Description of Instructor Intervention Using Individual Audio Data in Co-Located Collaboration

Abstract

Collaborative learning in face-to-face classroom settings requires instructors to monitor student groups and intervene with the group to ensure they make progress with the activity. One way learning analytics could help in facilitating such classrooms is by providing speech-based solutions to help instructors monitor this. In this paper, we investigate audio data collected from individuals in group work, processed using voice activity detection (VAD), used to describe student collaboration before, within, and after instructor intervention. Individual audio data were collected from classes focused on collaborative problem solving that were part of an introductory undergraduate engineering course. Analysis of 22 groups of individual audio data using VAD indicate that individual audio data could provide critical information on how students interact before, within, and after intervention from a facilitator using metrics like turn taking, turn overlap, and turn duration of individual students.