Using Diffusion Network Analytics to Examine and Support Knowledge Construction in CSCL Settings

Abstract

The analysis of CSCL needs to offer actionable insights about how knowledge construction between learners is built, facilitated and/or constrained, with the overall aim to help support knowledge (co-)construction. To address this, the present study demonstrates how network analysis - in a form of diffusion-based visual and quantitative information exchange metrics - can be effectively employed to: 1. visually map the learner networks of information exchange, 2. identify and define student roles in the collaborative process, and 3. test the association between information exchange metrics and performance. The analysis is based on a dataset of a course with a CSCL module (n = 129 students). For each student, we calculated the centrality indices that reflect the roles played in information exchange, range of influence, and connectivity. Students’ roles were analysed employing unsupervised clustering techniques to identify groups that share similar characteristics in regard to their emerging roles in the information exchange process. The results of this study have proved that diffusion-based visual and quantitative metrics can be effectively employed and are valuable methods to visually map the student networks of information exchange as well as to detect and define students’ roles in the collaborative learning process. Furthermore, the results demonstrated a positive and statistically significant association between diffusion metrics and academic performance.