Using Predictive Analytics to Allocate Educational Resources: An Application to Career and Technical Education

Abstract

Education leaders increasingly seek to integrate data into decision-making processes as they confront difficult choices about how to invest public resources. This paper assesses the consequences of data-driven decisions related to systems-level investments in career and technical education (CTE). Using Massachusetts administrative records, we evaluate various decision rules—from simple population-based approaches to sophisticated predictive models—for forecasting CTE concentration. We find investing in data quality yields greater predictive accuracy than complex modeling alone. Simulations indicate allocating resources based on model predictions could increase representation of educationally disadvantaged students in CTE, underscoring predictive analytics’ potential to enhance equity and efficiency simultaneously.