Part I: How Predictive Analytics is Transforming Higher Education EnrollmentReading time: 4 minutes
Most institutions have access to a wealth of data that can help meet enrollment goals, boost retention rates, and drive more effective decision making. However, this data often resides in silos around campus, making it very challenging for institutions to glean insights or develop action plans.
This is the first post in a two-part series that will explore the benefits of using predictive models on enrollment and retention, and provide specific examples of how universities are accomplishing this work. Let’s dive into this post focused on enrollment.
Harnessing Predictive Analytics to Improve Enrollment
Enrollment accuracy is essential to the survival of any university. Under-enrollment can lead to unfavorable financial complications and over-enrollment can lead to overcrowded housing and a shortage of available classes. As such, most universities are under intense pressure to make accurate and informed enrollment decisions related to their college data.
This is where predictive analytics comes into play. Predictive models can help institutions meet their enrollment goals by answering the question: “What is the likelihood of a specific applicant enrolling at the university?” The answer can help institutions optimize their admissions outreach, focus their efforts on the right students, and even shape the characteristics of the incoming class.
Predictive models can also help forecast headcount and make prudent decisions on financial aid. This model answers the question: “What is the incoming class size and how will financial aid need to be allocated to support it?” The outcome can inform financial decision-making to help the university reach its revenue goals.
Predictive analytics can even be used to better understand competitive programs. “Which applicants will choose another university and where will they go?” By running enrollment data through a predictive model, universities can learn how they stack up against competitors and develop more effective programs for converting applicants into enrolled students.
As you can see, there are plenty of ways to use the outputs of a predictive model to improve enrollment. By summing the probabilities, an institution can better predict its incoming class size. By weighting financial aid with enrollment probabilities, an institution can predict expected financial aid outlay. By scoring the waitlist, institutions can decide which students would be most likely to accept an admission offer—and have a better idea of how many students to admit in order to fill necessary seats. Aggregating these results can even provide institutions with insight into specific sub-groups of students and how to best meet their needs.
Kenneth from Kutztown, PA
“I think Rapid Insight has one of the most responsive support teams I have ever worked with, be it via email, Twitter, or on the phone. On top of that, they really know their stuff and are always willing to help no matter how small or large the problem is.”
Real World Predictive Analytics for Enrollment
Let’s take a look at two universities that were able to leverage enrollment data to predict outcomes and make more informed data-driven decisions.
Fairfield University took a unique and truly creative approach to developing an enrollment predictive model. By incorporating high school and admission statistics, Fairfield was able to develop an “IR Score”, an original and all-encompassing enrollment metric. The IR Score assigns a single number to each student that is a measure of their predicted future performance.
The IR Score is broken down into high school GPA, SAT Math and Verbal Comprehension Equivalence, and an admission rating (based on interviews and college essays). The IR Score can also include any data that was used to investigate the student’s prior performance. Alternative information would include metrics such as class percentile, AP data, and entrance exams.
If a student has an IR Score on the low end of the predictive model, the university may decide to track the student’s GPA and provide support as needed once enrolled. If a student has an IR Score on the high end of the predictive model, the university may predict that she is likely to enroll at another institution. Regardless of where a student falls on the spectrum, the IR Score enables the university to know and understand the student’s capabilities and to execute the most effective means of reaching him.
The IR score may have a complex background of information, but its simplicity as an indicator for a student’s potential makes it a key player in assisting decision makers. In fact, this metric has helped declutter the report system at Fairfield. By assigning every student a single score, reports are easier to read and faster to comprehend by non-data-oriented staff outside of the Institutional Research department so everyone can make more informed enrollment decisions.
Let’s quickly explore one more example.
The University of Pittsburgh was able to synthesize historical data to predict future behavior by building a predictive model. The question they were trying to answer: “Will a prospective student pay their deposit?” Michael Seidel, a Senior Data Analyst in the Office of Admissions and Financial Aid, built weekly reports that compared students who had deposited in the past with similar students. One data pool consisted of applicants, while the other consisted of current students. The institution was excited to see that they could successfully predict results for the upcoming fall term applicants based on this model.
The examples I shared with you today are all Rapid Insight customers who are leveraging our analytics platform to meet enrollment goals. Visit our higher education solutions to learn more about the platform, access video resources, and hear from other higher education customers who are meeting their enrollment goals with Rapid Insight.
If you’re inspired by the results and ready to learn more, please request a demo with one of our expert analysts.