Walking the Line of Predictive Analytics in Higher Education



The ethics of predictive analytics in higher education has been a hot topic recently, and for good reason. The age of big data and advanced analytics thrusts higher education institutions into uncharted territory. Predictive analytics opens a Pandora’s box of controversial questions: What are the ethical guidelines for telling students a discouraging prediction? Would discouraging information decrease a student’s chance of completion or will it compel the student to work harder to beat the odds? Who decides what information is shared with students and when? Taken to the extreme, if predictive analytics could be self-fulfilling prophecies when shared with students, instructors, and advisors – should they be withheld?

Our Stance on Predictive Analytics

Predictive analytics are not about perfect future outcomes; they are about analyzing historical and current data to project a likely future outcome. Imagine my doctor has a test that uncovers I have a 70% chance of developing heart disease over the next few years. While there is still a 30% chance I wouldn’t develop heart disease, I’d want my doctor to use this information to provide choices and treatments to prevent this outcome from ever happening.

The power of predictive analytics in education isn’t determining a student’s future in advance. It’s helping shape positive outcomes while there is still time to act. With large class sizes and growing advisor to student ratios, identifying students in need of help can be a difficult challenge. Advisors and instructors can see current grades or whether students complete assignments on time, but this limited view does not capture the students who might be likely to struggle later in the semester even though they are doing fine now.

Rich historic data sets and technology allow predictive models to use years of student demographic and performance data to predict which students are likely to struggle and ultimately may be at risk. In essence, the model reflects how similar students with similar attributes performed in the past, in order to make predictions about current students in the future. A predictive score is not a proclamation set in stone. It is a tool to identify at-risk students early on to help shape better outcomes.

They say an ounce of prevention is worth a pound of cure. The reason I get so excited when I talk about Blackboard Predict is not the accuracy of the model, or the logic behind the algorithms. It is the ability to give instructors and advisors the opportunity to proactively help students and facilitate their success.

Driving Effective Interventions

Our goal for Blackboard Predict is to help power this impact in student’s lives and make the information actionable. James Wiley, Principal Analyst at Eduventures, says in a recent report, “Its [Blackboard Predict’s] key differentiator is how well it allows for sharing of output analytics to enable institutions to act. For institutions considering a solution that would fit within established analytics processes, Predict is a strong choice.”

Effective interventions cannot happen without student engagement. How and when we share information with students is important and nuanced. Every student is unique and will react to information differently, which makes how information about academic progress is delivered key to success. Simply telling students they are at risk is far less valuable than providing valuable guidance. Implementing predictive analytics comes with a responsibility to define a process for appropriate and timely intervention responses. The prediction itself isn’t the end game. The end game is success, driven by valuable guidance that empowers a student to act to improve their trajectory.

While predictive analytics are still quite new, I am excited about the opportunity institutions have to drive student success in a more personalized manner. To learn more about Blackboard’s predictive analytics offering, please visit our Blackboard Predict web page or sign up for a demo.