Why the Left has Criticized the New Data Journalism

Jonathan Chait has an excellent column arguing that the "data journalism" currently in vogue at sites like Vox.com, the new FiveThirtyEight, and some new site the New York Times is shortly going to open up is in fact an ideological/partisan endeavor, because commitment to empiricism as the basis for public policy is an ideological issue. It is, roughly speaking, the foundation of the original progressive movement of approximately a century ago, and it continues in the progressive movement of today. Now, it shouldn't be a partisan issue, because it seems kind of, y'know, obvious that we should base our public policy in an accurate understanding of what the world is like, but it is. This fact became blindingly obvious during the Bush Administration, with the whole "reality-based community" line and all. The modern conservative party is deeply against government use of empiricism to guide its decision-making. I might argue, analogously to how I did in this post, that this disregard for facts is another consequence of the Republican Party's commitment to an agenda which is deeply, deeply unpopular; in order to promote policies which will achieve their actual agenda without revealing what that agenda is, they need to get people to stop focusing so much on the empirical issues about the causal mechanisms. But one way or another, Chait is right: data journalism is inherently a left-wing enterprise in today's world.

This, however, leaves open a question: why has there been a substantial amount of criticism from the left, too? From, say, Paul Krugman, most famously, in his brief feud with Nate Silver over the new FiveThirtyEight. One reason is the oft-mentioned "hack gap": conservative pundits are a lot more hackish than liberal ones, and therefore one can expect them to toe the party line much better. But another, I think, is that data journalism really, really, really wants to pretend that Chait isn't right, or at least that they don't know Chait is right. The ideals of empiricism and neutrality are very deeply tied to each other in our society. If one party's ideas systemically fail all empirical testing, well, there's a limit to how much you can do about that, but you certainly, certainly can't let yourself act on the presumption that, because the Republicans have a long history of anti-empiricism, their ideas will continue to be empirically wrong going forward. So a data journalist like Nate Silver or Ezra Klein is going to have to try to frame what their doing as being completely non-partisan and non-ideological.

This, in turn, is aggravating for someone like Krugman, who occupies a position from which he can and does observe just how deeply anti-empirical the Republican Party has become. That fact is an incredibly important one about American politics, and forms a very high item on the indictment of the Republicans as being unfit to govern. It is also deeply underappreciated by the American public. To those of us within the partisan struggle, it feels remiss of anyone ever to miss an opportunity to point it out. And I think this probably explains a lot of the backlash to Vox and FiveThirtyEight. Even if all their actual analysis is completely sound, and therefore comes repeatedly to the conclusion that Republicans are wrong yet again, they feel compelled to at least pay lip service to the kind of Very Serious People-ish ideas Krugman rightly despises about how no party has a monopoly on the truth, blah blah blah. So what Krugman objects to isn't mainly the content of their analysis, although of course the whole global warming controversy at 538 was substantive. It's about the framing.

I don't say this to take a stand on whether the data journalists' framing decisions were correct; I can see the arguments both ways. If it really does enhance their credibility, and if they use that credibility to accurately portray the facts (which will reliably be detrimental to Republicans), then it would seem like a good idea. Merely an observation that I think this is what's going on, rather than (by and large) an actual substantive disagreement.