Georgia State is the No. 2 most innovative university in the nation. The university loves to share this metric, so much in fact that it adorns the walls of the Hartsfield-Jackson International Airport and is listed as the very first bullet point on the university’s about page.
But what does “No. 2 most innovative” mean and how did Georgia State grow to such staggering stature? Predictive analytics is part of the answer to what propelled Georgia State to the top.
Georgia State began using predictive analytics in 2012 to track students and maintain their paths to academic success. On a basic level, the university now collects and assesses hundreds of potential data points per student, carefully monitoring them in order to provide relevant and helpful advisement.
The program has been hailed as a success for eliminating achievement gaps for black, Latino, first-generation and Pell-eligible students. Georgia State now graduates more black students than any nonprofit university in the nation.
But despite the attention this program has garnered in higher education circles nationwide, students have yet to see what data is actually being collected on them and how it’s being used. For the first time, The Signal is the only source with access to this information and the story behind the student perspective shared by the more than 50,000 enrolled at Georgia State – and it’s time that story is told.
Iris Palmer, an expert in predictive analytics and ethical use from New America in Washington, D.C., can answer the basics about what predictive analytics is and the four main ways it’s used.
The first use is enrollment management, which includes targeting recruitment and financial aid offers based on collected data. Adaptive technologies employ predictive analytics in the classroom by using courseware to personalize learning to each student. Predictive analytics is also used alongside facilities to determine the number of classrooms needed to fulfill demand or how to optimize water usage.
But targeted advising is the facet Georgia State is known nationwide for pioneering.
“That’s using the data and the performance of students to target different resources for them and advise them how to proceed through their college career,” Palmer said.
Georgia State worked side-by-side with the data firm EAB to develop the models that are now used in the web platform Navigate for Georgia State’s GPS Advising initiative. This system works to monitor, alert and respond to potential threats to a student’s success.
“I would even go so far as to say that Georgia State was really the one who helped EAB develop their targeted advising system,” Palmer said.
Some students are concerned about the ethics behind the expansive data being tracked. In 2016, Palmer co-wrote a report that examined and compared examples of ethical and unethical uses of predictive analytics. Her finding? Georgia State was doing it right.
But in that same report, Palmer discovered that Mount Saint Mary’s University conducted a freshman survey and used data from the survey to determine who was most likely to drop out. The catch was that MSMU then encouraged these students to drop out earlier to improve their retention rates.
Georgia State, on the other hand, only uses the data to advise and guide students toward graduation – not away from it.
One of the perils universities can slip into is basing their data too much on personal information, like race, ethnicity or economic background. By doing so, Georgia State could run the risk of perpetuating disadvantages that students may already face, since racial identity and a lack of resources already are tracked very closely.
“If your predictive analytics will only tell you who is a minority student or who is low income, that’s not a helpful predictive model,” Palmer said.
According to Georgia State’s Vice President for Student Success Tim Renick, Georgia State’s model is based entirely on academic data, such as grades, GPA, courses and majors. Student demographics and academic performance before college, like high school grades or SAT/ACT scores, are never used.
“As such, the analytics are based entirely on data that Georgia State always has collected about its students and indeed must collect about its students to be an accredited institution,” Renick said. “As a result, when we discussed the matter with our legal department, the determination was that there are no new legal issues posed by the data in GPS Advising.”
Given this, there is no way for students to stop the university from collecting data on them since this data can be as rudimentary as one receiving a “B” in Chemistry.
FIVE WAYS STUDENTS AND ADVISERS CAN USE PREDICTIVE ANALYTICS
When it comes down to it, what exactly can advisers and students do and see with predictive analytics? At Georgia State, the university uses a web platform called Navigate. Carol Cohen, assistant vice president for University Advisement, said that this program could be coming to students on a mobile platform as early as this summer semester, rolled out on the Perimeter campuses first.
Risk Scores
Georgia State’s algorithms will predict at what rate it believes a student will successfully graduate on time. This metric combines all of a student’s academic data along with 800 potential “missed markers” that would identify at-risk behavior, like being a senior and not having completed a math credit.
This risk score will be displayed as low (green), medium (yellow) or high (red) on the homepage for a student.
Academic Planner
This web feature allows students to drag and drop courses into a four-year projection, focusing on when students should take certain courses. Students can then share this online planner with their adviser to receive feedback.
This also helps the university recognize course demand, so they can know in advance whether to have more faculty to teach additional sections of courses.
Major Explorer
The major explorer feature allows students to select a major and dive into the potential and most common careers they can get with their degree. The system shows what kind of skills, level of experience and coursework certain careers typically want. Real-time data from the Bureau of Labor Statistics is updated daily to provide median wages and employment trends.
This helps students to be more strategic in their college career so they can work toward earning a job post-graduation. Advisers can use the data from other features of Navigate to determine if they should guide you towards a different career.
Adviser Communication
Students will soon be able to “instant message” advisers in the Academic Planner to ask for assistance. Instead of digging through their student email account, students can log into one system, send a message and expect a response on the same system from an adviser – business hours applied.
“Walk-ins have increased because students are busy,” Cohen said. “This will be a lifesaver for commuting students, as it will allow advisers to talk with students about things other than scheduling when they do come in for personal appointments.”
Advisers will soon know students beyond their name, GPA and major since they can note their goals and what they like and don’t like so their adviser can get to know them on a more personal level. To ensure advisers have been reaching out to students, Navigate displays all email history between students and advisers – even when students don’t respond.
Academic History
All of a students info and data is now in one place. Instead of going through multiple different screens on PAWS, students and their adviser can now see everything from their high school GPA and SAT scores to transfer credits and courses they are currently enrolled in.
Progress reports submitted by professors or teaching assistants are another feature. Other than grades, these reports track attendance, so if a student is having trouble in or withdrawing from a class, the adviser can have better insight into what went wrong.
STUDENT AWARENESS
All this being said, what do students know about predictive analytics at Georgia State? While not an exhaustive determination of student awareness, The Signal conducted a survey of 50 students, consisting of two simple “yes or no” questions.
“Do you know what predictive analytics is?”
The majority of students polled said “no,” though only by a slight margin; 58% responded “no” and 42% responded “yes.”
“Did you know Georgia State is using predictive analytics to measure your risk level of not graduating?”
This question showed a much larger divide among students; 86% of respondents said they weren’t aware of the practice, while only 14% answered “yes.”
Students were given the option to leave a comment on their submission. These open-ended responses ranged from “seems sketchy” and “seems very concerning” to “exciting and efficient” and “seems cool and very useful.”
Given the perceived benefits of the new Navigate system and the success metrics the university is advertising, it seems some students are still concerned about the process and unaware of what is offered to them.
“We have had a couple of students ask questions of the sort that you are asking … but they are comforted once they understand how the platform really works,” Renick said.
