“Earlier” Warning Systems: Making the Most out of the First Signs of Student Underperformance

This Research-to-Practice Work in Progress addresses the need for predictors in educational environments to make reliable predictions of student performance as early as possible. This is the main foundation of early warning systems, tools that aim to detect students at risk of failing or dropping out of a course, and do so at a point in time at which it is still possible to execute effective interventions to help these students. This study, carried out in the context of a first-year university course, presents a classifier of students depending on whether they are expecting to pass or fail the course, using exclusively data available before any assessment activities are performed. This classifier is based on the Random Forest algorithm. The obtained results, while limited by the scarcity of data, show promise that it is indeed possible to detect students at risk of failing with acceptable reliability in the studied educational context.