You have /5 articles left.
Sign up for a free account or log in.
Colleges see plenty of potential in predictive analytics -- using data to make informed decisions about institutional strategy and help students progress through their courses and toward graduation.
But there is also potential for abuse, namely that data could instead be used to limit students’ opportunities.
New America hopes to avoid that. The Washington-based think tank on Sunday released a five-point framework intended to help colleges navigate the ethical concerns surrounding predictive analytics as they flip the switch on adaptive learning, early-warning and other data-driven systems.
“Using data ethically is complex, and no magic formula exists,” the report reads. “This ethical framework is meant to start conversations on campus. It cannot address all possible issues surrounding the use -- and potential abuse -- of institutional data.”
Some colleges already have systems in place that notify advisers when students are in danger of dropping or failing a course, for example, or recommend majors to students based on their interests and academic performance. Many more colleges are not that far along, however. The framework therefore targets those institutions that now are considering predictive analytics.
“We’re reaching a point where the idea of predictive analytics is really catching on,” said Iris Palmer, a senior policy analyst at New America. “As institutions are pressured to improve completion and graduation rates and provide more support to students, it’s going to become more common.”
The framework introduces questions that colleges should debate as they proceed through the planning, design and implementation phases of data-driven systems. It offers advice on how to involve faculty members, staffers and students, how to keep data private and secure, and how to use that data to meet institutional goals.
“Have we set a goal/vision for using predictive analytics and/or adaptive technologies?” an early question reads. Follow-up questions encourage colleges to view their plans for predictive analytics from different perspectives. “Have we considered the unintended consequences predictive analytics may introduce? When drafting our vision and goals, have we made any assumptions about students or data?”
New America created the framework with help from an advisory council consisting of ed-tech vendors, higher education analysts and researchers. Broadly speaking, the advisory council endorsed a vision of responsible use of predictive analytics where technology plays a supporting role to advisers and faculty members who know how to interpret the data and explain it to students. Colleges should be open about how they collect data about students, and mindful of whether the predictive models they have built address existing biases about age, gender, race and socioeconomic status.
“As we predict the future, we want to make sure that we’re not just repeating the past,” said Manuela Ekowo, a policy analyst at New America.
The framework joins a growing body of best practices concerning the ethical use of data. In the summers of 2014 and 2016, for example, administrators, ed-tech vendors, faculty members and higher education associations met to discuss the topic during the Asilomar conventions.
During both meetings, attendees produced a set of guidelines that, like the New America framework, stressed using data in transparent ways to improve student outcomes (Ekowo and Palmer said they were familiar with the work and had spoken to the organizers).
Mitchell L. Stevens, associate professor of education at Stanford University, in an email called the framework a “very useful document.” Stanford, along with the research and consulting group Ithaka S+R, helped organize the Asilomar meetings.
“We are in early days of a wholesale transformation of postsecondary delivery, in which distinctions between physical and digital platforms are dissolving,” Stevens said. “Ubiquitous digital traces of learning activity raise very large new questions about the nature of the student record and instructional propriety. We need a hundred reports like this one -- and collegial conversations among leaders of colleges, education businesses, engineers and ethicists to think them through.”
Martin Kurzweil, director of educational transformation at Ithaka S+R, said the report’s straightforward instructions will help colleges make sense of predictive analytics. “I expect this paper to serve as a valuable resource and a good entry point for institutions pursuing efforts to use data to better serve their students,” he said.
