Online learning

Online learning is becoming more common in all levels of education, but questions remain about how best to use video and other tools to educate students in an equitable manner. Using the large datasets available from Harvard Medical School's online learning platform, HMX, we sought to answe rquestions such as "do online courses lead to real knowledge gains?" (they do, as shown in the graph below) and "what student behaviors can we use to predict whether they will succeed or fail in a course?" Many of these courses have been taught for several years with thousands of students enrolled, allowing comparisons to be drawn where smaller courses would be unable to spot differences. Similarly, online learning platforms collect a large amount of data, such as how many times a person (re)watches a video, whether they answer qustions right on their first attempt or if they use multiple tries, and the order in which they navigate the course. This fine-grained data allowed us to investigate correlations between a variety of student behaviors and their performance on summative assessment. While we used data from online courses in the biomedical sciences, many of the lessons learned are relevant to in-person teaching across disciplines, and can hopefully help all teachers improve their classrooms.