Report Proposes Ethical Practices for Use of Predictive Analytics in Higher Ed

• By Dian Schaffhauser
• 05/17/17

The productive use of predictive analytics in higher education is almost a foregone conclusion. Being able to predict whether a student will enroll in your institution, stay on track in his or her studies or need extra support to succeed seems like just the kind of data that can help colleges and universities meet their enrollment goals, better target recruiting efforts and more strategically apply their institutional help.

However, the application of data in this way also cries out for a set of ethical practices to prevent its abuse. For example, the same data that can help students succeed could also be used to pinpoint which low-income students not to bother recruiting because their chances of enrollment are smaller than more affluent candidates.

Last year New America set out to explore the use of predictive analytics. Authors Manuela Ekowo and Iris Palmer interviewed 30-plus college administrators, experts and educational technology vendors and also worked with several institutions to write "The Promise and Peril of Predictive Analytics in Higher Education: A Landscape Analysis." This year the policy analysts developed a framework of ethical practices for the use of predictive analytics. Their newest report, "Predictive Analytics in Higher Education: Five Guiding Practices for Ethical Use," examines the ethical concerns involved in using data to make predictions and its impact on underrepresented students. New America is a thinktank that undertakes projects bridging technology and policy.

The report's five guiding principles build on each other as the institution grows and matures in its use of data:

• Have a vision and a plan. Bring together "key staff" to make the important decisions and examine the "purposes, unintended consequences and outcomes" when developing the plan.
• Build a supportive infrastructure. Communicate the benefits of using predictive analytics and create a climate where it can be embraced. That includes setting up "robust" change management practices.
• Ensure the proper use of data. This encompasses making sure the data is complete and of sufficient quality to answer the questions being asked, that data privacy is guaranteed and that data security is monitored.
• Avoid bias in models and algorithms. That calls for testing, being open about the models and choosing vendors "wisely."
• Intervene with care. That requires communicating to staff and students about the changes in intervention practices, acknowledging that predictive-driven interventions can do harm if not used with care, training staff on the limits of data, training students on the use of their own data and evaluating and testing interventions.

"Predictive analytics are already changing how institutions recruit and support students," the report concluded. "As use of these tools become second-nature, addressing their ethical use will become even more important."

"Predictive Analytics in Higher Education: Five Guiding Practices for Ethical Use" is available on the New America website here.