Characterizing MOOC Learners from Survey Data Using Modeling and n-TARP Clustering

MOOCs (Massive Open Online Courses) attract a diverse and large set of learners, with largely unknown learning needs and expectations. Researchers have been exploring why learners enroll in MOOCs and have found that learners enroll for a variety of reasons. Knowing who these MOOC students are is an important step in improving their educational experience and the value of MOOCs. It is therefore vital to identify and understand what distinct student groups exist in a MOOC, that is, to learn who they are and what they want. Pre-course surveys try to collect this information by asking students about who they are and what they want from the MOOC they are enrolling in. However, making sense of this survey data is challenging. Machine learning clustering techniques are a standard tool for identifying groups within data; however, two problems exist when trying to cluster survey data: (1) it is often not in a form easily interpreted by clustering algorithms and (2) survey data is frequently high dimensional, which standard clustering techniques cannot handle well. We describe a technique for converting survey data into machine interpretable feature vectors. We then propose analyzing the data using the nn-TARP clustering technique which is capable of efficiently finding multiple different cluster solutions and is scalable to high dimensional data. Using the proposed analysis approach on pre-course survey data from four MOOCs resulted in multiple distinguishable groups (i.e., clusters) of learners in each course, thus confirming the existence of many different survey response patterns. Additionally, these criteria persist between STEM and non-STEM courses. That is, we found learners grouped into similar clusters regardless of the course topic. The ability to separate learner types into distinct categories within and across courses is an important step in furthering the goal of enabling MOOC designers to design better online open educational systems that serve their diverse set of learners.