Ethics and Predictive Analytics in Higher Ed
“Use your powers for good and not evil” It’s a famous line used in many superhero stories to remind the superhero to keep their focus on fighting the good fight.
At Rapid Insight, we love to call our customers data superheroes. It’s one of the ways we celebrate their accomplishments in using data to conquer organizational challenges. They face the same challenges as other superheroes do. They always have to be focused on working with data and using their insights in service of their students.
Frequently the work of our Higher Education Data Superheroes involves predictive modeling to better understand student recruitment, enrollment and potential student success or failure. They’re passionate about this role and it includes educating stakeholders on how to interpret the information and related insights. A customer summed it up with “we’re here to help the students. Sometimes predictive models show us a future that we want to help change. We can help change it by providing at-risk students the necessary help they need to succeed and graduate.”
These predictive models typically incorporate student data collected throughout their college experience from admissions to graduation. This can include GPA, residential distance from campus, engagement with University resources and many others. Because student data is used, there is always the overarching concern that the data is used without bias and in positive ways. We have begun to observe evolving conversations on this topic among data scientists, administrators and students.
In 2016 this began to take more formal shape when Rapid Insight was selected to contribute to the development of a framework for the ethical use of predictive analytics in higher education by New America, a non-partisan think tank,. Rapid Insight was excited to lead this conversation with other key experts in predictive analytics, learning analytics, adaptive technologies and representatives from institutions. We were proud to be the primary predictive analytics software vendor chosen and to have the most prominent representation on the 13 person panel including Rapid Insight’s Director of Strategic Partnerships and a Senior Data Analyst, as well as the Vice Chancellor for Enrollment Management at a Rapid Insight customer site. Our commitment to this process was significant. This included initially participating in interviews to develop New America’s Landscape Analysis in the spring of 2016 and conducting internal discussions to educate our team and distill a clear position for our organization. In September of 2016, we were invited to participate on the New America Advisory Council convened in Washington, DC to provide feedback on the landscape analysis and draft version of the ethical framework. The framework, entitled Predictive Analytics in Higher Education: Five Guiding Practices for Ethical Use, was released as part of SXSWEdu in Austin, TX in March of 2017.
We’ve been enthusiastic to engage in this dialogue and shape best practice on the ethical use of data. Every day we speak with colleges that are under increasing pressure to retain students for the good of the individual’s education and for the good of the institution’s bottom line. Predictive analytics is playing a vital role and is being used by colleges and universities around the country not only for enrollment management, but also to enhance student success and retention. This includes predictive analytics driving academic early-alert systems, school recommender systems and adaptive learning technologies. One key to the ethical use of data is a point that Rapid Insight has long held near and dear and is reflected in the operation of our software. That is, the process of building and sharing predictive models must be transparent and not in a “black box”. If you are asked what variables are used or how the model was calculated, do you know? Can you defend the results of your model and how it has driven policy decisions? Transparency is the only way.
Upcoming Online EMchat – April 27th, 9 PM ET
We want to see the dialogue around the ethical use of data continue- it’s important. We have a shared interests as software vendors, higher education institutions and students. This is why we are excited to share that we are hosting an exciting Twitter event very soon and we hope you’ll join us. We have teamed up with the folks at EMchat to host an online panel discussion with three higher education leaders titled, “EMchat Live: The Ethical Use of Predictive Analytics in Higher Education”. The session will take place on April 27th at 9 PM ET and is open to all. If you can’t attend live, we’ll also have a recording available for anyone who wants to watch an on-demand version. You can learn more and register at this link.