Leading With Predictive Analytics
Georgia State was one of the first universities to sign on with the Education Advisory Board in 2011, and we began the use of predictive analytics to track all students in 2012. Since the start of GPS Advising, we have had more than 250,000 one-on-one meetings between our advisers and students that were prompted by alerts generated by the system. We are tracking 800 different risk factors for more than 40,000 students every day, and we had 90,000 interventions based on the alerts over the past year alone.
We have begun to deliver common-sense advice to our students at scale. For instance, every year we have thousands of instances in which students register for courses that do not apply to their degree programs. In the past, because no one was watching, students often took the wrong course. Someone would notice after the fact, and the student would have to go back and take the right class, wasting time and money.
With our new tracking system, when a student signs up for the wrong course, an alert goes off, an adviser reaches out to the student the next day and we get the student in the right course before the first day of class. Over the last year, we had more than 2,000 of these corrections.
Intervening Early
We had more than 10,000 instances when students underperformed in a pre-requisite course.They may have received a “C” in a math course—a passing grade—but that’s not good enough to indicate they will do well in the upper-level chemistry class they need to take next. In the past, these students would have been allowed to take the chemistry class, and after earning a “D” or “F” (or more typically, two or three) an adviser would finally reach out to help.
Now the analytics alert us to the fact the student is at risk when the first grade is recorded, and an adviser reaches out immediately to help. Why wait until the student makes more costlier mistakes instead of helping at the first sign of a problem? At universities like Georgia State, where 58 percent of students qualify as low-income by federal standards, thousands of students were dropping out because they were running out of funds.
Now, we have begun to deliver the guidance students need in a timely fashion, and we’re doing it across the board for every student. It’s having a big impact, especially for low-income and first-generation students.
Seeing Improvement
Because we have been using our system more than four years, we have results based on graduation rates. Our four-year graduation rates have improved by seven percentage points since we started the program. Students are also graduating more quickly.
The average time to degree for graduating seniors has decreased by more than half a semester, saving the graduating class of 2016 saving $18 million in the costs of tuition and fees compared to the class of 2012. This means lower debt for these students at graduation. It also means more students are graduating, because more students can afford to do so.
Since we started our use of predictive analytics, we are graduating thousands of additional students every year, with the biggest gains being see with students from underserved backgrounds. For every one percent increase in retention, Georgia State has a $3.18 million increase in revenue.
Over the past seven years, the number of bachelor’s degrees conferred to African-American students by Georgia State has increased by 103 percent and Georgia State is now ranked first in the nation among non-profit colleges and universities in the number of bachelor’s degrees conferred to African Americans each year. This is the first time in U.S. history this distinction has gone to an institution that is not a historically black college or university.
We have eliminated achievement gaps. For the last four years, we have been the only national university at which black, Hispanic, first-generation and low-income students graduated at rates at or above the rate of the student body overall. Georgia State is showing, contrary to what experts have said for decades, that demographics are not destiny.
Students from all backgrounds can succeed at comparable rates. Predictive analytics have helped all demographic groups graduate at higher rates from Georgia State, but just as critically, they have helped to level the playing field for all of our students.
Tim Renick
Taking the Next Step
We have always known what works in higher education: personalized attention. This is one reason students do well at small liberal arts colleges. Someone is always watching, and students are receiving timely help when they get off path. Until now, it has been impossible for universities such as Georgia State with tens of thousands of students and modest tuition levels to deliver this kind of personalized attention. It is likely the big gains in graduation rates and the number of students completing their degrees we have made are attributable to a simple fact: someone is now watching, and we are reaching out to help when students first go off path.
We have also introduced a number of complementary programs over the past five years. In 2012, we introduced Panther Retention Grants, a micro-grant program that helps prevent students from dropping out by helping students who have trouble covering the full costs of their tuition and fees. Since 2012, we have awarded 12,000 of these grants, with an average grant of $900. Last year, more than 1,300 of our graduating seniors had been recipients of one of these grants during their time at Georgia State. In the past, many of these students would have stopped their studies or dropped out.
Sources:
GSU Office of Finance and Administration
GSU Office of Institutional Research
US News and World Report 2019 Best Colleges
Integrated Postsecondary Education Data System