Retaining Students: A Field-Tested Framework of Prediction and Prescription скачать в хорошем качестве

Retaining Students: A Field-Tested Framework of Prediction and Prescription

Student retention is critical to the success of both students and institutions. Through a collaboration with all seven Wyoming community colleges, we have developed accurate predictive models of student retention. The models use existing data from student information systems (SIS) and learning management systems (LMS). The goal is to predict whether a student will persist into their second term with the target of providing this prediction within the first weeks of the semester. Models based on SIS or LMS data alone are less successful and we will share the engineered features that proved most contributory in distinct student populations. We will share pain points and learnings from developing data pipelines, engineering features, and developing and testing models within the community colleges. We will also demonstrate an applied theoretical framework for a prescriptive action recommendation engine designed to help select interventions for students identified at risk via the model.