It’s not often you get to discuss the negatives of having too much data, but Viktor Mayer-Schönberger (Oxford) and Kenneth Cukier (The Economist) do just that. In their recently published book, Learning with Big Data, they explore the dystopian extrapolation of the current movement towards data gathering in education. In some ways, theirs is a warning as old as Nineteen Eighty Four (or even Brave New World); when authorities have at their fingertips every type of data they could desire, how can we trust them to act in our best interest? (Maybe this goes all the way back to Juvenal?)
Learning outcomes are not only being tracked at rates previously unimaginable, we’re also doing our best to build correlations and causations between specific curricula and end results. We’re convinced we can optimize every student’s learning potential just by serving him / her the right knowledge in the right manner. But as Mayer-Schönberger and Cukier point out, the incentives for greater aggregate learning aren’t always so clear cut.
“Consider if the technical system made predictions that tried to improve the school’s success rate not by pushing students to excel, but by pushing them out, in order to inflate the overall grade average of students who remained.”
As much as I hate to think about it, a vast number of educational institutions aren’t truly incentivized to provide the best education possible for the largest number of people. In fact, one could argue that incentives in education are skewed imperfectly across the entire K-20 spectrum. Public school districts expend their efforts cheating on standardized tests instead of modifying their curricula to improve student learning. In higher education, an embarrassing number of people are discussing lower interest rate loans for people who choose “better” (i.e., more lucrative) majors. Post-graduate programs are ranked in part based on the ability of their degree recipients to command premium salaries.
It’s not crazy to think that, should any educational institution be able to avail itself of any metric it liked, the ability to act on disincentives – to engage in discriminatory selection or teaching practices – is dramatically increased. With public school districts, they may devote additional resources and care accelerating the best learners if they feel confident that “lesser” students will only reflect negatively on the districts’ ability to hit statewide assessments. Universities may build filtering systems or disincentives to funnel more students into majors that are deemed more financially viable and therefore more “deserving” of student loans. Graduate programs may force churn matriculated students using preliminary tests to predict those least likely to be successful. Not that I watched Divergent (Kurt Vonnegut’s “Harrison Bergeron” is a much better story), but it portrays a society where, once it has a view on how to optimize each individual’s contribution to the greater whole, entire life paths are enforced based on it.
A second complaint is just about the availability and longevity of a person’s entire academic history. We wonder frequently about how those of us belonging to the Facebook generation will ever manage to run for office with the explosion of candid, less-than-professional content that we’ve created – voluntarily and involuntarily – in our younger and more vulnerable years. The same could be said for academics and resumes. My faithful readership knows just how important the idea of credentialing is to me. What happens when you’re credentialed not just for your successes, but forever branded for your failures?
None of these are valid reasons to stop pushing the envelope on what we can track and understand about ways to learn. To be fair, Mayer-Schönberger and Cukier didn’t come to that conclusion, either. Instead, it serves as a great admonition that as we procure for ourselves the superpowers that will let us teach and learn with efficacy that would put to shame the pedagogies of the past, we ought to be hyper-attuned to the unintended consequences. Just like economic progress, education needs its own Gini coefficient and anything that increases that value should be mercilessly questioned before being pursued.