Supporting Student Success with Predictive Analytics
Colleges and universities are under intense pressure to maintain enrollment, increase student retention and ensure student success. Predictive analytics can play a crucial role in these efforts at institutions of any size, by providing actionable data that can drive more effective strategic decision-making. In this web seminar, originally broadcast on May 5, 2015, an executive from The University of North Carolina at Greensboro described how the institution is using predictive analytics to increase student retention, and how it is positively impacting strategic decision-making in both enrollment and student success efforts.
Senior Statistical Analyst
Rapid Insight Inc.
Rapid Insight was founded over a decade ago with a focus on making it easier for anyone to use data to make better decisions. We have customers in many different markets. Specifically in higher education, our customers include two- and four-year institutions, as well as graduate schools and law schools.
Our software is being used by over 200 institutions nationwide. Our higher education customers are using our tools to cover the entire student relationship life-cycle—from recruiting and enrollment, to helping identify at-risk students and ensuring they achieve success. Then, after completion, our solutions help identify alumni with the most potential for becoming donors.
Our toolsets were designed to help you integrate data from any sources, do the necessary cleanup, automate mining and predictive modeling, and then output those results into any format needed—and all of this without needing to be a database expert or to have a degree in statistics.
Associate Provost for Enrollment Management
The University of North Carolina at Greensboro
Data has been my friend throughout my career. We are truly developing a data-driven culture here at UNCG, and predictive modeling has certainly helped with that. At the beginning there were a lot of questions but very few answers about why our students were leaving. Many folks had theories about why things were going wrong.
First, we realized that we needed a culture change. And one way to change that was to show the data and make sure that the message was clear. But strategic enrollment management is not about just running numbers; it’s about running numbers and then doing analysis. Once we started doing that analysis, we got together with our staffs. When I first came to UNCG, we moved away from the traditional enrollment management perspective where it’s just admissions, financial aid and registration, and we started focusing on the idea that student success is everybody’s job. The university is fully integrated—everybody is involved, from recruiting all the way through to our alumni.
How have I used Rapid Insight? Well, we talk about how many people enroll, who will enroll, who we’ll retain, and we have a future plan. We’re starting to predict graduation through a four-year predictive model.
One of the greatest things about becoming a data-driven culture is it allows us to take advantage of all the benefits of predictive modeling. I started using predictive analytics back at Seton Hall in 2006. At the time, we brought in our largest class there, plus we maintained a high retention rate, and a lot of that had to do with the fact that we were using data to make decisions. Most of what we were doing was paying attention to our territory analysis. We were looking at what yields we were getting from different territories; the higher the yield we were able to see, the more people we could target. I then got into predictive modeling for financial aid because our population was very price-sensitive, as it is here at UNCG.
We’re able to use Rapid Insight and analytics to home in on financial aid awards without overspending. We’re able to meet our revenue goals and our target goals, and that’s important for private institutions. To date, we’ve received almost 1,000 more in-state first-year applications than we did last year. And our first-year admits have increased by over 650 students. Our confirms are up by about 9 percent. So we’re feeling fairly confident that we will have an even larger class than we had last year, which was the largest class in UNCG’s history. Our overall enrollment has increased by 6.2 percent. UNCG has some wonderful programs, and we have a lot of students who apply from out of state, but it’s very tough to get that population to enroll because they have so many different options. As we started looking at this and started looking at our competitors, we said, “We need to be careful here. We are doing a lot of out-of-state traveling, but what kind of yield are we actually getting for it?” This is where you have to be careful. You should use predictive modeling only as a tool, not as gospel.
The best way I’ve heard it put is this: Predictive modeling is not destiny. So we didn’t look at this and say, “Let’s stop out-of-state travel.” Instead, we used this to assist us in our decision-making, and we started looking at just those states we were close to. From a financial aid perspective, we also used Rapid Insight for FAFSA location. Our data indicates that students who list us first on the FAFSA are more likely to enroll than those who did not. So we had our financial aid professionals contact that group first, because we wanted to make sure they will have a very smooth process. Financial aid can be very busy at times, so when the financial aid office wants to talk to a student, they want to talk to a student who already knows they want to be here. For the other population—i.e., those for whom we were not the first choice—we would sometimes use admissions counselors or faculty, depending on some of the other data we gathered with Rapid Insight. A lot of people also ask me the question: “How long did it take you to do this?” Well, once you start doing it, it’s easy. But you do need to take some time learning how to do it and understanding what analytics you can get. How are we using analytics now? One example is with our faculty outreach. We show them this information. Instead of sending each professor a file of 2,600 students and saying, “Let us know how these students are doing,” we can say, “Based on our data analysis, this decile of students may be having trouble in your classes. Can you let us know if there is something we can do for them?”
Our Students First office, which houses our professional advisors, then does the legwork to make sure that the students get the tutoring they need. It is a very focused agenda. Our campus loves it because they don’t have to call all the students, but rather only those we were identifying for them. We’re very excited about where we are, and where we’ve gone, and where we’re going. But what I want you all to know is that you can use predictive analytics to suggest where you need to be. We’ve taken a proactive approach with a lot of these students and reached out during critical times. Using predictive analytics, we’ve been able to identify students who are at-risk and spend a few more minutes with them to make sure they understand the rules and regulations, how to register and what they should be registered for. That is far superior to trying to spend five minutes with every student and hoping we were able to make the rounds. Now we can be targeted in our messaging.