Learning experiments in a MOOC (Massive Open Online Course)

Abstract

We present results from two treatment / control experiments in the 8.MReV: Mechanics ReView massive open online course (MOOC) run on edX.org during summer 2014. We compare the efficacy of physics homework problems: (1) traditional physics problems involving many skills, (2) deliberate-practice activities that train individual skills using the drag-and-drop format, and (3) analogous deliberate practice activities in multiple choice format. Using a common assessment, our results suggest that traditional instruction is more effective than deliberate practice activities cast in the multiple-choice format; comparison of traditional problems and drag-and-drop deliberate practice is so far inconclusive. Some evidence suggests users prefer the drag-and-drop format over multiple-choice and are more engaged in such problems. In a separate experiment, we investigate the validity of the pre-test/post-test methodology in a MOOC environment where students receive feedback on the pretest and can view the correct answer after finishing a pre-test problem. It seems that little learning occurs during the pre-test and that exposure to a problem on the pre-test usually does not provide students an advantage on the post-test.