Posted on | April 9, 2012 | 9 Comments
In 2008, a 30-year-old baseball stat nerd looked at the reams of public research product churned out by the nation’s 1,500-plus daily newspapers, and concluded that, though “there is nearly as much data as there is for first basemen,” the “understanding has lagged behind.” So Nate Silver launched 538.com (named after the number of votes in the Electoral College), and through sheer intellectual rigor and superior numeracy went on to outperform all comers in the political prediction business that year. As Silver later explained in The New York Post, too often “polls are cherry-picked based on their brand name or shock value rather than their track record of accuracy,” and “demographic variables are misrepresented or misunderstood.”
Silver, who was later hired by The New York Times (a blogger-to-riches story that would have made headlines a decade ago but is no big deal nowadays), is a living refutation of the Labor Theory of Value. All those thousands of big-media reporters and commentators and pollsters, paid full time to analyze and interpret political information, got their clocks cleaned by a sports geek blowing off steam after hours.
You should read the whole thing. By the way, Nate Silver is a data omnivore and, in recent months, I’ve had the sneaky feeling he’s been reading my blog, although he’s never cited me.
In early December, when Santorum seemed to be permanently stuck in the single digits, I kept watching the Iowa polls looking for some indication that Santorum was ready to break out of the “second tier.” My reasons for suspecting that this would eventually happen:
- Santorum had finished fourth in the Ames Straw Poll, despite his cash shortage, demonstrating a strong organizational ability, which matters in caucuses. The third-place finisher (Tim Pawlenty) and fifth-place finisher (Herman Cain) in the Ames Straw Poll had both dropped out of the race.
- Santorum is a pro-life social conservative, which is always a winning formula in Iowa (see Mike Huckabee, 2008).
- Santorum had done more retail campaigning in Iowa than anyone, and Iowa is a place that puts a premium on retail campaigning.
- Santorum had the best scorecard on immigration of any of the GOP candidates and, although it seldom gets mentioned by the political press, illegal immigration is a huge issue in Iowa.
- Once Newt Gingrich started melting under the avalanche of Romney attack ads, the compelling question became: Where would Newt’s lost support go?
A lot of people thought they’d go to Rick Perry, but (a) Perry’s campaign had foolishly snubbed the Ames Straw Poll in August, and (b) Perry’s debate flubs — especially his “you don’t have a heart” blunder in Orlando — had marked him as damaged goods.
What about Michelle Bachmann? She had already had her moment in the Iowa spotlight, winning the Ames Straw Poll, and if she was still in the “second tier,” there must be some explanation for the resistance. Was this the result of right-wing sexism? I dunno.
The point is, Bachmann had her chance already, whereas Santorum had been largely ignored by the major media (certain people at Fox News really seem to hate Santorum) and he was therefore obviously overdue for his Flavor-of-the-Month moment.
Anyway, as in so many cases over the past year, my gut hunch was more than just a gut hunch, and from time to time I would notice Nate Silver pointing out at Five Thirty Eight something about the Iowa polls that I’d mentioned a day or two earlier. Maybe Silver started checking this blog because, after Cain dropped out, I became a clearinghouse for Rick Santorum news.
Were we just coincidentally watching the same phenomena? Maybe, but Silver is shrewd, always looking for some analytical edge. If he was reading this blog, he would have seen headlines like these:
Nate Silver knows that (a) polls are a lagging indicator and (b) if you’ve got a series of different polls to look at, the thing to look for is the trend.
The question in December was, “Which of the ‘second-tier’ candidates is making a move?” And there was a series of polls — including both PPP and Rasmussen — that showed Santorum making a move in relation to Bachmann and Perry. When the CNN/Time poll was published on Wednesday, Dec. 28, showing Santorum going from 5% to 16% in the span of three weeks, moving into in third place ahead of Gingrich, Perry and Bachmann, the momentum became unstoppable.
But that poll was still more than a week away from publication when the Ron Paul panic briefly took hold among Republicans, some of whom badly overreacted to this item:
“The model gives [Ron Paul] a 44 percent chance of winning Iowa based on the current standing of the candidates and the historic uncertainty of polling-based forecasts. Mr. Romney has a 32 percent chance of winning, while Mr. Gingrich’s chances have crashed to 15 percent.”
– Nate Silver, New York Times, Dec. 19
My reaction on Dec. 19: “Are you people on drugs?”
Ron Paul winning Iowa never made sense as a possibility, and the absurd fear of it was mainly a panicked reaction to propaganda from the Romney, Gingrich and Perry campaigns. Nate Silver was merely reporting the results of his forecasting model — “designed to be quite aggressive” — and the ridiculous overreactions by some Republicans had to be taken with a grain of salt.
All the while the Great Pre-Christmas Ron Paul Panic of ’11 was taking hold, I kept looking at the exact same poll data and smiling, because I had more than a gut hunch about the final results.
Santorum won Iowa without a single poll showing him better than third.
The experts never admit they were wrong, because then they wouldn’t be regarded as experts anymore. But the New York Times is never going to hire me — hell, I’m not even good enough for BlogCon — so it’s time to ask you to hit the freaking tip jar again.