Most of the more than 40,000 students enrolled at Strayer University in any given semester take classes online. They’re typically working adults who learn asynchronously, participating in lectures and assignments at different times that fit their busy schedules. Strayer is betting big on data analytics to help more students persist with their online studies throughout the years it takes to get their degree.
“Many online classes are too lonely and it’s hard to stick with them for years,” says Joe Schaefer, chief information and technology officer for Strayer Education, parent company to Washington, D.C.-based Strayer University. “We’re using data science to change the culture and feel, and the total product of being in an online class.”
For the past three years, for-profit Strayer has worked with analytics provider Civitas Learning to identify factors that contribute to student success in online learning. This week the university announced what it’s learned so far: The data captured through students’ digital footprints is a better indicator of success than other traditional standards, such as race, income or part-time status.
From Dashboard to Action
Strayer captures behavioral data on hundreds of thousands of students, Schaefer says. The university takes daily snapshots of students’ progress by tracking indicators including time spent on learning activities, use of optional resources and the relative depth of discussion board posts. Strayer uses Civitas’ “Student Insights Platform” to combine these disparate data points and create a picture of a student’s work style, communication preferences, and opportunities for support and motivation.
Faculty also use predictive analytics software from Civitas called “Inspire for Faculty,” which gives them a dashboard view of student performance data and helps them target outreach.
Schaefer says Strayer is studying how faculty keep students engaged—or not. “What are the faculty behaviors that cause positive student behaviors?” Instructors can view a dashboard of student performance data and see how their actions influence learners’ behavior. “If [faculty] send a positive message to students, they’ll see results of that in dashboard the next day, relative to all other student behavior,” Schaefer says.
Early Results
This week Strayer reported how its use of predictive analytics in an online business course improved outcomes for at-risk students. Specifically, students who were identified as at-risk and received highly targeted, personalized outreach from their instructors experienced a 5 percent increase in attendance, an 8 percent decrease in course drop rate, and a 12 percent increase in course success. Overall there was a 17 percent decrease in the students considered to be at risk.
Schaefer is quick to point out that this was just one experiment of the hundreds that Strayer is running to understand student behavior, and every case is different. “There is no silver bullet to improving outcomes,” he says. “We’re figuring out how to get 1 percent better at hundreds of things—not 17 percent better at just one.”
A Team of Care
All of Strayer's online students are paired with a coach for the duration of their studies. These coaches have a dashboard view of student performance, but it’s different from the faculty one. Whereas instructors see student data for a particular course, coaches have access to how students perform in all of their classes over time. It’s a coach’s job to help students select classes and reach out if they see changes in behavior patterns. “In the kind of cold world of online learning, we want personal connections to be there,” Schaefer says.
The situation has some parallels to shifts happening in the healthcare industry, where doctors increasingly share patient data to improve outcomes. There’s also similarity to healthcare’s emphasis on preventive care—positive behavioral shifts are more impactful in the first few terms a student is enrolled, Schaefer says.
In the future, he’d like to see more emphasis on setting students up for success, not just helping them when they start to get off track. “All of Civitas' models start measuring in the classroom. We need to be able to that same thing prior to the class starting, so we can intervene before we even start using Blackboard.”
