How predictive analytics are opening up new possibilities in student success

Data analysis has traditionally been backward-looking—especially in higher education. But a cultural shift is afoot that’s focused on using predictive analytics to develop policies and interventions to help avoid potential obstacles to student success.

“The majority of data work has been focused on compliance and accountability—getting data to accreditors, the federal government, trustees, and agencies,” said Mark Milliron, Co-Founder and Chief Learning Officer, Civitas Learning, a company that offers a predictive analytics student success platform. “We joke that it’s ‘autopsy data.’ Imagine how much more useful that data would have been during the surgery, or even if there’d been diagnostic data to help prevent the surgery. That’s what we’re talking about—moving from accountability analytics to action analytics to help students avoid or handle these challenges before it’s too late.”

Predictive analytics has demonstrated its usefulness in helping students succeed. The University Innovation Alliance, a coalition of eleven public research institutions collaborating to accelerate student success innovation at scale, is working on a major predictive analytics initiative. Using lessons learned from collaborating campuses, the UIA helps schools avoid making the same mistakes so they can “move faster by moving together,” said Bridget Burns, UIA Executive Director.

When the UIA first launched in 2014, 3 campuses were actively using Predictive Analytics (Georgia State, Arizona State, and UT Austin).  In less than one year, 9 campuses within the UIA are now using Predictive Analytics.  One member, Georgia State University, has successfully employed a predictive analytics tool to develop policies and interventions that have successfully eliminated race and income as predictors of outcome, and doubled their graduation rate.

Burns and Milliron led an engaging and dynamic analytics panel discussion, moderated by Monique Snowden, AACRAO VP for Access and Equity, during the Monday plenary at the 2016 AACRAO SEM Conference in San Antonio. 

3 key concerns

For decades, the policy and praxis of SEM has been guided by “best practices”—What are leading SEM institutions doing, and how can you adapt and adopt those practices for your campus?

“A lot of the work in higher education in terms of data has been generalized data,” said Milliron. But the rapidly evolving field of predicative analytics is allowing more individualized and nuanced data analysis, which means that institutions can analyze and predict outcomes better than ever before—and act on that knowledge to improve student success.

Other considerations include:

  1. Choose partners wisely. “It’s important to find the right fit with your analytics provider,” said Burns. “There’s no panacea; no holy grail.”

Finding the right fit for your campus means asking questions like, “Who will be using the system?” and “What is your objective?” It’s important to connect the use of predictive analytics to your goal.

In addition, project management and onboarding are essential.

“You may have the most amazing product, but if the vendor doesn’t provide clear and useful project management and onboarding, it doesn’t matter,” said Burns.

  1. Visionary leadership is necessary. “Leaders must be experts in change management,” Burns said. “They need to know how to build a team and get people to get excited about solving the problem.”

“Leaders need to think about how to inspire teams, how to build a culture of evidence, and how to think collectively about policy change, practice change, getting data to the front lines—faculty, advisors, and so on,” Milliron said. “In some ways, institutional leaders have to become student success scientists, able to instrument and measure in close-to-real time.”

  1. Tailor your design for the front line. Know how to communicate with students and those who are talking to students.

Your data may identify an at-risk population, but it’s necessary to bring that data to the front lines in the right way.

“You don’t want a flashing red light saying, ‘This student won’t succeed,’” Milliron said. “You need to consider how to show the student what the next milestone is, what’s the path students like them have taken to succeed.”

“One of the most telling questions I’ve found for when I meet with academic advisers is ‘how many screens do you have to look at before you can advise a student?” Burns said. “I’ve heard numbers as high as nine. Academic advisors have a 30 minute appointment window. For some campuses, it makes a huge difference if we can get that number down to one or two screens for the folks who are coaching the student.”

Apps, swipes, and optimizations

These are only a few of the exciting conversations developing in the field of predictive analytics, which promises to explode in the next few years. And the useful application of information derived from this data could have a dramatic effect on student persistence and completion rates.

“It can mean the right nudge at the right time,” Milliron said. “Students and professionals can be more empowered than ever before.”

“We aim to build a movement where any institution can transform to be more effective and anticipatory on behalf of students,” Burns said, “If we can do that, as a country we stand a chance to close the achievement gap.”