Numbers overload: Polling data hype sways voters

Treating presidential polls as gospel is a little like placing political faith in the lifespan of a fruit fly.

“People tend to subscribe a more durable nature to polling data,” said Russ Baker, a political science professor at Rutgers University. “It’s more ephemeral.”

That’s because polls, as those who conduct them stress, are simply snapshots.

For example, two national polls released Monday suggested Republican presidential nominee Mitt Romney got a big bounce from last week’s first presidential debate – viewed across the political spectrum as a victory for him.

Two very different polls in battle for White House

A Pew Research Center survey showed Romney taking a 49%-45% lead over President Barack Obama after trailing by similar margins for weeks. The Pew survey was conducted over the four days following the October 3 debate.

Gallup’s daily tracking poll of registered voters, which draws its numbers from a seven-day average, indicated the five-point gap that Obama held before the debate had closed to three.

But Romney drew even at 47%-47% if only the daily numbers from the days after the debate were used.

Pundits whip themselves up with every move in the polls, political experts say. And that leads to a climate in which even slight shifts in data that are well within the margin of error influence both coverage and conversation.

There was no doubt that Romney’s strong debate performance – or Obama’s weak showing – impacted the numbers. And the headlines and chatter following the polls concentrated on Romney’s apparent surge.

Some pundits reacted strongly.

“The Pew poll is devastating, just devastating,” Daily Beast blogger Andrew Sullivan wrote under the headline, “Did Obama Just Throw the Entire Election Away?”

Others were more restrained.

“I’m not sure I would look at this Pew poll alone without the broader context of data,” New York Times blogger Nate Silver said Monday on CNN’s “Piers Morgan Tonight.”

“We’ve seen some other polls … for example, the Gallup tracking polls actually move slightly toward Obama, so right now it’s close enough that you’re going to have some polls showing a lead for either candidate and that makes it more dramatic and makes for big headlines but we think the consensus of data shows what will probably settle into being a very narrow lead for President Obama, probably one or two points on average, that is an educated guess.”

Silver said it’s not just pundits who feed on movement in polls. The campaigns feed the frenzy by pushing them – when they like the results.

Obama above 50% in some battleground polls

“I saw some Democrats in my Twitter feed complaining about the Pew poll when they were lauding it, of course, when they had Obama up by seven,” Silver said. “I think we have a tendency now in our culture to just get rid of data we don’t like. But what I do try to do is look at what all the polls say and you see some where Romney has got a gigantic bounce, some where you haven’t seen much of a bounce at all on average.”

An Obama campaign official pointed to a shift in the party identification percentages used in Pew’s mid-September poll compared to the poll released Monday.

The earlier comprised 39% Democrats, 29% Republicans and 30% independents, while Monday’s poll included 31% Democrats, 36% Republicans and 30% independents.

“This is far bigger than any one-month change in party ID ever reported by Pew in the past,” the Obama campaign official said.

A Romney campaign official said the GOP effort was “encouraged by the enthusiasm we are seeing from supporters who are energized, as well as undecided voters who are now giving the governor a new look.” They are “continuing to work hard to bring home persuadable voters,” the official said.

Analysis: Polling criticism unfounded

Just a few weeks earlier, Romney’s campaign dismissed the impact of public polling data showing Obama building on leads.

“Polls go up and down, but frankly you’re going to see the support that I need to become president on Election Day,” Romney told CNN a week before the debate.

Those on the sidelines are more influenced by the subsequent deluge of media coverage, said Michele Swers, a political science professor at Georgetown University.

“There’s a general mood that’s created,” Swers said. “Right now there’s more coverage of these polls regarding the debate than the September jobs numbers.”

Placing such a heavy emphasis on polls, she said, has a general impact on the mood of the voter.

Polls may make for great stories, but voters should be wary of the hype. It’s important to be a sophisticated consumer of polls and ask “what does that number mean,” said Marc Hetherington, a political science professor at Vanderbilt University

“If you were able to poll voters on what they think about polls they’d think they’re crazy,” Hetherington, said. “The idea that 1,000-1500 people can make a prediction on what people think in the country is crazy.”

Jim Acosta, Dan Lothian and Paul Steinhauser