2019 Impact Awards
How Personalized Support and a Culture of Data Can Jumpstart Student Success
When graduation rates plateaued, the University of South Florida turned to predictive analytics and a case management approach to student support — with striking results.
Category: Administration
Institution: University of South Florida
Project: Persistence Committee
Project lead: Paul Dosal, vice president for student affairs & student success
Tech lineup: Civitas Learning, Ellucian, Instructure
The University of South Florida has had formal programs to improve graduation and student success rates for a decade. The institution launched a Student Success Task Force in 2009, and was able to raise six-year graduation rates from 48 percent to 68 percent in 2014. But those gains plateaued two points shy of the state of Florida's benchmark to measure excellence in student success.
So in 2014, the university was looking for a way to jumpstart its student success efforts and continue improving student completion rates. For one thing, it formed a partnership with Civitas Learning to use the company's predictive analytics platform to better understand students' needs and enable more precise, effective student support. At the same time, the university created a Persistence Committee that developed a case management approach to addressing students at risk of leaving school. The task force's early work laid the groundwork for success with predictive analytics, according to Paul Dosal, USF's vice president for student affairs and student success. He recalled that the group worked to "create the institutional culture and build the data infrastructure and research capacity to support our student success initiative. All three of those things are interrelated and laid the foundation for our implementation of predictive analytics," he said.
USF's work with Civitas involves integrating and analyzing data from USF's Banner student information system (SIS) and Canvas learning management system (LMS) to gain insight into the factors that impact student success, making it possible to identify the individual students who are at the highest risk of dropping out based on past student behavior. Dosal said getting the needed data elements fed into Civitas only took a few months. "It took us longer to figure out how to act on it," he stressed.
The Civitas platform looks at more than 300 variables. For instance, it pulls data from the Canvas LMS about how often a student is logging in. "If the average student is logging in once per day, the platform finds students logging in only once per week, he said. "That is a powerful predictor of persistence or lack of persistence."
By adopting a case management model, the Persistence Committee was able to provide students with more personalized support, taking advantage of pre-existing relationships students might have with particular faculty members or staff. (Case management has been used by several universities, including USF, but mostly for handling students who are displaying signs of emotional distress.) Teams work to triage each case and determine what is going on with a particular student: Is the issue financial, academic or emotional? Then they try to figure out how to intervene and who should intervene.
"We began to generate lists from the Civitas platform of at-risk freshman students who are not likely to persist into the next semester or year. It might have represented only 2 percent of our entire freshman class — about 80 students," Dosal explained. "I thought we could send that list to advisers, counselors and coaches and ask them to intervene. That is what got us started. We began to realize if we pulled together the right people who might know the students or know how to work with them, once we realize which issues they were struggling with, then they could develop the appropriate interventions to support that student."
Today, USF has already surpassed its goal of achieving a 90 percent first-year student retention rate and a 70 percent six-year graduation rate. USF's first-year retention rate now stands at 91 percent, the highest in university history. Its six-year graduation rate recently reached 73 percent, an increase of 22 percentage points since 2009.
Impressively, using analytics and case management tools to understand risk patterns and student engagement, USF has been able to zero out the achievement gap between students of color and white students in their first-year class. "We are thrilled that we are helping underrepresented minorities and lower income students achieve at higher rates," Dosal said, "but that is not the intent of what we are doing. We are using analytics to identify all struggling students. In the process, we are helping underperforming students, who often are minorities or limited-income students."
Dosal said there may be other sources of data that universities could use to assess persistence. "One of our powerful predictors of persistence and success is engagement with campus recreation," he noted. "Basically, the more students visit campus recreation, the more likely they are to graduate on time. It is a signal of student engagement with the university as well as health and attitude. I love that. If it could be incorporated into Civitas, great. The challenge is that the data set has to be large enough to be meaningful."
The success at USF is the result of cultural change as well as a more sophisticated use of data, Dosal emphasized. USF faculty, administrators and staff began creating new student support services in their own departments. For example, library staff members designed a "personal librarian" program so that each student would have a source of support in the library to help him or her conduct research. In another example, one of the offices that asked to be on the Persistence Committee was the cashier's office. "That represented a dramatic change in institutional culture and practice," Dosal added. "They started thinking if we have a student who owes money, that student is a retention risk. They started steering that student toward the financial aid office or the financial literacy program to find ways to help that student pay his or her bill."
Universities both large and small can take advantage of this type of analytics, Dosal believes. "It is a fairly low-cost approach. The investment has been in the platform itself and about 12 new positions, staff members who serve as case managers. Those were added over time. It was never one big chunk of money we had to put down to make this happen. I think it can be replicated at low cost."
Dosal recommended, however, that universities lay the groundwork of institutionalizing success efforts and use of data before jumping into predictive analytics: "We have had that conversation with a few universities that want to adapt our Persistence Committee technique," he said. "If the right culture isn't in place — a culture of accountability and planning for performance — it may not work. People have to lay the foundation first."
Return to Campus Technology Impact Awards Home
