David Brooks becomes a rebel without a reality?
February 20, 2013
It’s rare that you’ll see me tangling here with any op-ed writers. But this odd David Brooks column that attacks data, Big Data, and data analysis is so much juicy comment bait.
Brooks dresses up these data-related ideas in straw man clothes and knocks them around for a while. And then he admits at the end that as a tool, data analysis can’t do anything it wasn’t meant for. Why not state that up front and move on? Or show us some cases where data analysis has been used inappropriately?
Brooks describes several ways in which human thought processes trump data analysis. Some are valid, but fall into the “don’t rely on a tape measure as an interior decorator” trap.
His thesis boils down to this: values are more important than data analysis. He thinks this is a good thing, obviously, but all of his examples show otherwise. Let’s run through them:
- Data obscures context (it doesn’t tell a story). Data doesn’t, but data analysis sometimes provides too much context and too many stories.
- Data creates bigger haystacks (our analytical tools are inadequate): embarrassment of riches here, not a problem. Did we stop using MRIs because they are so much more detailed than an x-ray that x-ray analysis is inadequate to interpret them?
- Big Data has problems with big problems: he cites the debate over economic stimulus and blames Big Data for people’s minds not changing in a polarized political environment. This is like blaming the smoke alarm that everyone ignored when the house burns down. Also, I don’t think he understands that Big Data isn’t really involved with the macroeconomic indicators related to the economy.
- Data favors memes over masterpieces (because it reports instant popularity): a meme or a masterpiece is a value judgement, so other than documenting what people consider something, data has no opinion. It documents the box office popularity of Avatar, while also showing that there are Big Lebowski fan cons, but no Avatar fan cons.
- Data obscures values: this is laughable. Values obscure, trump, substitute and flat out decimate data analysis in business, politics, medicine, parenting and any number of other fields on a daily basis. Here’s some glaring examples: the housing bubble and climate change.
Why does Brooks swing a wifflebat at data, Big Data and data analysis? Why does he conflate these three separate data-related concepts? Does he fear we will all become emotionless data processors without values? Has data analysis lead to some horrible tragedy that requires that it beat about the head?
No, quite the opposite. In the real world, show me a situation where someone said “this plan is crazy enough (refuted by data drawn from the real world) that it might just work” and then have it, you know, actually work.
Given the political context that Brooks operates in and his thin veneer of even-handedness, his piece seems like backdoor, pseudo-intellectual support for those who ignore evidence and empirical data because reality conflicts with their values.
I’ll go on record here as saying that ignoring reality and all its manifestations is mighty unhealthy of them, for their own sakes, as well as everyone else.