AS THE US re-elected President Barack Obama, mathematics fans crowned their own king: statistician Nate Silver.
He was the most prominent of a handful of political forecasters who applied sophisticated techniques to polling data from individual states to reveal an early Obama lead that never went away (see charts).
That was in stark contrast to the findings of traditional opinion polls and the predictions of political commentators and pundits. Almost invariably, they declared the race too close to call, providing a neat, but ultimately false campaign narrative.
The new breed embodied by Silver is clearly a force to be reckoned with, so what impact will their predictive power have on the elections of the future?
Obama's win alone does not validate Silver's techniques. Nor would a Romney win have disproved him, as his final figures showed a 10 per cent chance of the race going the other way. Still, there's no doubting the accuracy of his method.
Opinion polls are noisy due to quirks in the sampled population that are not representative of the whole. Each polling company has its own methods to deal with this, which in turn introduces bias. Silver, whose blog is hosted by The New York Times, adjusted for these effects by aggregating many polls and weighting them according to historical accuracy.
Rather than a simple national average, a figure that media commentators often rely on, this provided a range of possible vote splits for each state, some more likely than others. By simulating the election thousands of times, Silver calculated the ways these could combine and their likelihoods. The result was a prediction for how each of the 50 states would vote - every one of which proved to be correct - plus an overall probability of 90 per cent that Obama would win.
Is this the end for pundits? Not so fast, says Brendan Nyhan, a political scientist and media analyst at Dartmouth College in Hanover, New Hampshire.
Pundits often serve to convince one side or the other that there is still a chance of winning, or can spin a yarn that makes for good headlines.
By contrast, stats that predict an election result far in advance could suck excitement out of the campaign. "Pundits will have a job for a long, long time," says Sam Wang of the Princeton Election Consortium, which ran Silver-like analyses that predicted the outcome in 49 out of 50 states.
Sound stats don't necessarily affect voter turnout, at least not directly. "People follow politics like sports," says Nyhan, and stick with their side even if it is the underdog.
Instead, sophisticated stats may have most influence on which voters are targeted, similar to the way they now help baseball managers pick winning players, as popularised in the book and film Moneyball. "The use of statistics is very similar whether you are talking about sports, business or politics," says Ben Alamar, a sports analyst at Menlo College in Atherton, California.
Could this affect results? In this election, the Obama campaign employed its own version of Silver's methods to focus efforts on areas the president had a chance of winning. "We ran the election 66,000 times every night," one Democrat told Time magazine. Expect this approach to intensify in future, and for both sides to do it.
One campaign's more sophisticated use of statistics may not have been the deciding factor this time round, but in future elections it could be. Memo to wannabe presidents: hire geeks, not pundits.
If you would like to reuse any content from New Scientist, either in print or online, please contact the syndication department first for permission. New Scientist does not own rights to photos, but there are a variety of licensing options available for use of articles and graphics we own the copyright to.
Have your say
Only subscribers may leave comments on this article. Please log in.
Only personal subscribers may leave comments on this article
Subscribe now to comment.
All comments should respect the New Scientist House Rules. If you think a particular comment breaks these rules then please use the "Report" link in that comment to report it to us.
If you are having a technical problem posting a comment, please contact technical support.
andrew young real life barbie zipper armenian genocide asteroid mining memorial day ivan rodriguez
No comments:
Post a Comment