Designing early detection and intervention techniques via predictive statistical models—A case study on improving student performance in a business statistics course

ABSTRACTThis article presents a comprehensive study of factors that potentially impact student performance and success in a “bottleneck” college-level course in Business Statistics, with the goal of devising effective intervention methods to provide additional support to students who are at risk of failing. The latter are based on statistical models that predict the probability of failure based on relevant factors identified earlier. These models report high accuracy in detecting at-risk students as assessed by cross-validation techniques. Moreover, implementation of our techniques yielded positive outcomes, indicating that those students who took advantage of the intervention significantly improved their performance.