Enhancing Student Success by Combining Pre-enrollment Risk Prediction with Academic Analytics Data
Is Version Of
For nearly a decade, our institution has used multiple-linear-regressions models to predict student success campus-wide. Over the past three years, we worked to refine the success prediction models to the college of engineering (COE) students in particular, and to explore the use of classification and regression tree (CART) approaches to doing the prediction (e.g., Authors, 2016). In a parallel effort, our institution has contracted with an academic analytics company to do a retrospective analysis of student performance in every course as the university in relation to graduation rates. Here, we report on recent work we have done to make synergistic use of the results from the COE CART model and the academic analytics. Specifically, we have been able to examine student performance (i.e., grades) in core “success marker” courses as a function of the risk-grouping into which the CART model places them. We are now using this information to inform our advising. We provide details on these efforts, and on the opportunities and challenges provided by data-driven approaches to enhancing student success.
This proceeding is published as Raman, D. Raj, and Amy L. Kaleita. "Enhancing student success by combining pre-enrollment risk prediction with academic analytics data." Paper ID #18536. In 2017 ASEE Annual Conference & Exposition. 2017. DOI: 10.18260/1-2--28281. Posted with permission.