You are here

Feature

Data analytics rising in higher education

A look at four campus “data czars” and how they’re promoting predictive analytics
University Business, June 2016
Clockwise from top left: Lisa Daniels (Excelsior College), Vince Kellen (University of Kentucky), Thomas Blum (Sarah Lawrence College) and Elaine Lewis (Washburn University).
Clockwise from top left: Lisa Daniels (Excelsior College), Vince Kellen (University of Kentucky), Thomas Blum (Sarah Lawrence College) and Elaine Lewis (Washburn University).

The momentum building behind higher ed analytics is hardly surprising, considering the explosion of tech tools available to make sense of institutional data.

What might surprise some, though, is who on campus is responsible for predictive analytics. Rather than assigning the task to institutional research or creating a new department, schools are tapping administrators in a number of different departments.

“The interesting story here is about the sheer number of people playing what would often be considered a traditional role in the institution—vice president of undergraduate affairs, for example—who are now in charge of designing, implementing and integrating predictive analytics,” says Bridget Burns, executive director of the University Innovation Alliance.

One of the goals for this coalition of 11 public research institutions is to replicate the success that one member, Georgia State University, has had using predictive analytics to increase retention and on-time completion.

Internal and external pressures also make it impossible for schools to ignore the potential benefits of analytics.

Internally, institutional leaders often want to move from “best guesses” about budget impacts to more accurate predictions about issues that may range from financial aid levels to course enrollment to retention.

Externally, there’s growing government pressure to meet mandated retention levels. Thirty-two states have either a formula or policy that allocatesa portion of funding based on course completion, time to degree, transfer rates and other performance indicators, according to 2015 National Council of State Legislatures research.

Lisa Daniels, assistant vice president for analytics, Excelsior College: The position was created just over a year ago to provide the time and freedom to use predictive analytics for student success and retention.

To get a picture of who is responsible for predictive analytics on campuses and what their jobs look like, University Business interviewed four campus “data czars” to learn more about their work, how it impacts their institutions and how they make it all happen.

How does your institution use predictive analytics?

Academic advising is the starting point for analytics at many schools. At Washburn University—an open-access public institution in Topeka, Kansas with nearly 7,000 students—the focus is on retention and graduation.

“We are continually working to better understand which students persist and which don’t,” says Elaine Lewis, director of success, evaluation and retention in Washburn’s Center for Student Success and Retention. “We’re learning, for example, that students who work on campus persist at higher rates.

“If we can get others employed in meaningful work on campus, we might be able to help them stay here,” Lewis adds.

The school has started by identifying students eligible for the Federal Work-Study Program who don’t have work-study jobs and contacting them about open positions.

Lewis is also working to expand the budget for student employment. “We try to be very intentional about new student positions created,” she says. “Employment on campus should always be a learning experience for students and an opportunity to gain new skills to expand academic critical thinking and workforce skills.”

Similarly, data analysis has revealed that students who live on campus are less likely to leave than are those in off-campus housing. So, the university is exploring ways to make it possible for more to live on campus.

Thomas Blum, vice president for administration, Sarah Lawrence College: Job responsibilities changed in September 2015 based on the need to expand the use predictive analytics to forecast enrollment.

Student success is also the goal at Excelsior College in Albany, New York, which is geared toward for adult learners. Assistant Vice President for Analytics Lisa Daniels is using analytics to quickly identify students who are at-risk yet engaged—versus at-risk and disconnected. Those who are most likely to respond to a coach’s outreach, for example, are engaged and will be more responsive to coaching and other assistance.

A decentralized analytics program at the 30,000-student University of Kentucky in Lexington touches on, among other areas, student success, financial aid decisions, course enrollment and staffing.

“The data can give you clues that guide some institutional decisions,” says Vince Kellen, senior vice provost for analytics and technologies. For example, students who are late filling out the FAFSA form or who register for fewer credit hours may be indicating they don’t plan to return and then the school can be more thoughtful about financial aid allocations. After all, aid awarded to a student who doesn’t return is aid that isn’t offered to a new student.

Better, more precise enrollment forecasting is the goal of Sarah Lawrence College’s expanded predictive analytics effort. Thomas Blum, vice president for administration, leads an initiative to predict the number of accepted students who will not only enroll, but also attend.

This is important for the budget of this small, tuition-dependent liberal arts college located just north of New York City.

“We had a clear signal from the president and the trustees last fall when we fell short of our target that we could implement what we needed to in order to have more confidence in the probability that the class we want to enroll will actually do so,” Blum says.

Vince Kellen, senior vice provost, analytics and technologies, at the University of Kentucky: His chief information officer position was expanded three years ago to reflect the growing importance of data and to include institutional research.

Who do you interact with and how?

Building relationships across departments, educating colleagues about the value of data, and reporting on outcomes are important parts of the job.

“A typical day for me involves a lot of conversations,” says Kellen. “I’m sharing our goals and objectives with others and getting consensus about what we want to do. I’m identifying organizational issues that are preventing progress and finding a way to deal with them.”

Blum invests time building and maintaining relationships with the people who provide the high-quality data that supports his mission. These include staff in admissions, financial aid and IT. They determine how to deliver the right data elements—fields and factors to analyze—and set it up so it’s automated.

Building a model to forecast enrollment took “a solid two months of effort,” he says.

More recently, before releasing the class list and sending out notifications, Blum met with the president, chief financial officer and dean of enrollment to discuss their confidence on the accuracy of the predictions.

“In the past, we reviewed the estimate for the class, but this year we were focused on probabilities in way we hadn’t before,” he says.

Lewis, who focuses on campuswide relationship-building at Washburn, has more student interaction than most because she also teaches a first-year student transition course. Her overall role is to ensure student success, it’s so it’s not all analytics. She meets regularly with military-veteran students to understand how the college can better support them. And she works through problems that may come up with her supervisor, the dean of the university library, as well as colleagues in IT and academic affairs.

Elaine Lewis, director of success, evaluation, and retention, Washburn University: Position was created about 18 months ago to allow more focus on using predictive analytics for success and retention.

At Excelsior, Daniels meets regularly with the two people she reports to in this new position, the chief information officer and the provost. Meetings with the CIO tend to focus on issues surrounding data transfer and process automation, while during sessions with the provost, they tackle how to integrate the processes they’re designing with other day-to-day operations in academic programs as well as possibilities and priorities for potential future projects.

Daniels also works with the IT department, deans and associate deans, as well as colleagues from her previous position in the strategy and institutional effectiveness department. Much of her time in her first year in this position has been spent developing and implementing models for identifying at-risk students.

After getting current plans to refer students as needed to the Student Success Center for help, she plans to shift to other projects. “One nearing completion now will provide deans with forecasts to assist in course section planning,” she says.

What advice do you have for those who are new to using predictive analytics?

The pioneers of predictive analytics have a keen sense of what has helped them the most—and what might help colleagues elsewhere. Building cross-disciplinary teams so that as many relevant people as possible understand the goals and process are key.

But, advises Blum: “Don’t do it in a vacuum. Everybody involved needs to be part of a team that’s working to reach the objective.”

“Everything is so new that when you make a request, people can be apprehensive because they don’t quite understand how you’ll use the data,” Lewis says. For example, she’s working to get access to student race and ethnicity.

“Our equal opportunity people are understandably cautious about sharing this,” she says. Daniels, at Excelsior, recommends networking with colleagues at similar institutions. Her school joined the PAR (Predictive Analytics Reporting) Framework, a nonprofit provider of predictive analytics.

“Members contribute data and get information from the aggregated data,” she says.

Daniels has also joined related professional associations including the Society for Learning Analytics Research and the Association for Institutional Research, and attended industry conferences. But she especially benefits from participating in relevant LinkedIn groups such as “Data Science Community,” “The R Project for Statistical Computing” and “Data Science & Machine Learning.”

Burns at the University Innovation Alliance underscores the importance of networking with colleagues at other institutions.

“If you can, find partners who are also undertaking this significant change. One of the benefits is that you have people you can call when you have questions,” she says.

At the University of Kentucky, Kellen stresses the need for collaboration within the institution.

“You don’t have to centralize everything,” he says. “The question is, how well are the people at the ground level enabled to collaborate freely? And how well are their leaders facilitating that?” Data analysts in different units at the university meet in person every two weeks to share information and problem-solve.

While colleges and universities may have different approaches to predictive analytics and how the information is used, they do tend to have one thing in common: A mandate from the top to solve a problem.

By using discussion, collaboration, training, software and communication, predictive analytics leaders are moving toward the goals they’ve established—and looking past them to how they can use what they’ve learned through predictive analytics in other areas of the institution to better predict behavior and outcomes.

Predictive analytics tools of choice

Burns, of the University Innovation Alliance, cautions administrators that “there are no plug-and-play” software solutions for predictive analytics.

“It’s not about chasing the shiny object, but about finding the system with effective project management and onboarding for your campus,” she says. “There’s no one-size-fits-all option. It’s about the right fit.”

  • At Excelsior College, Daniels uses QlicView from Qlik for business intelligence and R open-source software for predictive modeling.
  • Lewis at Washburn University uses Tableau from Tableau Software, while a research analyst in the school’s Institutional Research uses SPSS.
  • Blum, who uses Rapid Insight at Sarah Lawrence College, says the company’s customer support is a significant asset for him.
  • At the University of Kentucky, Kellen uses SAP HANA from SAP, which he describes as “ridiculously fast.”

Sandra Beckwith is a Fairport, New York-based writer.

Taxonomy: 

Register now for UBTech 2018

Register now for UBTech 2018, June 4-6 at the Mirage, Las Vegas. At UBTech 2018, you’ll network with a dynamic community of higher ed leaders who are shaping the future of campus technology and explore topics like cyber security, distance learning, campus learning space design, communications, personalized learning and more. Your UBTech registration also includes a free pass to the InfoComm exhibit hall.

Register now>>