Using Learning Analytics to Predict At-Risk Students in Online Graduate Public Affairs and Administration Education

In this global information age, schools that teach public affairs and administration must meet the needs of students. Increasingly, this means providing students information in online classrooms to help them reach their highest potential. The acts of teaching and learning online generate data, but to date, that information has remained largely untapped for assessing student performance. Using data generated by students in an online Master of Public Administration program, drawn from the Marist College Open Academic Analytics Initiative, 1 we identify and analyze characteristics and behaviors that best provide early indication of a student being academically at risk, paying particular attention to the usage of online tools. We find that fairly simple learning analytics models achieve high levels of sensitivity (over 80% of at-risk students identified) with relatively low false positive rates (13.5%). Results will be used to test interventions for improving student performance in real time.