Yesterday’s announcement of Blackboard Predict might sound a bit familiar. That’s because in December 2015, Blackboard acquired an innovative startup called Blue Canary, and it is upon this basic technology that Blackboard Predict was built. But Blackboard Predict isn’t just a new name applied to an old product. Rather, since Blue Canary was acquired, our team of data scientists and engineers have worked diligently to take that fundamental technology, and on the basis of extensive user feedback, designed a product that we believe will be a game changer for students, faculty, and advisors. All this has taken place under the guidance of Mike Sharkey, now Vice President for Analytics at Blackboard, whose team was responsible for developing the technology. Mike has also ensured that our continued innovation around predictive analytics is in line with the original vision: using data to promote student success.

So, what is Blackboard Predict?

In a nut shell, Blackboard Predict consists of: (1) a custom predictive model that makes use of an institution’s existing data to identify students at risk; (2) engaging visualizations that help to drive action on the part of students, faculty, and advisors; and (3) a communication tool that enables faculty to easily pass early alerts along to professional academic counsellors in support of proactive and intensive advising practices.

How does it work?

The implementation process for Blackboard Predict sounds complicated, but it’s actually pretty simple. In fact, it can be up and running in just three months. The first and arguably most challenging step is to understand the problem. There is no such thing as an at-risk student. There are only students at risk of engaging in specific types of behavior or achieving particular outcomes. This is a crucial distinction that has major implications for how we treat and engage with our learners. The language of at-risk students assumes that there is something about the students themselves that makes them less likely to succeed. If students are inherently more or less at risk, then there is nothing much that we are able to do about it. The minute we shift our perspective away from students and toward behaviors, however, we begin to look at things that we can change. Examples of questions that Blackboard Predict has been used to address include: What are the chances that a student will pass a course with a grade of C or higher? What are the chances that a student will attend class next week?

The second step is ingestion. Once we have worked with an institution to clearly identify the kind of behavior they want to predict, we access, combine, and clean data from multiple sources, including the student information system and the learning management system. It is this aggregated data set that forms the raw material for our data scientists. It is important to note here that this process (as well as Blackboard Predict in general) is platform agnostic. Whatever your SIS, whatever your LMS, we can work with your data as a partner in support of your students.

The third step is modeling (there’s a reason why data science is called the ‘sexiest job of the 21st century’).  I’d like to say that ‘this is where the magic happens,’ but there’s nothing magical about this process. There is no black box. There is no secret sauce. There is only high quality data science. In fact, when we have generated the predictive model, we freely share it with the institution. It’s your data. They are your students.  It’s your model.

The final step is distribution. What good is it if you can predict that a student is likely to fail a course but nobody takes action? The distribution part, putting predictive information in the hands of people who can do something about it and in a way that facilitates appropriate action, is absolutely vital. That’s why Blackboard Predict features dashboards for instructors, advisors, and students. In each of these visualizations, careful thought and attention has been given to the types of information that each type of user needs in order to take action. We have further integrated communication between instructors and advisors so that educators and student success advocates can more effectively collaborate in support of student retention and graduation.

Why should anyone care?

There are an increasing number of student analytics and retention products on the market. This is a good thing. Proactive and intensive advisement works, and we know that something as simple as a well-timed email from a faculty member can make a significant difference in the life of a struggling student. Too often, however, analytics used by integrated planning and advising systems produce alerts only once a student has already gone off path for graduation. The student has earned a low grade. They have registered in the wrong course. These kinds of systems have been a godsend for advisors, who can now proactively reach out to students who have made suboptimal decisions, and get them back on track. But with access to data and a recognition of the power that this information has to increase the effectiveness of student services, has come a desire for more. What if an advisor could reach out and intervene with a student before they went off path in the first place? What if they could receive, not just information about grades in progress, but also risk probabilities based on other factors, including in-class engagement? By putting the results of high-quality data science into the hands of expert practitioners in a way that promotes timely action, Blackboard Predict is a truly innovative complement to proven 21st-century approaches to student success.

Related Posts

Share This Article

Twitter Facebook LinkedIn Pinterest Email