The Data Revolution Comes to Higher Education: Identifying Students at Risk of Dropout in Chile
13 Pages Posted: 27 Nov 2017 Last revised: 14 Jan 2018
Date Written: November 19, 2017
Higher education has begun to use data and analytics to identify students at risk of dropout and to target and evaluate interventions. In this article, we present initial results from a model that uses data to predict student persistence at a Chilean university. The information available when students begin college predicts success poorly, and high school grades are more predictive than test scores or family background. Data from the first semester of college improves prediction dramatically. Data from the second semester improves prediction only a little more, and many students have already dropped out by the time second semester data are available. Some predictive variables suggest mechanisms for intervention, such as optimizing the distribution of fellowships and allowing more students into their first choice major. Colleges and universities should use first semester data to identify at-risk students and initiate targeted prevention.
Suggested Citation: Suggested Citation