Predictive models for higher education dropout: A systematic literature review

School dropout is considered a complex problem and one that cuts across several levels of analysis. The development of predictive models has been a more dynamic and proactive response to tackle this problem. This research offers a systematic literature review on dropout prediction in higher education. The analysis period was from 2010 to 2020, searching scientific studies in six databases and working with a sample of 48 studies. The results indicate methodological and contextual characteristics of the cutting-edge literature on dropout prediction and enable the proposition of a research agenda for future studies. The analysis revealed an absence of research reporting or proposing management actions and educational policies that go beyond applying dropout predictive models.