Evaluating Online Engagement and Narrative Feedback as Indicators of Student Performance

Abstract

AbstractLearning Management Systems for online learning in higher education allow for the collection of data that can describe student engagement and online behaviour. Student engagement refers to the degree of participation and active involvement with content by students and is often linked to academic achievement. Narrative feedback given by students can be used by academic staff to review and improve learning and teaching. Feedback can also be used as an indicator of student engagement and shown to correlate to student performance. Machine learning algorithms that predict whether a student will pass or fail, are trained by using student engagement features. Text mining techniques that help extract variables from narrative feedback are shown to improve these models for identifying at-risk students. The feature distribution of engagement factors indicate that there are patterns amongst the variables between the students who fail and those that pass.KeywordsLearning AnalyticsStudent engagementFeedbackText miningMachine learning