July 8, 2015 - 8:00amRyan Baker (Teachers College, Columbia University, USA)Plenary SessionPlaza Ballroom A-C60 MinutesVirtual Session
In the last decades, the power of data mining and analytics has transformed practice in field after field. In this talk, I will discuss how this trend is transforming instruction in higher education. Increasingly, large-scale data is available on student learning and interaction online. Much of this data represents student behavior in a fashion that is both longitudinal and fine-grained. This has allowed researchers to model and track many elements of student learning that were not previously feasible at scale: engagement, affect, meta-cognition, collaborative skill, and robust learning. In turn, these models can be used in prediction of long-term student outcomes, and to analyze the factors driving long-term success.
I will illustrate this potential with examples of how our lab has studied engagement and learning in several contexts, including students learning data mining from a MOOC, biology undergraduates learning genetics from intelligent tutoring systems, undergraduates learning how to program, and military cadets and medical residents learning core medical skills. I will discuss how we track disengagement, the processes leading to robust or shallow learning, and the factors leading learners to long-term participation in a community of scientific practice.