The science of predicting retention

To help boost retention rates, advisers and colleges use predictive analytics of students.

The science of predicting retention

Anne Millerbernd

The University of Minnesota predicts which students are most likely to graduate.

By measuring incoming students’ high school grades, test scores, financial background and other factors — and comparing that information with similar students’ records from the past — colleges and advisers work to retain students.

Predictive analytics is the analysis of data to make predictions about the future. The data-gathering strategy has been around for decades, but as technology grows more sophisticated and retention becomes a bigger topic in higher education, more institutions are using predictive analytics in an attempt to improve retention and timely graduation rates.

The most recent available data show that nearly 60 percent of University students who enrolled in 2009 graduated in four years.

Director of Undergraduate Analytics Peter Radcliffe said students who are more likely to stay at the University share similar characteristics, like living on campus for their first year, getting into their program of choice and having family members who attended college.

“Virtually all students struggle at some point in some way,” he said, “and it’s helpful if you’ve had family [members] who have been through that.”

And the University can do a lot with statistics on freshmen — in 2009, University researchers published a study that measured how many times incoming freshmen swiped their U Cards at the University recreation center.

The results showed that the University is more likely to retain those who went to the rec center often.

That study became a factor in the institution’s decision to more than double the rec center’s size last year, said Tony Brown, associate director for the Department of Recreational Sports.

Predictive analytics are also a part of the University-wide APLUS system, a program that allows faculty members and advisers to see how a student is performing while they’re at the school.

Though advisers can’t see all student information, they get alerts from the program about things like if one of their students is repeating a course or if he or she hasn’t registered on time. They can also share notes with other advisers and faculty members about the same student, College of Food and Natural Sciences senior academic adviser Sheryl Bolstad said.

She said she’s been advising at the University for about 15 years and the system has simplified her job and allowed her to be more helpful to her students since it’s been in place.

“I think the students get better service because it’s easier for us to have access to that past information,” Bolstad said.

But colleges can use the analytics to benefit themselves, said Christopher Coogan, deputy director and chief of staff of the association for institutional research.

Some schools may use data to accept more students who seem likely to stay and graduate on time, he said, in an attempt to boost their rankings.

But Rachelle Hernandez, associate vice provost for enrollment management and director of admissions, said the University doesn’t use predictive analytics to keep students out.

Even though the school uses the data to make decisions about admissions, she said, staff members only use positive markers — like high test scores and high school achievement — to influence their decisions.

“We’re looking to make decisions that are good decisions both for the student and the University,” she said, “but not in a way that’s self-serving.”

Beyond colleges

Predictive analytics are not limited to colleges and universities. In fact, businesses have been using these methods for analyzing customer data for years.

Netflix, for example, uses information from its viewers to recommend television shows and movies that they would like based on content they already viewed.

Amazon and other shopping websites also use information on what a shopper has clicked on to recommend similar items.

The goal is to get a returning customer. Similarly, colleges are trying to get students to come back for their second year.

The University has a one-year retention rate of about 90 percent.

The University of Wisconsin-Madison has a similar formula for tracking graduation and retention.

Based on information about incoming freshmen, the school tries to understand what can impact students’ potential to stay at a university, said Margaret Harrigan, the school’s distinguished policy and planning analyst.

Even though her office isn’t involved with how the institution uses the data, she said, the information is important to faculty members and administrators.

“We know that they pay attention to the data,” Harrigan said, “and the goal is to have the students [have] a successful experience.”

But Coogan said the colleges that can perhaps benefit most from predictive analytics are those that have fewer resources, like community colleges.

Normandale Community College will receive a software system that could allow the school to monitor different factors that may affect their graduation and retention rates to see where the school’s programs can improve, said Normandale’s Director of Institutional Research Monica Haynes.

“We’re often asked to help those programs to make predictions,” she said, “and it’s not something that we really have the capability to do at the moment. So that would be really useful.”

The use of predictive analytics in higher education is growing as it gets easier to access the technology and record the data for it, Coogan said.

Schools could even have the ability to track students on their cellphones as they walk around campus, he said.

The University hasn’t taken that step yet, but Radcliffe said his office might consider doing so if the tracking would produce important results.

“Until we actually have a kind of explanation of what we wanted to do with [the data],” he said, “We wouldn’t just grab it for the sake of having it.”

He acknowledged that there would be ethical concerns with cellphone tracking, which the office would discuss before putting the system in place if they chose to do it.

Retention and graduation is important for students and universities, Coogan said, because it stimulates the economy, improves students’ chances in the job market and allows schools to serve their students effectively.

“In essence, your college is a business,” he said. “Getting people to graduation is the business they’re in.”