Fostering Online Discussions Through Social Learning Analytics
Abstract (101 words): Learning analytics are commonly reserved for the teacher. This paper reports on the initial phase of a design research program that aimed to devise social learning analytics to be used by students. The developed tool, CanvasNet, aimed to foster online discussions by turning discussion data into actionable insights for students. The first pilot, implemented in an undergraduate course, showed both early promise and significant challenges facing this effort. Future efforts in this area need to strive for tool simplicity, scaffold students to develop procedural and conceptual knowledge necessary for interpreting provided analytics, and bring analytics and pedagogical designs into greater coordination.
Word Count: 1995
This paper reports on a design-based research study aiming to devise a social learning analytics toolkit to foster online discussions in undergraduate classrooms. The developed toolkit, CanvasNet, combines social network analysis with lexical analysis to produce actionable insights into students' forum participation. By providing students access to CanvasNet and corresponding pedagogical supports, it is hypothesized that students would become better aware of their participation patterns and adjust participation accordingly. In the paper, we ground this study in relevant literature, introduce the research design, and report results from the first pilot.
Learning is fundamentally a social and cultural process (Vygotsky 1978). Knowledge is not discrete objects that can be obtained out of context; rather, it is situated in unique sociocultural scenarios and is negotiated within communities (John-Steiner 1996, Lave 1991). Therefore, in online courses where forum participation is deemed important, learning takes place as part of forum discussions and interactions.
Social learning analytics (SLA, Buckingham Shum 2012), a subdomain of the emerging field of learning analytics (Siemens 2013), is informed by the sociocultural perspective. As an effort to foster social learning, SLA draws on a “substantial body of work demonstrating that new skills and ideas are not solely individual achievements, but are developed ... through interaction and collaboration” (Buckingham Shum 2012, p. 5).
Analytical techniques applied in SLA---e.g., social network analysis and text mining---are not necessarily new. What is relatively new is applying these techniques to providing immediate feedback as learning is taking place. For example, SNAPP (Social Networks Adapting Pedagogical Practice) produces real-time visualizations of social interactions in discussion forums (Dawson 2010); CourseVis helps instructors to monitor student participation in online courses (Mazza 2007). What is still missing in this line of research is attempts to give students---instead of teachers---access to such analytics. As students as agents is becoming an ethics principle in learning analytics (Slade 2013), designing tools that directly engage students in choice-making seems imperative (Authors, 2016). This shift will incur significant changes with power dynamics in classrooms (Buckingham Shum 2012), and its effectiveness would rest on a range of contextual factors (Ali 2012). To ground future development in this area, it becomes crucial to explore challenges and issues in such efforts.