Why scholars should use big and little data to study complex issues

Last week, I read an article in the Chronicle about the growth of data-related graduate programs. Big data are big, both in scope and popularity. Large datasets require data analysts with specialized skills. As a result, universities are creating more and more graduate programs to address market demand.

As colleges adjust to shrinking budgets and dwindling enrollments, data-centric programs serve as an exception to the new higher education normal. Schools are receiving an abundance of applications from energetic, highly qualified individuals. And—imagine this—graduates are getting jobs.

Like many tech nerds, I am excited about the potential applications of big data. But, these types of articles—and the underlying assumption that all data equal numbers—also cause me pause. I worry about a balance of perspectives and skillsets. When is the last time you read an article about soaring enrollments in the humanities or saw a headline like “Critical Thinkers Untangle the Most Knotty Social Issues”? Of course, some of my colleagues will take issue with the above statement. Look at design thinking, they may argue. Projects require individuals with diverse backgrounds. Project managers place a premium on critical thinking skills. Unfortunately, such movements, particularly in education, are relatively nascent and minute.

Not all data are equal. But, they are all important (and often complimentary). The Obama campaign used big data to forecast results and target resources. They also used call centers to interview undecided voters and assess their preferences. A big box store like Wal-Mart uses big data to analyze millions of customer transactions. Observational data of in-store traffic patterns provide different—and also valuable—insights. Both examples illustrate the value of multiple data sources. 

I imagine other industries and disciplines are much better at balance than education. We have a gold standard. Quantitative data, along with experimental and quasi-experimental designs, rule. Such a love affair does not always permit scholars to creatively analyze the most difficult social issues. For example, check out Caroline Hoxby’s unproductive response to Michael Bastedo’s critique of undermatching studies. 

Until we recognize the value of all data, we handicap our ability to design effective studies and develop productive reforms.