Case Study: Crown College Uses Predictive Analytics to Retain At-Risk Students

By on

Crown College in St. Bonifacius MN embraced a five-year persistence and completion project four years ago. The college’s Presidents’ Cabinet embraced persistence and completion as an improvement initiative to retain at-risk students and the school had been accepted into the Higher Learning Commission (HLC) Persistence and Completion Academy. HLC accredits colleges and universities in a nineteen-state region.

Critical to the initiative was to use six years of data from first-time, full-time degree-seeking students who entered the college in fall semesters of 2009 – 2014. It mined this data and identified nine factors that contributed toward making a student at-risk at the school, and it used this information to create a logistic regression model to predict the retention outcome of a particular student population. It ran this model in the fall of 2015 for the first time against its incoming full-time freshmen in its School of Arts and Sciences (SAS), and continued to use the model for the next six semesters.


Learn new analytics and machine learning skills you can put into immediate action with our online training program.

Starting in the summer of 2018, the school analyzed its data and determined that the model still was accurately predicting students that needed assistance to persist at the college, said Patty Pitts, Institutional Director of Persistence and Completion at Crown College.

“The Persistence and Completion team strategically developed intervention strategies to assist these students to thrive on our campus and retention has improved,” Pitts said. In the fall of 2018, its improvement initiative was extended to its School of Online Studies and Graduate Studies (SOS/GS).

In spring 2015, she said, the school retained 90 percent of all eligible students, and 84 percent of the freshmen who were eligible to return. Spring 2019 saw it retain 94 percent of all eligible students, and 89 percent of the freshmen who were eligible to return, Pitts said. The program continues on to this day successfully.

Predictive Analytics Steps In

A major part of the initiative was to perform an extensive data discovery effort in partnership with Dr. Gesele Durham, HLC Persistence and Completion Academy mentor, and Product Consultant Jason Ackman at Jenzabar, which provides predictive analytics software and services designed to help higher-education schools reach their goals.

Taking a proactive stance in dealing with possible at-risk students, as Crown College did, is more important than ever. “The college needs to organize for student success by looking at the bigger picture,” said Meghan Turjanica, Jenzabar Student Success Product Manager. An element of this organization is for higher-ed institutions to start thinking about the fact that they are businesses, too, and just as with business customers, it’s less costly to keep a student in school than it is to recruit a new one.

In many cases, colleges and universities tend to have the same probational criteria practices as they’ve had for 25 years or longer. “These may not be useful anymore for understanding what students will be successful or not,” she said.

Eric Nilsson, Jenzabar Vice President of Corporate Development, expanded on that point, explaining that there’s no time like the present to change how colleges think of themselves—and how they act on new ways of thinking. “It is a big industry challenge because historically colleges and universities didn’t have culture of using data-driven decision making,” he said.

“Colleges’ and universities’ overall structure has been around longer than most countries. So, schools are almost looking backwards about how they did things before. They use what they think are best practices based on history but are not looking at current data.”

At Crown College, purchasing the Jenzabar Retention product happened at the same time as its enrollment in the Persistence and Completion Academy and the creation of the persistence and completion project, Pitts said:

“The retention product was an important piece of the project that helped launch us and allowed us to experience some gains in our retention. We also formed the Persistence and Completion Committee that has been instrumental in developing strategies and activities to work with the students identified as at ‘risk’ in the retention product. Using the Jenzabar Predictive Analytics allowed us to focus our efforts to improve persistence across the campus at a higher rate than we had prior to the purchase of the product.”

Colleges and universities such as Crown College that have turned to predictive analytics for student success modeling, are ahead of the game. Others have some catching-up to do, and the sooner the better. One of the problems Jenzabar has seen is that by the time they are working with schools about currently enrolled at-risk students, it’s often too late to affect their decisions about leaving, said Turjanica.

“Mid-term is too late, so what our most successful institutions are doing is to take the information of who is at-risk before the term begins and design intervention plans for those students,” she said.

Jenzabar gives predictive scores of student who were retained or not, and works with schools to build out engagement plans from there—assigning at-risk students to a particular advisor, for example. 

Crown College has two individuals on staff who have skills and knowledge in the areas of statistics, testing, and measurement, Pitts said.

“The value of using internal staff is that they are able to do the work and understand the context and can draw meaning from the analysis that an outsider could not do without significant time and effort.”

Building Buy-in

Crown profited by having buy-in from senior leadership from inception, which made it easier to then get buy-in from administration, faculty, and staff. That minimized challenges, though of course, some remained.

“Our biggest challenges were effectively communicating, increasing the awareness on campus about the importance of improving the persistence and completion rates at Crown, and engaging every staff and faculty member across the campus,” she said.  

Others have had it tougher when it comes to getting more use out of their data. It goes back to the issue that schools don’t necessarily like to think of themselves in terms of being a business, yet it is important to their finances. Nilsson points to a session he attended at a Council of Independent Colleges conference, where one attendee—a college president—had a hard time getting the board and stakeholders to agree that they actual had a business model. “It’s not necessarily accepted to have a business language in higher ed, but they do all have business models like those to create tenure and so on,” Nilsson said.

In addition to its solution for predictive analytics for student success, Jenzabar also focuses on institutional finances for everything from managing cash flow to tracking procurement and streamlining asset depreciation.

Next Steps for Jenzabar and Crown

Looking ahead, Jenzabar aims to implement more best practices to data models for its customers, applying the power of the cloud to them. Nilsson said Jenzabar can help a school with predictive analytics by using multiple third-party data sources along with the school’s own data. “We have a data lake to absorb third-party data they can use without mashing it up with their own data,” which helps with providing additional insight.  

Crown currently is discussing the development of a model for sophomores that may be at risk, but no decision has been made yet. It also has been in communication with the company about the noncognitive assessments it can purchase for use with its students, though a final decision on that still hasn’t been made either. “We are continuing to make retention gains with the products we are currently using and will continue to analyze data before we make any decisions about additional solutions,” Pitts said.

Crown is pleased with what they have in place now, noting that Jenzabar’s technology has been instrumental in retention gains. Pitts pointed out that the college only has been using the first alert solution in the SOS/GS Division since the fall 2018.

“We are excited to see the impact the implementation will have on the gains we will see in our SOS/GS program,” she said.

Image used under license from

Leave a Reply

We use technologies such as cookies to understand how you use our site and to provide a better user experience. This includes personalizing content, using analytics and improving site operations. We may share your information about your use of our site with third parties in accordance with our Privacy Policy. You can change your cookie settings as described here at any time, but parts of our site may not function correctly without them. By continuing to use our site, you agree that we can save cookies on your device, unless you have disabled cookies.
I Accept