Massive Open Online Courses (MOOCs) are online courses that are offered for free using several platforms (such as Coursera, EdX, and Udacity). While learners in these courses do not receive official university credits, hundreds of thousands of learners register for these courses. The MOOC phenomena has led to a vivid discussion regarding their impact on higher education and the job market. However, one variable that is often left out of the conversation is quality of learning. As each MOOC often has tens of thousands of students, there is only limited or no interaction between learners and the course staff. Instead, learners interact with each other via discussion boards and peer feedback. Furthermore, learners often come from diverse populations and are not traditional university students. Thus, learners have a variety of goals from these courses.
In this talk, we seek to put learning in the forefront of the MOOC discussion. After reviewing existing research on learning with MOOCs, we will propose several metrics for defining successful MOOCs. Specifically, we argue that successful courses are ones in which students persist in the course, report to achieve their personal goals, and engage with a variety of learning activities (such as quizzes and discussion boards). We will demonstrate the usefulness of these metrics and begin to establish their validity. Second, we will identify course elements that correlate with productive learning. For example, as previously shown, longer videos seem to lead to reduced engagement. However, a closer examination suggests that it is the overall length of the videos per week, and not their individual length (or overall length per course), that matters. That is, courses with several short videos per week show similar engagement levels to courses with fewer longer videos. Another one of our findings suggests that size of support team does not appear to be critical. While some courses attempt to support learning by offering extensive support, our data suggests that the available support does not improve engagement, probably due to the overwhelming number of leaners. Interestingly, and somewhat counter intuitively, it seems that geographical location matters. That is, learners tend to prefer courses from nearby institutions, even though the courses are given online with no on-site component. We intend to demonstrate these findings using data from several institutions.
Participants in the session will gain better understanding of the varied institutional goals for MOOCs and means to achieve these. Findings that will be shared in the talk will facilitate the formulation of design guidelines for MOOCs. In addition, we anticipate that our methodological contributions will help participants in evaluating their own MOOC data, whether MOOC instructors, researchers, or designers. The session will include built-in time for comments and Q&A in order to receive feedback on our suggested measures of success and productivity of course elements. We hope that encouraging additional institutions to use similar metrics will facilitate a broad set of findings that will improve our collective understanding of the MOOC ecosystem.