So tweeted Nate Silver at 2.29AM the night of Obama’s re-election, thereby allowing liberals like myself to safely breathe a huge sigh of relief. In the run up to that night most pundits and news organisations repeated ad nauseam that the election was too close to call. Silver on the other hand, using his model which aggregates polls giving extra weight to the more reliable ones, had given Obama a greater than 65% chance of re-election for the preceding couple of months. And in the end the New York Times blogger not only called the presidency right, but also correctly predicted each of America’s fifty states.
His approach earned the ire of both the conservatives he predicted were going to lose, and the pundits who relied on more old fashioned forms of psephology. Joe Scarborough, the conservative host of MSNBC’s Morning Joe, had dismissed Silver as a “joke”, whilst websites like Unskewed Polls purported to correct for the liberal bias supposed in Silver’s model. In her Wall Street Journal blog Peggy Noonan, after cautioning that “nobody knows anything”, predicted a Romney win on the basis of a number of arguments, including that Obama seemed “tired and wan” whilst Romney looked like “someone who’d just seen good internals”. A special mention must go to conservative political consultant Dick Morris who predicted a Romney landslide, and whose tweets as the results unfurled make for amusing, schadenfreude-inducing reading.
I was as delighted as any of Silver’s many readers – his blog at one point accounted for 25% of the NYT site’s traffic – to see his empirical approach win-out against these tea-readers, but I think there are several downsides to his growing popularity that have been overlooked by relieved liberals.
One straightforward downside is that people’s focus on polls often comes at the expense of a focus on more substantive issues. Rather than learning about, say, a political party’s economic policies, we learn what other people think about those economic policies, and what other people think is usually reduced to a simple ‘boo!’ or ‘hooray!’.
But worse than this, polls can actually influence public opinion, creating a positive or negative feedback loop as voters jump on the bandwagon of a winning candidate or leap off that of a losing one, a phenomenon that has been well-documented by academics. Moreover polls can also drive the policy agenda, as politicians become oversensitive to the shifting whims of public opinion.
Worse still, polls can be used to manipulate public opinion by partisans of a particular cause or party. Silver’s own prognostications prompted many other blogs, newspapers, and cable news shows, to devote time and space to asking ‘Why is Romney losing?’, which in turn had a damaging effect on Romney’s campaign. As conservative blogger Matt Lewis noted, “when the public sees that a prominent New York Times writer gives Barack Obama a 70 percent chance of winning, that can become a sort of self-fulfilling prophesy. It has consequences. It drives media coverage. It dries up donations”. To avoid problems like this countries such as France, Mexico, and South Korea actually ban the reporting of opinion polls immediately prior to an election.
So whilst Silver introduced statistical rigour to a previously speculative domain, and at the same time gave succour to anxious liberals, there are good reasons to feel ambivalent about his growing prominence.