Election aftermath: How’d pollsters do?

Poll aggregators weight individual polls according to credibility

Perhaps the only people who had a better election night than President Obama are the poll analysts who predicted his win. After all, the pollsters’ work is now done.

Polling aggregators such as Nate Silver, who runs the FiveThirtyEight blog for The New York Times, and Sam Wang of Princeton Election Consortium correctly predicted Obama’s victory, the states he carried and his margin of victory.

Silver had been under fire from Republicans for consistently putting Obama’s chance of winning in a range of 60% to 90%.

Poll aggregators average poll results but also weight individual polls according to their own analysis of each poll’s credibility— each aggregation has its own “secret sauce.” Wang, for instance, says he nailed Obama’s popular vote margin by looking at state polls rather than national polls.

But aggregators were right partly because the underlying polls they used were also solid.

“Despite efforts by both the right and the left to discredit the pre-election polls before the election when they didn’t like the results, the final results were generally right on,” says Jon Krosnick, a political scientist at Stanford University.

Pre-election surveys “were about as good in predicting the actual margin as they were in 2008, even though this was a much closer election,” says Costas Panagopoulos, a political scientist at Fordham University. He compared 28 polls to actual election results and found on average they predicted a vote gap of 1.07% between the two candidates. The gap ended up being 2.2%.

Obama’s popular vote margin was larger than polls suggested: “There seemed to be some movement at the end, and I’m not sure the polls had time to catch up,” said polling analyst Sean Trende.

During the campaign, Obama strategist David Axelrod accused Gallup of having flawed methodology when its polls showed Romney leading. Romney pollster Neil Newhouse said that Democrats were overrepresented in polls that showed Obama ahead.

Polling analysts have been critical of Gallup — one of the largest and best-known polling companies, because its pre-election surveys consistently

“Despite efforts by both the right and the left to discredit the pre-election polls ... the final results were generally right on.” Jon Krosnick, Stanford political scientist

gave Romney a strong edge. Gallup is USA TODAY’s partner in the Swing States Poll.

Gallup stands by its polling, says director Frank Newport, noting that its final survey, out Sunday, showed the two within 1 percentage point of each other in the popular vote.

Pundits should give more weight to polls even when they don’t like the results, Wang says — comparing it to ignoring hurricane forecasts because you don’t like the idea of your house blowing down. “Their gut instincts were wrong, and it would be a good idea for them to ground their judgments in quantitative data.”