To avoid a VR hype cycle, learn from edtech
December 11, 2017
Vendors and product designers could learn a lot from the much-hyped educational technology that came before them.
Data experts at SXSWedu point to power of analytics to improve learning and intercept struggling students before they fail.
Wyatt Kash is an award-winning editor and journalist who has been following government IT trends for the past decade. He joined Scoop News Group in...
Advances in predictive data analytics are propelling colleges and universities into a new “golden age of learning,” education experts said Monday at the SXSWedu conference in Austin, Texas.
“What began as a trickle and then a stream of data,” in education “has become a river,” said Mark Milliron, chief learning officer of Civitas Learning, a provider of predictive analytics software that helps institutions assess student behavior and engagement.
“I think we’re entering the golden age of learning. The lights are being turned on … in ways we haven’t ever see before,” in understanding the factors that help students succeed — and explain why others fall away from their pursuit of a college degree, he said on a panel on how data innovation is changing higher education.
Milliron compared the torrent of data now available from learning management systems, and the insights they can offer, to the evolution of data in the health care field and the medical discoveries that arose as powerful analytic tools identified patterns otherwise lost in a sea of databases.
He also likened college faculty members to battlefield surgeons who rarely knew the eventual outcome of their work, but who now have data in hand to recognize when students are in trouble and can intervene before they fail.
“Before, we were looking at one or two sets of data; now we’re looking at a much wider picture,” said Laura Mercer, director of research, analytics and reporting for Sinclair Community College, which serves 32,000 students in southwest Ohio.
“A lot of universities may have an LMS, but they may not own that data, or don’t have data warehouses,” she said. But having those assets has made a substantial difference for faculty at Sinclair.
“Understanding that a student is in trouble before they fail a class or drop out of school is really important. We can see that as soon as two weeks into the school year. It’s a different way of doing business,” she said.
Data analytics has begun to reveal fissures in common assumptions — and provide a window onto new predictors of student success, according to Milliron.
A December 2016 study by Milliron’s company examining 3.97 million students’ records at 62 institutions, for instance, found that virtually every institution in the study was losing more students with a GPA above 2.0 than below, Milliron said. More worrisome: Almost half of students who left their college or university had GPAs at or over 3.0.
Plying through a variety of other LMS data, the study discovered four other variables — attendance, interim LMS grades, course material engagement and discussion board engagement — were all better predictors of student’s chances of success than traditional GPA averages.
Data analytics has the power not only to identify predictive variables, said Milliron, but also derived variables, such as the distance between a student’s activity measures and his or her peers.
“These variables are dramatically more predictive of persistence than anything that is reported in the LMS itself,” the report found.
“We’re finding a lot of challenges are life and logistics issues — programs and degree maps that don’t make sense that we have to fix,” Milliron said.
As data analytics shifts from being a forensic tool for accountability to more of a predictive tool to enhance learning achievement, it also has the potential to change the “culture of blame” to a “culture of wonder” around the science of learning at many institutions, he added.
Mercer added that data analytics is also helping colleges and universities get beyond handling students “like cattle calls” and instead, providing the tools to help individualize efforts to support students and ensure they’re making proper progress from the time they enter college to the time they graduate
Data analytics also has a darker side, warned Jon Daries, associate director, MIT Institutional Research Office. Colleges put a tremendous amount of effort in balancing their admissions, and so-called yield rates. That can put added pressure on admissions offices to predict which students will succeed and fail in order not to lose the institution’s standing in U.S. News & World Report’s ranking, he said.
But there’s equal incentive for colleges and universities to help students “learn well, finish strong, and help the institutions to perfect the learning journey,” said Milliron.
At the same time, he suggested, data analytics is helping institutions to move beyond an era of continually trying new best practices and focusing instead on initiatives with greater precision.