Supporting Social Interaction in an Intelligent Collaborative Learning System

TLDR

Collaborative learning improves learning outcomes when teams are active and well‑functioning, with students encouraging each other to ask questions, explain, justify, and reflect on their reasoning. The study proposes a collaborative‑learning model to enable an intelligent system to detect and address interaction problems within groups. The model defines indicators of effective collaboration, recommends improvement strategies, and guided the creation of two automated tools that code and analyze group conversations, comparing supportive and unsupportive groups. Empirical tests show that active, demanding teams produce effective learning, and that high‑level conversational analysis can automate group interaction assessment, supporting the feasibility of an intelligent collaborative learning system.

Abstract

Students learning effectively in groups encourage each other to ask questions, explain and justify their opinions, articulate their reasoning, and elaborate and reflect upon their knowledge. The benefits of collaborative learning, however, are only achieved by active, well- functioning teams. This paper presents a model of collaborative learning designed to help an intelligent collaborative learning system identify and target group interaction problem areas. The model describes potential indicators of effective collaborative learning, and for each indicator, recommends strategies for improving peer interaction. This collaborative learning model drove the design and development of two tools that automate the coding, and aid the analysis of collaborative learning conversation and activity. Empirical evaluation of these tools confirm that effective learning teams are comprised of active participants who demand explanations and justification from their peers. The distribution of conversational skills used by members of a supportive group committed to their teammates' learning is compared to that of an unfocused, unsupportive group. The results suggest that structured, high-level knowledge of student conversation in context may be sufficient for automating the assessment of group interaction, furthering the possibility of an intelligent collaborative learning system that can support and enhance the group learning process.