Ray Jay wrote:I think you need to read the op ed. There is a tremendous amount of detail; it doesn't seem like a puff piece to me.

One potentially dispositive question is what mix of Republicans and Democrats will show up this election. On Friday last week, Gallup hinted at the partisan makeup of the 2012 electorate with a small chart buried at the end of its daily tracking report. Based on all its October polling, Gallup suggested that this year's turnout might be 36% Republican to 35% Democratic, compared with 39% Democratic and 29% Republican in 2008, and 39% Republican and 37% Democratic in 2004. If accurate, this would be real trouble for Mr. Obama, since Mr. Romney has consistently led among independents in most October surveys.

To a political world unfamiliar with mathematical methods that are normally taught in an introductory statistics course, Silver’s prophecy was nothing short of miraculous.

But was it? To find out, I spent a few hours re-building Nate Silver’s basic Monte Carlo poll simulation model from the ground up. It is a simplified version, lacking fancy pollster weights and economic assumptions and state-by-state covariance factors, but it contains the same foundation of state poll data that supports Nate Silver’s famous FiveThirtyEight model. That is, they are both built upon the same assumption that state polls, on average, are correct.

After running the simulation every day for several weeks, I noticed something odd: the winning probabilities it produced for Obama and Romney were nearly identical to those reported by FiveThirtyEight. Day after day, night after night. For example, based on the polls included in RealClearPolitics’ various state averages as of Tuesday night, the Sean Davis model suggested that Obama had a 73.0% chance of winning the Electoral College. In contrast, Silver’s FiveThirtyEight model as of Tuesday night forecast that Obama had a 77.4% chance of winning the Electoral College.
. . .

Thus, of the five major state races in which polls were wrong over the last four years, Silver only got one right. I’m no baseball scout, but batting .200 when it counts won’t get you into the big leagues, let alone the All-Star game.

Silver has made a big deal this election cycle about how state polls are generally more accurate at forecasting the winner of the Electoral College and the popular vote than are national polls. That may well be true, although a Monte Carlo simulation of the final week’s worth of Florida polls in 2000 suggests otherwise.

But assuming it is true, how much actual data in the era of modern polling do we actually have? Maybe three presidential elections’ worth, going back to 2000. Or, if you want to be really generous, maybe eight if you go back to 1980.

Wall Street had exponentially more data when it incorrectly bet on the housing market than we have today when it comes to presidential election polling data. Although we pundits may think a handful of elections qualifies as a robust data set, political polling data simply pales in comparison to the wealth of data we have on the stock market, economic output, or life expectancy, going back to the example of the insurance actuary. But is the science settled on how Stock X will perform tomorrow, or what precise economic growth we’ll see next quarter, or exactly how long each of us will live? Of course not.

freeman2 wrote:By the way, given how different the composition of 2010 voters was compared to 2008 one might have expected the RCP to have gotten it wrong in 2010 (given that polling might have based its composition of voters based on the composition of those who voted in 2008) But the RCP still did pretty well, at least according to Mr Davis's figures. The fact that RCP did pretty well in 2008, seems to indicate that it would.be unlikely for the RCP to become inaccurate in 2012

freeman2 wrote:Mayor Bloomberg seems to think Obama is going to win (else why the late endorsement)...

Doctor Fate wrote:What's more impressive? A nanny-state-lover endorsing Obama or people on Obama's Job Council endorsing Romney? I'll take the latter.

Furthermore, I will say that most people voting for Obama are not thinking of future generations. That's why he does so well with single women: no kids and no worries of the future.

First, all things I complain about in coverage of UK polling are the same in US polling. Most notably warnings about cherry picking, comparing like to like and being aware of methodological differences and house effects from different pollsters. For example, I keep seeing people cherry picking out Rasmussen polls and Gallup polls to claim that Romney is doing better. Rasmussen are one of the most prolific polling outfits in the US, but also tend to produce some of the most Republican results. Gallup use a very tight screen for likely voters that also tends to produce favourable figures for the republicans. Look at most other polls and Obama is doing better.

Secondly, remember that the person who gets the most votes doesn’t necessarily win, it is who wins states with enough electoral votes to win a majority (270) of the electoral college. The average picture across all the national polls in the US has Romney and Obama very much neck and neck. However, polls from the key swing states, which themselves have become very regular as the election approaches, have Obama clearly ahead in terms of electoral votes.

There are various US websites (I’ve already mentioned Pollster.com, though fivethirtyeight tends to be the best known these days) that make projections based on state polling, and these all show Barack Obama with large leads in terms of electoral votes.

This has, in turn, produced some (generally pretty poorly informed) criticism of the projection sites, normally based around what sort of weights they give to different polls, what polls they include and so on. I don’t think these criticisms carry any weight, however even if one is sceptical about the weightings, filters, trends, house effect adjustments or whatever that the various projection sites make, the bottom line is that even if one takes just a crude average of state polls, Obama is still ahead.

The headline figures from YouGov polls over the past couple of weeks show that Obama leads nationally by 49 per cent to 46 per cent and holds the advantage in enough key swing states (including Ohio) to win the election.

It is the detail behind these figures, in particular the state polling numbers, that goes some way to explaining why there is such a large discrepancy between polls.

The surveys of 25 states before and after the first televised debate were true panel surveys. We contacted the same people twice so we could look at actual change among the 25,000 people recontacted (from an initial sample of 33,000).

What we found was only a tiny movement - Obama's lead was cut by 1 per cent. But 80 per cent of previous Romney supporters responded to the second wave of surveying, compared to 74 per cent of Obama supporters.

So the bigger swings that some polls have shown are not caused by voter shifts, but by Obama supporters becoming less likely to respond to surveys after the debate