Now that the presidential debates have begun, everybody’s got an opinion about how this thing is going to turn out. Some have loud, kind of paranoid ones. Others have more rational and informed ones, including media wunderkinds Nate Silver and Sasha Issenberg — pros at crunching quantitative data, employing scientific methods and applying historical perspectives to analyze political races.
Their areas of expertise make them a bit of a wonk dream team: Issenberg examines how campaigns try to influence voters before elections, while Silver writes about how to predict the results.
Want to know whether Romney’s or Obama’s camp is doing a better job at mobilizing voters — actually getting them to register and turn out on polling day? Issenberg — Slate columnist and author of “The Victory Lab” ($26, Crown Publishers) — is your guy.
Need to know which polls to trust or which presidential candidate to pick for an Election Day wager? Turn to Silver, stats star behind the New York Times’ FiveThirtyEight blog and author of “The Signal and the Noise” ($28, The Penguin Press).
As of our interviews with Silver and Issenberg on Sept. 20 and Sept. 24, respectively, both saw a greater potential for Barack Obama to win the presidential election than Mitt Romney. Read on for their insights.
Nate Silver finds himself answering one question fairly often these days.
“People say, ‘Would you place money on your election odds?’ ” says Silver, a statistician, writer and founder of the political blog FiveThirtyEight.com, which forecasts, among other things, the likelihood of victory for each of the two main presidential candidates. (The original site, which was picked up by the New York Times as a blog in 2010, first started with presidential predictions in 2008 and takes its name from the number of electors in the U.S. electoral college.)
“I probably do have a lot of money riding on” the outcome of Nov. 6, Silver says. He does, after all, stake his career on the reliability of his forecasts.
Silver gained attention around the 2008 elections, when he correctly predicted the presidential voting results in 49 of 50 states — missing Indiana, which ended up going for Barack Obama by 1 percent — and all of the Senate races. (Prior to that, Silver dealt primarily in sports odds and developed PECOTA, a system for forecasting baseball performance.)
As this year’s election draws closer, Silver realizes a lot of eyes are again on him. The Times slot has given FiveThirtyEight a much wider audience, and he’s become a nerd-chic hero in political circles. He’s also just published his first book, “The Signal and the Noise,” about the science of predictions — what makes some forecasts succeed and others fail.
Silver formulates forecasts for FiveThirtyEight by compiling polls, weighing their data and running the numbers through complex calculations. Because so many polls are factored in, “our predictions tend to be relatively stable” from day to day, Silver says. Still, it’s possible that an October surprise could skew the numbers dramatically for the candidates, and by that point, he says, it could be too late for the loser to catch up.
That element of chance is what keeps Silver from feeling overly confident in his forecasts — a good thing, in his eyes, since he says those who are too confident in their predictions are the ones you should trust the least.
And, as Silver would know better than most, “people tend to make better predictions when they have financial incentives to get them right.”
On his wonk fan club:
“It’s not too crazy. Every couple of months, I’ll get interrupted when I’m eating a slice of pizza or something. But it’s usually a 60-year-old dude.”
On FiveThirtyEight’s election prediction:
“We estimate that if the election were held [Sept. 20], Obama would have about a 94 percent chance of winning,” Silver says. “But, hey, that 6 percent is going to come up often enough.”
Politics and Prose, 5015 Connecticut Ave. NW; Thu., 7 p.m., free; 202-364-1919. (Van Ness)
At one point in Sasha Issenberg’s new book, “The Victory Lab,” 2008 Obama campaign consultant Ken Strasma says something unsettling: “We knew who … people were going to vote for before they decided.”
Sci-fi as this sounds, it may actually be true, thanks to a sophisticated algorithm that predicted individuals’ views based on weekly polling calls to thousands of voters in the months before elections, among other data. There’s nothing illegal or unethical about it, but “there are all sorts of reasons why this stuff can sound creepy,” Issenberg says.
“I am not a number! I am a free man!” you may be saying. Well, even if you don’t want political campaigns making assumptions about you and your ballot, there’s not much you can do about it.
“If you vote, or if you’re registered to vote, your voter registration information is public,” Issenberg says. “And that already is the basis for campaigns’ starting to draw conclusions about you. And they will draw conclusions about you.”
Ick factor aside, the “secret science” of how campaign strategists have worked to win votes over the past decade is important and interesting — and it hasn’t gotten much notice in the mainstream media, Issenberg believes. (Which is what spurred him to write “The Victory Lab” in the first place.)
Rather than trying to turn out supporters by “brute force” using ads, as campaigns traditionally did, strategists are shifting to subtler forms of persuasion. They’re “microtargeting” smaller pockets of desired voters, whose behavior they believe they can influence: Individuals who seem likely to sympathize with the candidate but unlikely to actually make it out on Election Day are the focus of voter registration drives; individuals who appear undecided or likely to shift allegiances are solicited with personalized, persuasive direct mailings or calls.
“Campaigns are a lot less focusing on changing people’s minds than they are on changing their behavior,” Issenberg says.
So, you can be sure someone out there will be pleased if you make it to the polls. That is, of course, if they didn’t already know you were going to go before even you did.
On voter behavior and campaigns:
“Almost all innovation I write about has taken place since 2000. Some things people would have laughed at in the ’90s. Then, Florida got decided by 500 votes, and nobody’s laughing anymore.”
On his election prediction:
“I’ve been doing reporting for Slate this year about the Obama campaign, and I’m incredibly impressed with [the campaign’s] empirical sophistication. … Right now, I’d bet on the Obama campaign.”
Politics and Prose, 5015 Connecticut Ave. NW; Oct. 20, 1 p.m., free; 202-364-1919. (Van Ness)