A systematic analysis of the problem of MOOC learner dropouts

The openness and large-scale nature of massive open online courses (MOOCs) have attracted learners from all over the world. However, MOOCs are now facing the problem of high dropout rates. This paper uses social science systematic review methods to summarize and analyze 70 journal articles on MOOC dropouts from 2014 to 2024 to answer four questions: What is the research direction of MOOC dropouts? How to study the MOOC dropout problem? What are the factors that affect MOOC dropouts? How to predict MOOC dropout rates? Through sorting, it is found that the literature on MOOC dropouts is mainly concentrated in three aspects, namely, studying relevant policies or implementing corresponding measures to reduce MOOC dropout rates (accounting for 34.72%); studying important factors affecting MOOC dropouts (accounting for 27.78%); and predicting MOOC dropout rates (accounting for 37.50%). In the study of influencing factors, dropout factors are divided into 4 categories. Among them, the literature involving motivational factors among personal factors accounts for 73.08%. In the study of dropout prediction, researchers' research perspectives have also changed from the early simple binary classification to the current time series problems. In recent years, neural network algorithms have become the most mainstream prediction algorithms with a usage rate of 57.69%.