But candidates take note as well: the new methods get results. As public polls proved to be miles off in the wake of New York City’s mayoral election, the race’s private pollsters—who had the race pegged—showed that new hurdles like cell phones and low voter turnout are not insurmountable.
Bernard Whitman, who polled for the Bloomberg campaign, gathers data by hiring the services of an online consumer survey company. Respondents have agreed to answer several surveys about their shopping habits in exchange for frequent-flyer miles, coupons for cash, or other rewards.
Voters, meanwhile, who live in a certain district, or fit a certain socio-economic pattern, or are prime voters, are slipped a political poll that asks decidedly non-political questions.
“We ask them a whole range of questions about their interests, their activities, their associations, their attitudes, as well as demographics,” explained Whitman (who spoke generally, declining to explain his polling for Bloomberg). “What we’ve been finding is that the differences that are surfaced via behavior are oftentimes much more significant than those differences that simply show up demographically.”
The online services company monitors the surveys against traditional telephone polling to make sure sample sizes are accurate.
Whitman said that new methods such as online polling are necessary since fewer people have landlines than ever before, and few reached on cell phones are likely to acquiesce to a poll.
“What we find is that with telephone interviews, cooperation rates are going down. This is due to both caller ID and mobile-only households,” he said. “At the end of the day, you have to ask people.”
The rise of the Internet has created new communities based around common interests and values that supersede some of the old polling data like race and income level.
“We’ve thrown out traditional demographic indicators,” said Whitman. “What I think is the really exciting new trend is using micro-targeting to identify new groups of voters that behave in similar ways.”
Still, the new methods are not without their disadvantages. They are expensive, for one thing, and for another, pollsters are unable to target voters in quite the same way as telephone polls can.
“The challenge is the fact that we typically, in politics, rely on voter lists and look at prime voters,” said Whitman. “The difficulty of taking a voter list and matching it to online sampling, I think, is something that’s the Holy Grail.”
As cutting-edge pollsters have learned to navigate masses of online data, their methods are trickling down to smaller campaigns. Micro-targeting voters with the help of consumer information has become cheaper and more common in lower-profile races, according to campaign veterans.
“Even at the local levels—municipal, local, state—it’s getting more and more sophisticated,” said Kyle Kotary, a Democratic political consultant. “Literally, I know how much coffee you drink, how much you use a credit card.”
Cash-strapped candidates who cannot afford online polls need not completely fret, however, if they know how to read the public polls that research institutions put out regularly.
Steven Greenberg of the Siena Research Institute said that public polls can be used more effectively by paying closer attention to questions on voters’ ideologies and pet issues, rather than which name they expect to check at the ballot box. He notices that each month when the Siena polls are released, the more substantive questions are largely ignored in the media.
“I think to many in the public, the analysis of the issues, where their neighbors and fellow citizens are on a particular issue, can be even more meaningful and more important than the horse-race three or six years in advance,” said Greenberg.
