EXTENDING LEARNING ANALYTICS WITH MICROLEVEL STUDENT ENGAGEMENT DATA
Student engagement is a known contributor to academic success in online learning, and particularly in the highly demanded and challenging Science, Technology, Engineering and Mathematics (STEM) education. However much of the engagement data from the three main types of engagement; behavioural, cognitive and emotional remain uncaptured in a timely fashion and in-context. In terms of measuring softer qualities of student engagement which are harder to infer solely based on online behaviours, self-disclosed student data has a high value. Our proposal is therefore towards a microlevel student engagement (MSE) data capture from online students in order to gain timely data that could provide better insights into student learning and compare the findings with the more traditional approach of measuring student engagement at the end of a semester. At Universitat Oberta de Catalunya (UOC), Spain, we have conducted the first tests of engagement data capture in virtual classrooms of a programming course, using an engagement data capture module where students report their engagement with components in their online learning environment. By complementing this data with the data from a set of short questionnaires placed during course elements such as practicals and continuous assessments and the system-level log data captured at UOC, we have generated actionable knowledge such as student interest levels in the ongoing lessons, most engaged types of resources (videos, images, algorithms etc.) and lessons, cognitive and emotional engagement levels required by these different types of resources and lessons. Our goal is to further understand engagement levels and behaviours that help achieve academic success and thereby encourage richer and timely feedback and informed decision making by the teachers and the institutions and also provide students with an understanding of their own engagement trajectories.