How can we use data to support student success? One way is by using University data to reduce uncertainty about the future.
The Business Intelligence Unit, Metropolitan State University of Denver’s analytics-and-data-science team, has been busy building a statistical model to better understand and make predictions about whether current students will stay or leave. With this “machine learning” modeling complete, a campuswide effort is underway (using this data as the basis for strategic interventions) to promote student success in the form of retaining students on their way to graduation. The initiative is first focusing on retention into the fall 2022 semester, and data will be updated every semester.
Understanding the data
The model is calibrated using 15 years of data on student retention. Those behavioral patterns are then coupled with data on students’ demographics, academic performance and socioeconomic statuses.
Once adjusted to account for the ways that students have historically come and gone relative to these variables, the model is used to make predictions as probabilities. Right now, that means looking at last fall’s students and predicting which students are likely to retain into next fall and which students may be less likely to retain, based on historical retention trends.
Using the data
Over the next five weeks, prior to fall registration, this data will be used to strategically intervene to help students who:
- Have registration holds on their accounts
- Have financial hurdles to continued enrollment
- May need additional academic support
This collective effort will engage advisors, faculty members and leaders of multiple student groups to proactively work with students to address registration barriers as early as possible.
A new round of interventions will begin once fall-semester registration opens. Using the predictive model, the team will engage with students who have yet to register (excluding graduating students) to nudge them toward registration. This will include personalized outreach from faculty members, advisors and staff members, tailored to students based on patterns seen in the data.
Want to get involved?
To support this initiative, share your ideas or learn more about predictive analytics to promote student success, contact Sean Petranovich, Ph.D., director of Data and Analytics, at [email protected].