Michael Sajor is Chief Information Officer of the Apollo Education Group
How Data Analytics Can Enhance Learning in STEM Disciplines
When I hear leaders speak of Big Data and the power of collecting massive amounts of data relative to student behaviors, I grow concerned. Perhaps my fellow IT professionals have the same reaction, especially when public- and private-sector leaders seem to speak of Big Data as if they’re speaking of a mythical King Kong-type creature. Even if leaders’ feelings toward the idea of Big Data are similar to the way they see a nuclear reactor or a hydroelectric dam, I still worry about the direction the conversation is headed in. Information Technology professionals know that data alone is useless without generating actionable insights into problem solving, and thence using those insights in a positive way to effect significant and lasting change – improved outcomes.
I lead a talented group of IT professionals who often hear me say that data for data’s sake is not helpful. We speak of outcomes, not Big Data, and we work with predictive data analytics and interventions in the field of higher education. We have heavily and thoroughly instrumented our online classroom—also known as our Learning Management System—and we currently collect over 500GB of data each day that we use to evaluate student behaviors and take actions to help them succeed academically. The data is processed and measured continually, and balanced against past performance of the student, as well as performance of the particular cohort they are in, and against the performance of all cohorts over time. Variances are identified, tracked, observed – and when they become statistically significant, reported to academic counselors or faculty members. The relative magnitude of the deviation is reported – signifying if the analytic model concludes that the student is at some minor risk of not meeting learning objectives, or if the student has some serious deviations in their behaviors that would indicate a strong likelihood of unsatisfactory outcomes without immediate academic intervention. And then, the intervention models are defined – based on the insights coming from the data – so counselors and faculty know how to address the issues that were uncovered in the most satisfying way that is most likely to result in positive change in outcome.
As IT professionals, we not only orchestrate data infrastructure and instrument applications to collect data, we have defined our motivations and desired outcomes, as well. We work closely with our University partners. We aim to improve the learning outcomes of working adult students and do everything with an eye toward helping University of Phoenix earn recognition as the most trusted provider of career-relevant higher education for working adults.
I also hire talented working adults who completed their degrees whose learning was enhanced by what predictive data analytics provided their faculty and academic counselors. I want to hire talented graduates, like those who come from the schools and colleges within University of Phoenix, so that our work at Apollo Education Group can help advance learning at institutions.
I believe all institutions of higher education must begin using massively scaled data analytics to improve student learning outcomes. Our teams at University of Phoenix have joined with others at The Bill & Melinda Gates Foundation, Eduventures, IBM and other universities, to share best practices and distribute case studies. Companies like Civitas are engaged in helping educators and institutions instrument, measure and take action on the right data. Collectively, we are on the right track – we just need to move faster to leverage the incredible resources that are available to us.
All IT professionals working to advance a larger talent pool must focus on the specifics of educating a STEM-ready workforce. Toward this end, by using predictive analytics we can more accurately identify how students are really performing, when they need support and additional learning resources. We can also see what is working and where areas for improvement exist within our physical and online classrooms. But data analytics only works as part of a comprehensive approach that has an emphasis on strong human academic counseling and — most importantly — engaged and experienced faculty members. Faculty are more important than any data, any software, and any application — that’s always true.