January 27, 2012–Using analytics to predict future issues is a great benefit. However, we shouldn’t get carried away and think big data will solve problems associated with human unpredictability. Sure, it can guide our choices, but ultimately, humans do what they want. Case in point, the recent Daily Camera article, “Data-Mining Project to Raise Red Flags for College Dropout Risks.”
The story focuses on a Colorado community college’s attempts to help advise students what majors to choose to avoid dropping out. According to the piece: “Through the predictive analytics reporting, researchers will have access to a database of records for more than 400,000 students. They’ll also be able to comb over more than 3 million course records, digging for hints about what causes students to quit college — which could include everything from courseloads to demographics to how often they participate in online course discussions.”
This, oddly, reminded us of a story many years ago about how data mining doesn’t help spot terrorists. This was back in 2008, and we honestly feel like the intelligence community has latched onto unstructured data in an innovative way to overcome this problem. But the core of the issue is that human nature is tough to pin down. We think that goes double for education analytics of this sort. You can lead a horse to water, but you can’t stop it from choosing the wrong major.