In particular, it explores how these processes unfold in relation to one another in different learning situations, and how task design shapes their (co-)emergence and quality.

The dissertation consists of three studies based on two empirical datasets. In the first dataset, 8th-grade students worked in small groups on a collaborative conceptual drawing task in physics. In the second dataset, 9th-grade students collaborated on a hands-on task in which they built a robotic arm designed to support the development of generic skills. Students’ interactions were analyzed using a process-oriented approach that combined detailed video analysis with speech and physiological data.

The results show that the relationship between knowledge construction and regulation of learning is dynamic and context-dependent. Regulation of learning often emerged in moments of cognitive conflict or socioemotional tension related to the task or to students’ attempts to build shared understanding. Task design played a central role in shaping when and how regulation and knowledge construction occurred, as well as how deeply students engaged in collaboration.

Methodologically, the dissertation demonstrates how combining video, speech, and physiological data can reveal patterns in both observable interaction and less visible psychophysiological responses. These patterns provide insight into how learners align, coordinate, and respond to one another during collaborative learning.

Overall, the dissertation contributes to a deeper understanding of how collaborative learning unfolds through the interplay of interaction, regulation, and task design. For educational practice, the findings highlight the importance of designing collaborative tasks that support not only cognitive learning but also the emotional and motivational dimensions of group work. The results also offer guidance for developing tools and scaffolds that better support students during collaborative learning processes.