Nate Silver's Blind Allegiance to Data

Statistical Whiz-Kid Nate Silver made waves last week with the launch of his new FiveThirtyEight website and his proclamation that opinion columnists are useless (The Week). As quoted in New York Magazine, he argued, “They don't permit a lot of complexity in their thinking. They pull threads together from very weak evidence and draw grand conclusions based on them... It's people who have very strong ideological priors, is the fancy way to put it, that are governing their thinking. They're not really evaluating the data as it comes in, not doing a lot of [original] thinking... We're not sociopaths, which means that we look at the world and have opinions. But we're not trying to do advocacy here. We're trying to just do analysis. [New York]”

To many this argument seems completely rationale and hard to find fault with. It is the classic positivist position that has been popular in a certain strand of academia since at least the rise in popularity of Karl Popper on our shores in the 50s. It claims that the only thing we can truly know is what we can discern from sense data. It discounts subjectivity and opinions in deference to what it calls “objective,” empirical research. It discounts Marxism, psychoanalysis, critical theory and any other theory that is not falsifiable (aka able to be “proven wrong.”) And it has benefited from dramatic improvements in econometrics and statistical analysis in general, allowing what is purported to be more reliable and valid research. The press has largely followed this charge in recent years, engaging in a “he said, she said” style of journalism that many critique – simply reiterating what pundits and politicians say without any fact-checking, or even reasonable argument checking.

So does Silver have a point? One could argue that much opinion journalism on the right and left (though the left in the mainstream media is really centrists liberal) does merely reinforce the orthodox ideology of their compatriots. In fact, the talking heads on the right and not that different than many talking and writing heads on the left, suffering in many cases from what statisticians call “confirmation bias” (the tendency to “see” what you are looking for, and ignore what you are not – or as documentarian Errol Morris put it, “believing is seeing”). So, yes, we could argue that we can predict with relative consistency what our most famous pundits will say or write about any given issue. And that is a problem. Yet does that mean we should embrace an objective, data-driven form of journalism instead? This is, in fact, what Bloomberg claims to offer – though many are less than impressed with the results.

I think here we fall perilously close to undermining the very point of the media in a democracy. Edmund Burke arguably coined the term  “The Fourth Estate” in a debate in parliament in 1787. The term has been used ever since to describe a media that challenges government and holds it accountable to the people and the truth. Throughout history, media has taken down politicians and governments (most famously Nixon’s here, but others across the globe), uncovered corruption and provided a platform for social justice and expanded democracy in every corner of the planet. But something changed starting in the 80s here and has slowly but steadily undermined its ability to serve this essential role. This change has, not coincidently, followed the incredible consolidation of power into five (or maybe six, if you count Vivendi) multinational corporations that control over 90 percent of the media Americans consume on a daily basis (following deregulation from first Reagan and then Clinton).

This is not to argue that facts and statistics should be lost in the debate. But opinion journalists are just that – media personalities that can look deeper into the facts and opinions and provide perspectives for the people to consider and debate. If an opinion journalist never shifts their position on anything, they do largely become worthless and if they never include strong evidence of their positions, they are simply storytellers (as, one could argue, David Brooks should be considered). Yet if we become so encumbered to facts that we ignore what those facts mean, then what purpose do the facts really provide. The problems with the positivist position have always been clear, and I’ll conclude with the questions we should all ask when statisticians like Silver come along to tell us objectivity exists and should be our only goal …

Do we really want to live in a world where we only hear about what is, not what can or should be?

What happens to our consideration of things like ideology or even cultural racism, which are all but impossible to measure?

Should we heed the call of Benjamin Disraeli that there are lies, damn lies and statistics – i.e., statistics can lie as well as Fox News pundits?

Who gets to decide what questions are important and whose voices we will hear? What happens to those voices not heard? Isn’t this already a big problem across the media and journalistic worlds?

Is there even such a thing as objectivity?

Nick Silver should be congratulated and revered for predicting all 50 states in the last presidential election, for choosing the Super Bowl winner this year, and for all the money he will make as a great predictor of the future. On the other hand, I’m more interested in those working to change the future than in those who tell me what will happen ceteris paribus.