Use of predictive analytics paying dividends at one university


The Degree Compass program was not meant to be a ‘self-fulfilling prophecy,’ Denley says.

Two years after first using predictive analytics to suggest courses and majors for students, officials at Austin Peay State University are reporting significant increases in passing grades across the institution.

Austin Peay, a four-year public university with more than 10,000 undergraduates, first used predictive analytics in 2011 to suggest academic paths for students, combining hundreds of thousands of past students’ grades with each student’s transcript to make specific recommendations for which courses and majors students would fit into best.

The technology is similar to the systems that generate recommendations on Netflix and Pandora based on a person’s movie and music preferences. Austin Peay’s predictive system, however, does not make recommendations based on the popularity of a certain class or major.

Tristan Denley, Austin Peay’s provost and vice president for academic affairs and designer of the Degree Compass program, said the school has seen notable increases in the proportion of students finishing courses with an A, B, or C, rather than a failing grade since adoption of the Degree Compass system in 2011.

This, Denley said, could be in part due to the software helping students avoid advanced courses that might prove too challenging before prerequisite classes.

Why did Pell grant recipients fare better in the predictive analytics program? See Page 2 for details.

Denley said there were no percentage breakdowns available, but passing grades had spiked by five standard deviations in a recent analysis of how the predictive technology has worked over the past two years.

Among Pell grant recipients, Denley said, the impact is even greater. There has been an increase of seven standard deviations in Pell students who have used the Degree Compass program.

“We can absolutely feel the impact on our campus,” Denley said. “It’s a real victory to see that providing these extra layers of information has helped students. … It was a logical step for us to inform choices within a major and to show students which majors to select, which majors they’d thrive in.”

The system selects classes that fit best with the sequence of courses in a major and are the most central to university curriculum. That ranking is then overlaid with a statistical model designed to predict in which courses the student will achieve the most success.

Students don’t simply plug in their academic information and wait for a series of viable options to pop up in the Degree Compass system. They work closely with campus advisors who guide students through the process of choosing classes and majors that fit their interests and skillsets.

The system predicts grades for each student in each class, but the student can’t see the prediction. “This is not intended to be about self-fulfilling prophecy,” Denley said.

“It’s not prescribing or restricting, it’s simply narrowing down choices and providing useful information,” he added. “It’s not just picking a major, it’s picking a future.”