Since 2008, Timothy Renick has directed student success and enrollment efforts at Georgia State University, overseeing among the fastest-improving graduation rates in the nation and elimination of all achievement gaps based on students’ race, ethnicity or income level, thanks to an innovative technology and advisory system. The Learning House spoke with him after his speech at Connect 2018 to hear more about it.
I’d like to know more about predictive analytics and data and how you use them in college retention strategies.
Well, I think our approach to data at Georgia State, first and foremost, is to empower students. We enroll a population that is predominantly low-income, first-generation students. We found a lot of students were not graduating and dropping out, not because they didn’t have the academic ability to succeed, but because we create such insanely complicated bureaucracies and expect students to navigate them. And if you don’t have brothers and sisters and parents and aunts and uncles who have gone to college, and you haven’t been raised in a context where college issues are discussed, this is a very foreign environment.
So, about 10 years ago, we began looking at our students’ data much more proactively to see what kind of things were tripping students up. And in the case of predictive analytics, about six years ago, we looked at around 140,000 Georgia State student records. The point was, couldn’t we find in the data, academic behaviors that predicted students dropping out and flunking out? Academic behaviors that occurred not right before they drop out, but six months before, nine months before, 12 months before, while there is still time to intervene and make a difference for the students.
And what we found in the data project was 800 different identifiable behaviors that had statistically significant power in predicting students dropping out and flunking out at Georgia State. So, every night for the last six years, we have been checking the records of every single student at Georgia State for all 800 behaviors. And if anyone is identified, the next morning the advisor or someone connected to that student gets an alert that this student, over the previous 24 hours, engaged in this activity or this performance that is risky and there’s an intervention within 48 hours with that student one-on-one where the advisor will talk to the student.
Last year alone we had 54,000 one-on-one meetings with students that were prompted alerts coming into the system. While it sounds intrusive, maybe an invasion of privacy, when you actually look at the details, it’s just common-sense advice that we’re finally able to deliver to the students at scale.
One of the things we’re tracking now is registration records. Do our students sign up for the right classes? At Georgia State last year, we tracked over 3,000 instances where students went in for registration, went into the system, and registered for courses that wouldn’t advance them any closer to getting a degree. Maybe they’re a STEM major and they signed up for a chemistry course for non-STEM majors. Or they signed up for the wrong pre-requisite and so forth. In the past, students would just be allowed to take the wrong course, pay for it and – how demoralizing, not to mention costly – told six months later, “Oh back in fall of 2017, you took the wrong course and now you need to go back and take the right one.”
So, we’re catching those sorts of things.
What we found was that in many instances, we can gauge a student’s eventual success by the grades they get early in their major. What we found, for instance, with political science majors was if they get an A or B in their first political science course, they have about an 80 percent chance of graduating on time in political science, but if they get a C in their first political science and they’re a political science major, they only have a 25 percent chance of graduating on time.
Historically, Georgia State had never engaged that C student and – actually most institutions don’t engage the C students – we just passed them on to upper-level work. But you’re really misleading them because even though they have a passing grade and are eligible to take a course at the upper level, if they only have a 25 percent chance of succeeding, they’re clearly not prepared and ready.
Instead of just letting the students go on, knowing that three-quarters of them will end up getting Ds and Fs, and they’re paying for these courses that are not bringing them any closer to a degree, and many drop out – now we engage them the moment they get a C. The advisor will sit down, maybe there are writing issues or reading issues, problems with exams. Or maybe it’s just the student is working full-time. But the idea is that now we’re going to diagnose the problem when it first occurs, rather than later.
We call this system GPS Advising. It stands for graduation and progression success. But it’s an obvious play on the GPS in your car because the logic is the same. It used to be that if you were driving and made a wrong turn and didn’t notice, it could be almost impossible to get back on track again. Now with GPS, the second you make a wrong turn, you’re turning and it’s already recalibrating and you’re back on the path you need to be on in seconds. And the same with our students. In the past, they would make their mistakes. It would be a STEM major and they take a math course, get a C, they think they’re fine but really, their math skills aren’t going to get them through organic chemistry next semester. And then they get a low grade in that and then next semester it steamrolls and nobody did anything.
Now, sometimes as early as two or three weeks into the semester, perhaps as early as the first quiz, we’re saying, “Oh, well, you didn’t do that well on your first calculus quiz, let’s have an intervention, let’s talk about resources are available like supplemental instruction, tutoring, so that by the time you get to the midterm, what might have been a D grade is now a B. By the time you take the final, it might be an A- instead of a C or D.” And that prepares them to do better in organic chemistry when they’re on to that.
Tell me a little bit about the advising system. Who are the advisors? How many advisees do they have?
Well, we’ve had to hire a number of advisors. One of the questions we often get is, “How expensive is this?” And the reality is, we’ve had external people look at the data and do some ROI. Most of these initiatives pay for themselves.
Once we launched the predictive analytics and advising, we knew that there were going to be thousands of alerts going off every week across campus. If you didn’t have someone to answer the alerts, and then meet with the students in a productive fashion, you may as well not have the system in the first place.
So, we’ve gone through a process over the last six years of hiring about 60 additional full-time advisors. These are mostly recent college grads, all of them have bachelor’s degrees, many of them have master’s. But there’s no credential you need to be academic advisor, in fact a lot of the academic advisors have degrees in the areas they’re going to advise about. So, if they’re going to advise in STEM, they’re all good students who have pursued STEM areas.
But they go through rigorous training – training that used to be less strict than it is now. What the advisors are critically doing is taking data and interpreting it in the way it will be most helpful to students. How do you engage the student, based on the data?
And it’s happening at scale, we’re graduating more than 2,800 students a year than we did before the launch and we don’t have achievement gaps anymore – black, Latino, first-generation students, low-income – all graduate at or about the rate of the student body overall.
We’ve lowered students changing their majors by 32 percent. We’ve significantly lowered the time-to-degree at Georgia State not because we’ve reduced any graduation requirement but because students are finding their right major early on and they’re getting through their programs much more efficiently. This year’s graduating class — all the students— saved an aggregate $15 million in tuition and fees when compared to the graduating class six years ago.
You know, was it the students or was it us? The students didn’t suddenly get better at navigating the institution’s insane bureaucracy. We got a little better about helping the students work through those various choices and that’s made all the difference in the world.