1022elnpollsaudMatt Towery, CEO of Insider Advantage, talks about the reliability of political polls.
If you believe the polls, it looks like one presidential candidate has moved ahead of the other and might be heading toward a victory on Nov. 4.
But can you believe the polls?
Nobody wants to repeat the mistake of 1948, when voter surveys indicated Thomas Dewey easily would defeat incumbent President Truman. The media fell for it, leading to one of the most embarrassing headlines in newspaper history.
But Rich Engstrom, assistant professor of political science at Georgia State University, said such a blunder wouldn’t happen today.
"(Polling) is really a science now," he said. "We can trust polls, but they always have to be taken with a grain of salt."
Maybe with more than a few grains of salt this year. The presidential race between Republican John McCain and Democrat Barack Obama has been so unpredictable that there is no historical precedent.
There will be more registered voters this year than ever before, and some will belong to demographic groups that haven’t voted regularly in the past. Pollsters may interview these people about their opinions, but they can’t predict how many actually will show up on Nov. 4 to vote.
And because this group trends toward younger people, some have questioned whether the views of these new voters are reflected accurately in the polls.
The reason: At least 15 percent of American adults have a cell phone but no landline phone. Traditionally, pollsters only have called people who have landline phones, so their survey sample may have skewed toward older, more affluent voters.
According to the Pew Research Center, people who have only a cell phone are more likely to be young, male, nonwhite and lean toward voting Democratic.
That sounds like a group that would favor Obama. But a Pew study last summer compared cell-only voters with a "weighted" sample of landline voters and found no significant difference.
Engstrom said pollsters weight the data by making adjustments to correct any potential biases. But these assumptions are based on factors that have been learned from previous elections.
"All bets are off if, all of a sudden, a whole new demographic group shows up to vote that never voted before," Engstrom said.
Henry "Chip" Carey, associate professor of political science at Georgia State, said the concern about cell phones has been overblown.
If the surveys miss some cell phone users, he said, it won’t necessarily lead to inaccurate predictions of election results.
"Most kids have cell phones," Carey said. "But kids are less likely to vote."
All this hand-wringing over cell phones may be a waste of time, according to Matt Towery, chief executive officer of Insider Advantage, a Georgia-based polling firm.
"The biggest urban myth of the year is that we’re not able to poll cell phones," he said. "It’s not true. (Cell phones) have no effect on polling whatsoever."
Towery said cell phones are only a hindrance if you’re relying on the phone book to find phone numbers. Most pollsters don’t do this anymore.
"We can get representative phone numbers from all age groups," he said. "We buy data from companies that collect names from voter registration rolls."
He said it is legal to do this for the purpose of scientific research, including political surveys.
Towery said pollsters do "oversample" young people, meaning they attempt to call a larger pool of individuals.
"But that’s not because of cell phones," he said. "It’s just because it’s difficult to reach them."
He added that most polling firms now use automated surveys, with a computerized voice instead of a live person at the other end of the line asking the questions.
"It’s much more accurate," Towery said.
That’s because people tend to tailor their responses to how they perceive the surveyor. For example, if they think the person asking the questions is African-American, they might answer questions about Obama differently than they would if they thought they were speaking to a white person.
And that brings up another possible variable in the presidential race. In recent weeks, political pundits have been debating whether there will be a "Bradley effect."
This theory goes back to 1982, when former Los Angeles mayor Tom Bradley ran for governor of California. He lost the race despite being ahead in all of the polls.
Political analysts speculated that Bradley, who was black, was thwarted by hidden racism. Supposedly, white voters told pollsters they supported Bradley but then voted for his white opponent.
Today’s political experts question whether the Bradley effect ever really existed. Carey said there certainly is still prejudice among some people, but probably not enough to decide an election.
"There’s also a ‘reverse Bradley effect’ — black people who will vote for Obama just because he’s black," he said, noting that the two biases may simply cancel each other out.
Towery said he doesn’t foresee a Bradley effect this time, because people are much more comfortable talking about racial issues than they were in 1982.
"I’m not buying into the idea that there’s a tremendous ‘lie factor’ now," he said.
So if all the polls are showing Obama leading McCain, can we trust that those numbers reflect reality?
Engstrom said if we were seeing a lot of fluctuation, he would question poll results. But the surveys all seem to be trending the same way.
"It would be very improbable if all the polls pointed in the same direction, and they were all wrong," he said.
Carey said that Obama’s lead has held steady for so many weeks that it would be difficult for McCain to close the gap at this point.
"Barring some unforeseen circumstance, Obama will win," Carey said. "His numbers are too high."