As President Barack Obama celebrated re-election, number-crunching geeks everywhere could revel in redemption.
For months, political pundits and reporters said that the race between Mr. Obama and Mitt Romney was too close to call.
Not so, insisted a new breed of political analysts who rejected intuition and relied instead on data - vast amounts of it, run through computer algorithms - to arrive at their election predictions. Many of these models showed that Mr. Obama would win with relative ease. The forecasts drew rebukes from Republicans and other critics who insisted that the statisticians were biased.
The highest-profile of these forecasters is Nate Silver, whose blog, FiveThirtyEight, is licensed by The New York Times. Mr. Silver, a former baseball statistician who made a name for himself by correctly predicting the results of the electoral college votes in 49 of the 50 U.S. states in the 2008 election, did at least as well this time around. If the c lose result in Florida goes Mr. Obama's way, Mr. Silver will have gotten all 50 states right.
Mr. Silver is not the only so-called quant who had a good night. Simon Jackman, a Stanford University professor whose work is published by The Huffington Post, and Drew Linzer, a professor at Emory University who runs a Web site called Votamatic.com, predicted the exact number of electoral college votes that Mr. Obama received - 332, assuming the Florida victory is confirmed.
Another widely followed analyst, Sam Wang of the Princeton Election Consortium, predicted 303 electoral votes, the total that Mr. Obama will receive if it turns out that he actually lost Florida.
Mr. Linzer appears to have been especially prescient, because he was predicting 332 electoral votes for Mr. Obama as long ago as June.
‘‘We are all using the same data (more or less) and so it doesn't surprise me that we're all reaching similar conclusions,'' Mr. Linzer wrote on Votamati c.com on Election Day. ‘‘The real challenge is producing meaningful and accurate forecasts early in the campaign.''
The number-crunchers made some old-fashioned pundits, relying on less scientific means, look silly. In the build-up to the elections, Joe Scarborough, a talk-show host on MSNBC, an American television channel, said of Mr. Silver and other data-driven analysts: ‘‘Anybody that thinks that this race is anything but a tossup right now is such an ideologue, they should be kept away from typewriters, computers, laptops and microphones for the next 10 days, because they're jokes.''
Instead of gut instinct, Mr. Silver relies on polls, emphasizing the data from individual states, in conjunction with the national surveys. This is because a U.S. presidential election is, in effect, a series of state elections, in which the winning candidate must post a majority in the electoral college, whose members reflect the choice of voters in the individual state s.
While the national polls showed a close race, with some favoring Mr. Romney and others Mr. Obama, the new-school analysts had been saying that the state-by-state calculus made it a near certainty that Mr. Obama would stay in office. On Election Day, Mr. Silver put a 90.9 percent chance on this probability. Mr. Wang was even more certain, predicting a 98 percent likelihood.
‘‘On the whole, the polls had it - were quite good,'' Mr. Silver said in a video posted on The New York Times Web site. ‘‘The polls are usually quite good. A number of cycles in a row now, any one poll is often a little bit dubious, but when you average them together they seem to do a better job of kind of calibrating where the average voter is.''
To avoid aberrations in individual polls, Mr. Silver's model crunches numbers from a variety of voter surveys, and also incorporates other data, including economic statistics.
Mr. Silver has not always gotten it right. In 2010, for example, he overestimated the number of Parliamentary seats that the Liberal Democratic Party would win in British elections.
During the 2012 U.S. campaign, he was rebuked by the public editor of The New York Times, Margaret Sullivan, over his response to the provocation from Mr. Scarborough. On Twitter, Mr. Silver challenged the television host to a $1,000 bet that Mr. Obama would win; he later raised this to $2,000, though Mr. Scarborough did not respond.
This prompted Ms. Sullivan to write that ‘‘the wager offer is a bad idea - giving ammunition to the critics who want to paint Mr. Silver as a partisan who is trying to sway the outcome.''
Mr. Silver writes for The Times and has a desk in its newsroom but is not on the paper's staff. He has said that in 2008, before his blog was licensed by The Times, he supported Mr. Obama's presidency.
On Monday, the day before the election, one in five visits to The Times's Web site included a visit to Mr. Silver's blog, the newspaper said.
After Mr. Obama's re-election, Mr. Silver was being congratulated, too. ‘‘Nate Silver wins & data is vindicated,'' the Poynter Institute, a journalism research group, wrote on Twitter. ‘‘So who's an embarrassment to journalism now?''
No comments:
Post a Comment