Predictive Social Science

One of the great sources of despair in the social sciences is the lack of predictive theories.

Physics can tell us what will happen when we throw a ball in the air, or when we drop two objects simultaneously. Why can’t the social sciences provide similar trajectories for human behavior?

Put another way by economist Richard R. Nelson, “If you can land a man on the moon, why can’t you solve the social problems of the ghetto?”

One argument is that the social sciences are quantitatively stunted compared to their natural science peers; that the science of social has not yet developed to it’s full potential.

Those feeling more kind may argue that human affairs are simply more complex than those of levers and pulleys; that civil society is infinitely more intricate than a Grand Unified Theory. It’s not so much an issue of scientific chops, but rather that there is so much more work to do to solve social problems.

I find both of these arguments rather uninspiring, but what’s notable is that they each lend themselves to the same solution: more data, more formalism, more math, more “science.”

As if predictive social science is just around the corner. As if the solution to poverty is one Einstein riding a wave of light away.

To be fair, the social sciences have made remarkable quantitative advances. In 2008, Nate Silver correctly predicted the presidential contest in 49 states, and the winner of all 35 U.S. Senate races.

Fueled by the promise of better sales and better customers, the field of predictive analyics is on the rise – helping companies better identify what their customers want. Or perhaps, more accurately, what they can get their customers to buy.

In 2012, for example, Target used their big data mining to figure out a teen girl was pregnant – before her father did. It wasn’t that complicated, as it turns out, just watch for the purchase of certain vitamins and you could have a lucrative customer for life.

But creeping on a teenager – or even predicting elections – is a far cry from solving our most pressing social problems.

Why can’t you solve the social problems of the ghetto?

Perhaps our first mistake is to think there is an analytical solution.

Bent Flyvbjerg, a Danish urban planner, argues that a predictive theory approach to the social sciences is “a wasteful dead-end.” Instead we should “promote social sciences that are strong where natural science is weak – that is, in reflexive analysis and deliberation about values and interests.”

Flyvbjerg calls this approach the phronetic model, explaining, “At the core of phronetic social science stands the Aristotelian maxim that social issues are best decided by means of the public sphere, not by science. Though imperfect, no better device than public deliberation following the rules of constitutional democracy has been arrived at for settling social issues, so far as human history can show.”

I’m not sure I agree with Flyvbjerg that “no predictive theories have been arrived at in social science, despite centuries of trying.” Surely, we have not solved poverty, but we’ve come disturbingly close to predicting the patterns of an individual.

But just because we could have predictive theories of social science does not mean that is all we should aim for.

There is important knowledge, valuable knowledge, in quantitative understandings of society. We should pursue those understandings fully, but we should not deign to stop there.

Why can’t you solve the social problems of the ghetto?

Surely, one white, male economist cannot. No matter how much data he has.

But perhaps we can.

Predictive social science, assuming it exists, is only one tool towards a solution. Without phronetic social science – dialogue and deliberation between all members of a society – it is worth nothing.

Of course, this phronetic social science ought to be informed by predictive social science, just as predictive social science ought to be informed by phronetic social science.

The two aren’t competing paths towards the same end – we must pursue them both.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.