Yeah, yeah, Obama vs. Romney, Warren vs. Brown, all very big deals. But one of the biggest battles that will be decided (for now, at least) in a couple of days is not on the ballot.
We talk a fair amount about polls here, as does just about everyone who follows elections closely. But there are a few people who know a lot about statistics and who crunch poll numbers in a much more sophisticated way than most of us can. Two of the best of them are Nate Silver, the proprietor of 538 at the NY Times website, and Sam Wang, who teaches at Princeton and runs the Princeton Election Consortium. These guys have been saying for quite some time that, far from the presidential race being a “tossup” or “tight as a tick,” President Obama is the clear, maybe even prohibitive, favorite in this race. As of right now, Silver (who considers both state and national polls as well as some economic and other external factors he thinks might impact the race) has Obama at an 83.7% likelihood of winning; Romney’s chances have faded to 16.3%. Wang (who considers only state polls) is even more bullish on Obama: Wang says that, at best, Romney has only a 2% chance of winning. Silver’s current electoral vote prediction is 305-233 for Obama; Wang’s is 323-215.
Silver, whose work is much better known than Wang’s because of its prominent place at the NY Times, has come under a lot of criticism in this cycle from Romney fans who don’t like what he has to say. That’s not terribly surprising – “kill the messenger” has been in vogue for a long time. What’s been more surprising is the intensity of the criticism of Silver from more centrist types who just really don’t like the geek approach to poll analysis.
Why has the criticism been so strident? In a word, because geeks make pundits look stupid.
Being a pundit is an awesome job, because you get paid for shooting your mouth off, and you really don’t have to have much to back it up. Here’s a fine example: on October 29, Cokie Roberts said on NPR that the “momentum behind Romney” in the race “is real.” But in fact, if you look at the data, what you find is that the race’s movement toward Romney – which did indeed occur after Obama’s epic fail in the first debate – had ended several days before Roberts said what she said. Silver crunched the numbers as of October 24 (five days earlier), and concluded that as of that day, Romney had stalled in the polls, was no longer gaining ground on Obama, and in fact had begun to cede ground back to him. And one day earlier, on October 23, Wang concluded that at that moment the actual momentum in the race belonged to Obama, and that Romney’s claims of momentum were a bluff, designed to take in a gullible corps of pundits who don’t bother to look hard at the data. And that is exactly what it did in the case of Roberts and many others, who for days after Romney started giving up ground in the polls still thought that he was closing the gap with Obama.
Here’s another excellent example from David Brooks, who regularly publishes at the NY Times and yaks on PBS. It’s particularly spectacular because Brooks started off this column, dated October 22, by claiming that “I’m a pollaholic. For the past several months I have spent inordinate amounts of time poring over election polls.” And then he offered up a couple of assertions about what the polls showed as of that date that were, to put it plainly, wrong. Sam Wang eviscerated him in a truly glorious post a couple of days later.
It was fun to learn of David Brooks’s addiction to polling data. He spends countless hours on them, looking at aggregators, examining individual polls, and sniffing poll internals. From all of this, what has [Brooks] learned?
- Today, President Obama would be a bit more likely to win.
- There seems to be a whiff of momentum toward Mitt Romney.
I am having a sad. All of that effort, and his two conclusions still have two major errors….
1. President Obama would be a bit more likely to win. This is false – he’s a lot more likely to win. Look at the Princeton Election Consortium’s EV histogram, which tabulates all 2.3 quadrillion possible combinations of states to give a clear snapshot of the race. [there follows some geeky analysis] …
2. There seems to be a whiff of momentum toward Mitt Romney. Ah, yes…Ro-mentum! Bobo has taken the bait…. As you can see, Ro-mentum ended around October 11th, the date of the VP Biden-Ryan debate and reversed around October 16th, Debate #2. Now the median EV expectation is at a plateau around Obama 293 EV, Romney 245 EV. [more geeky analysis ensues] … If anything, the race is starting to look a bit static.
Posts like that one, which make the target pundit look like an idiot, is why the pundits hate the geeks. Because of course, the pundits cannot possibly answer the criticisms. Instead, they trivialize what geeks do, as Brooks did when he said “[t]he pollsters tell us what’s happening now. When they start projecting, they’re getting into silly land.” (That’s true, but only in a trivial sense that misunderstands what people like Silver and Wang do when they assess probabilities – in fact, the geeks are quite careful about not “projecting” in the way that pundits like to do. More on that point in this interesting post.)
In these closing days of the election cycle, even the pundits have started to realize that the state polling is consistently showing an Obama win, and they are starting to back off their “tossup” predictions a bit. This quote from Mark Halperin at Time on Thursday is especially wonderful.
I’m going to take a risk here and peel back the curtain to tell you what insiders are thinking. This is not my opinion and this is not a projection of who’s going to win but insiders in both parties look at the last few days of swing state polls and say, “The President may have this.” … Right now, there’s a danger for Governor Romney that elites are starting to think, in the last 24 hours, that these leads are, as the Obama campaign has said for a long time, small but persistent and consistent. And, I think, it may not effect the coverage that much but there’s a bit of a tipping point here.
Amazing, right? “In the last 24 hours,” these mysterious “elite” “insiders” in both parties who apparently whisper to Halperin everything they are thinking have noticed what Silver and Wang have been saying for weeks: that Obama has consistently been ahead in enough swing states to win the election. My goodness – if those “elite insiders” only noticed the state polling a few days ago, they should all be fired. And yes, yes, polls are snapshots, so events can and do (e.g., the Democratic convention; the 47% video; the first debate) cause the polling to shift. But the point is this: if you look at the polling in an intelligent way, this race is not and has never been a tossup. It has always been at least “Lean Obama,” and even that assessment is probably too conservative by now.
My own beliefs are that (a) polling works quite well, especially when (as is the case now) there are a lot of polls being taken in important jurisdictions like Ohio by a lot of different organizations; and (b) people like Silver and Wang know what they are doing, whereas pundits often don’t. Silver argues that, given the current state of the race and the amount of time left, “for Romney to win, state polls must be statistically biased.” And it’s not just a couple of them – it’s practically all of them. The whole enterprise of statewide polling in the swing states would have to be seriously off. And I doubt that’s the case.
I therefore think that, on Tuesday, America is going to reelect President Obama, probably by an electoral vote margin that is very close to what Silver and Wang are predicting – i.e., Obama gets a number in the low 300s.
If that happens, the geeks will have won this round. But the battle will continue – don’t expect the pundits to give up their cushy gig quite that easily.