Abstract: Instructors of online courses do not face their students directly and thus must rely on a different set of tools for gauging how a class is doing. Furthermore, the measurement of “how’s it going” is vastly different between a group of 25 students and a group of 300. Noisy outliers may lead an instructor to believe things are going well when they are not, or conversely think the class is not understanding things and progressing when in fact they are. In this work, we present two approaches for assessing the sentiment of a large population of learners without the benefit of face to face interaction. The first is a process for analyzing a large body of learner generated posts and determining the overall sentiment. The second is an approach to analyzing individual learners in order to target specific interventions to maximize their success. Leveraging the combination of these approaches will enable instructors to know how a very large body of students are perceiving the work to be performed as well as personalize intervention techniques based on the situation an individual is facing.
The full paper “Measuring Learner Tone and Sentiment at Scale Via Text Analysis of Forum Posts” can be found here.