When the dust settled and the 2014 election finally reached its merciful end, it could have been argued that the only major institution that had a better day than the Republican Party was FiveThirtyEight, the new data-journalism website launched by ESPN and run by former New York Times superstar Nate Silver. Because while Silver’s model didn’t have quite the lights-out success with regard to its predictions as was the case in 2012, it came damn close. Not bad, considering the site had launched less than seven months prior.
Now that the election is in the rearview mirror, FiveThirtyEight has more of a chance to show that its data-driven approach to journalism can tackle subjects well beyond electoral politics, including entertainment, science and even food (burritos, to be precise). This year will also see the paperback release of Silver’s best-selling book, “The Signal and the Noise: Why So Many Predictions Fail — but Some Don’t,” which endeavors to help numbers geeks and laymen see predictions in a new, more complicated way.
Recently, Salon spoke over the phone with Silver, discussing lessons from 2014, the latest bits of suspect conventional wisdom emanating from D.C., what Silver thinks of his critics, and which GOP governor may be flying under the radar when it comes to nabbing the party’s nomination in 2016. Our conversation is below and has been edited for clarity and length.
So to start off, I want to ask you how you’d apply the signal/noise schema you lay out in the book to the 2014 election?
A couple of interesting phenomena about the past election. Number one is, you had what happened in 2012 in reverse … in 2012, of course, you had the unskewed polls crowd screaming out that the polls are biased against Mitt Romney, even up into the very last minute with Karl Rove’s Election Day tantrum.
Instead, on Election Day , you had Republicans across the board beat their polls by 3 or 4 points on average. That’s important, I think, in a systemological way … I think that development was maybe helpful to keep Democrats on their toes, potentially …
A lot of basic mistakes in prediction are made because people look at the most recent trend and assume that it will continue on indefinitely. For example, whereas in 2012 you certainly had a Republican skew in the polls relative to reality, that was still just a sample size of one. You had actually the reverse skew instead in the 2014. The pollsters corrected a lot of the issues [from 2012], and maybe overcorrected potentially as well.
So it’s a pretty esoteric statement in what I think was interesting about 2014, but I hope to emphasize that, hey, the world is a complicated place and it’s not as simple as averaging the polls and making a perfect prediction. There’s a lot of uncertainty in an environment where somewhere between only 2 percent and 10 percent of people actually respond to the pollsters' phone calls … That’s a challenging environment and one that people have to be a little bit worried about.
Sticking with 2014 for a bit more: You’ve been a pretty vocal critic of some parts of the punditry’s conventional wisdom, especially those narratives that are farthest removed from empirical foundations. Have you seen any bits of conventional wisdom about 2014 pop up that you’d recommend we interrogate a bit more before embracing?
I think we would have seen more of it if Democrats had had a good 2014.
I’m a little suspicious of the narrative about the emerging Democratic majority for a couple of reasons. One thing is that there’s a really long history of Republicans after 2000, 2004, talking about a permanent conservative majority and then falling totally flat. I think people overestimate how predictable politics is over the long term, maybe in part because the party can start to act smugly if they assume they’re going to win every election. I think that narrative [in the Democrats’ favor] took a hit with 2014. But in the long run, hopefully, it should keep Democrats thinking actively on how to win the majority in elections.
I think people maybe are a little too quick to point out the death of the so-called Tea Party, where people are trying to measure it based on, did any senators lose in primary challenges, like Dick Lugar did a couple of years ago? And you didn’t have any senators, but you had a lot that came close — and, historically, the vote share incumbents received was pretty low. You did have Eric Cantor, of course … so I think the GOP learned the lesson that, hey, if you nominate more competent, moderate candidates, it’s usually helpful.
On the other hand, from the very much unresolved question about whether the GOP establishment has reined in the unhelpful characteristics potentially of grass-roots conservatives and the Tea Party or not. And that’s a pivotal question, really, for 2016. You have what in my view looks like a pretty deep field of establishment candidates between Bush and Romney and Rubio and Christie and whatnot. Will those candidates gain enough support with the more conservative third or conservative half of Republican voters, especially in Iowa and South Carolina and states like that to win the nomination smoothly? Or would someone like Rand Paul actually have a chance?
Somebody who could bridge the two sides would be ideal for them.
I think someone like Scott Walker potentially has a good opportunity to win the GOP nomination in 2016 because he’s kind of on this bridge between being a true blood conservative, Tea Party candidate, and at the same time having a lot of credibility with the GOP establishment. So someone like that I might think is an underrated candidate for 2016 versus, say, a Chris Christie, who’s a very prominent and entertaining politician to those of us on the East Coast but does not poll that well anymore with conservative Republican voters, and has at the same time managed to turn off certain independents. So I’ll probably have as a guess for 2016 different candidates I think are overlooked or overrated by the media, but I guess it’s kind of too early for flimsy, overall thematics to develop. I’m sure we’ll get some of those, though, before too long.
About 2016, I wanted to ask how you guys are going to approach the issue of money’s impact on various races. How do you see its role, especially since Citizens United (and other subsequent rulings) changed the status quo so significantly?
There are narrow ways and broad ways to answer that question. In the narrow sense, obviously fundraising is a useful indicator, predictor of all types of different things, but when you have money coming in and there’s less public disclosure required then you’re a little bit more in the dark about it.
From a macro point of view, one of the more firm, robust findings about how campaign finance works and some of the effects on election campaigns is that you’ll encounter diminishing returns. If you have the Koch brothers or liberal groups or whatever else, if you take a candidate for Congress in some commercial district in North Carolina and she’s raised no money at all, and you give her a million dollars, that makes a huge impact. If you take Mitt Romney, who’s already raised $800 million, and you give him another million dollars, then you’ve already pretty much maxed out on exposure to voters … At some point, the presidential candidates, or maybe the candidates in the best-funded Senate races, have more money than they really know what to do with, potentially. So I’m not sure that money like the Koch brothers’ would make a lot of impact on the presidential race.
There are some trade-offs, too; think about the 47 percent tape, for example. If you’re a presidential candidate now who has to spend a lot of time catering to the whims of these extremely strange, extremely wealthy people — even relative to ordinary political donors who are themselves pretty high income — it could potentially affect the policy stances that you’re taking and the way you think about and describe the world in ways that could move you further away from the mainstream position. So it’s not without total risk either.
That’s a long-winded way of saying, I think this money of Citizens United has had a big effect potentially on races of Congress and less so the higher up the ballot you go.
To switch gears for a second and move away from electoral politics, I’m curious about how you respond to some of the chattering that’s come from people in the media that basically says FiveThirtyEight’s been influential overall but that the site itself isn’t having the impact that it “should.” Do you get frustrated with that? Or focus more on how the media in general is integrating more empiricism into its political coverage? Or do you feel more like the athlete who gives the clichéd answer about taking it one game at a time and just focusing on your business, etc.?
I feel sometimes like giving that Marshawn Lynch answer [laughs]: Thanks for asking, but we’re pretty busy. We’ve got a lot of planning to do for 2016.
I have different, somewhat contradictory responses to it. [If] you launch a new product, a new journalistic product, which FiveThirtyEight effectively was, last year, there’s two things [that’ll happen]: one is, there’s going to be a lot of hype about it, which you may contribute to yourself; and, number two, you’re not going to be all that great at what you’re doing for the first little stretch — and not just like for the first week, but for the first month. We learned a lot from Grantland; Bill Simmons told me it’s going to really take 12 months to 24 months before you feel like you’re really even firing on most cylinders; so a lot of the criticism was bullshit. If you feel like you’re not putting a great product out there, then maybe it’s not productive to really engage with [the criticism] too much.
I do think one of the problems with running a news organization and being a journalist in general in today’s age is that a lot of people sometimes write too much for other journalists and others in the industry. So people say, FiveThirtyEight’s not having as much influence; I don’t know how you measure that. I do know that we get about five times more people coming to our site now than we did at the New York Times on a typical day, and the media buzz [then] was different. (There are all kinds of problems with using that, too, as a benchmark.)
We’re thinking about this effectively, I would say … We think all the time about: Are there ways to be more on top of the news cycle at certain periods of time … But if you’re an NBA fan, you know that teams like the Houston Rockets have learned that [the best shots] in the NBA are layups and 3-pointers … You don’t want to take these long, contested [2-point] jumpers, the Carmelo Anthony amateur shots. I feel like FiveThirtyEight in its first year has been a little bit Carmelo Anthony [laughs], we’re trying to write the day-after story where it’s not going to be a huge slam dunk, and at the same time it’s not what people are demanding then and there… So I think we have to do more things on both of those ends and not just taking those long, contested jumpers.