Toward Using Multi-Modal Learning Analytics to Support and Measure Collaboration in Co-Located Dyads

Abstract

This paper describes an empirical study where the productive interactions of small collaborative learning groups in response to two collaboration interventions were evaluated through traditional and multi-modal data collection methods. We asked 42 pairs (N=84) of participants to program a robot to solve a series of mazes. Participants had no prior programming experience, and we used a block-based environment with pre-made functions as well as video tutorials to scaffold the activity. We explored 2 interventions to support their collaboration: a real-time visualization of their verbal contribution and a short verbal explanation of the benefits of collaboration for learning. This paper describes our experimental design, the effect of the interventions, preliminary results from the Kinect sensor, and our future plans to analyze additional sensor data. We conclude by highlighting the importance of capturing and supporting 21st century skills (i.e., collaboration and effective communication) in small groups of students.