One subject that may have gotten short shrift in my story on the Republican polling reboot in National Journal magazine (available for National Journal members) is the future of non-telephone survey modes, particularly internet polling, which is growing in use and popularity. Most of the pollsters I spoke to emphasized that they were still focused on phone polling, but some are at least beginning to experiment with web and mobile surveys.
Republican pollster Alex Lundry, who works at the Alexandria, Va., firm TargetPoint Consulting, thinks the GOP polling reinvention only scratches the surface by ignoring "emerging modes." "We are optimizing a failing but functioning system" in telephone polling, Lundry said. "We've also got to be putting resources into optimizing an emerging and innovative system."
Lundry, who led Romney's data team in 2012, said the lack of emphasis on internet polling makes the project shortsighted. "We can't neglect the online side as well," particularly for ad testing, he said. Doing a survey online allows a campaign to show its advertising -- whether television, radio or mailer -- to voters and capture their immediate reactions. But too few campaigns are utilizing this technology.
"Even among those campaigns that are moving ad testing online, the numbers aren't that large, and they're not uniform," Lundry added. "You lose the opportunities to compare from ad-to-ad."
But the recommendations proffered by the National Republican Congressional Committee and Republican National Committee (Lundry participated in the RNC's Growth and Opportunity Project, he said) focus exclusively on phone polling. "These discussions and these things will continue going forward. The internet-slash-kinda digital aspects of the surveys is kinda an exciting direction that some of these things are going," said NRCC political director Rob Simms.
"I think for us, in some districts, that's not going to work, practically speaking. In some other districts, it could be very useful," Simms added.
Any appeal of web-based surveys in the near future to GOP campaign pollsters might lie in it as a solution to the cell phone problem -- the nearly 2-in-5 adult Americans who live in a household without a landline phone. Since cell phones are more expensive to call (numbers must be dialed by a human being, not a computer), polling these respondents over the internetis a possible solution.
The future is "coming sooner than people realize," said GOP pollster John McLaughlin. "It's going to be more important to have [a voter's] email address or Facebook page than it's going to be to have their cell phone number."
Other pollsters are more skeptical of doing mixed-mode research. Public Opinion Strategies, in Alexandria, Va., is one of the largest firms on the GOP side, but it was their work for NBC News and the Wall Street Journal in producing their regular, bipartisan survey that allowed them to study this more closely. They, along with the Democratic firm Hart Research Associates, conducted 5,000 interviews with cell-phone-only respondents in 2012 as part of their NBC News/Wall Street Journal polling. "And so we have a pretty good profile of cell-phone-only respondents, in terms of who they are demographically and their attitudes," said POS's Bill McInturff, at an last week here in Washington on the future of polling, sponsored by the media firm Kantar.
McInturff and POS then conducted large panels on internet and mobile devices with cell-phone-only respondents to see how the people reached there compared with those actually reached by dialing a cell phone.
"They're interesting because they're much better educated than phone respondents, a lot less Latino, and, on most attitudes, they were the same, except for gay marriage and abortion, where the internet and the mobile cell-phone-only respondents were much, much more liberal than the phone folks," he said.
That made him doubt the validity of doing mixed-mode polls. "And so this notion that there's going to be a world where you can combine methodologies, where you're going to combine phone, internet or mobile, maybe that world will come," McInturff said. "But as I look at this data in 2013, I don't think that you can simply take phone landline and somehow combine it with a different methodology and create one unified survey."
"We are doing tons on the internet, we are ... starting to do substantial stuff on mobile," he added. "I see them as different products with different objectives. I don't see how, today, from the work we've done in 2013, they oughta be blunted into one survey response."
The biggest impediment to using the internet for campaign polls, McInturff said, is that there aren't enough potential respondents in your sampling frame -- that is, the group of people who've volunteered to participate in internet polls -- to survey at the congressional district-level. That means that the phone survey is going to remain the dominant mode for his firm and their competitors, even if other segments of the polling industry migrate to internet or mobile.
"There simply are not large enough cell sizes in the internet panels, mobile panels and others to do [congressional district] or [state legislative] work," said McInturff. "The political pollsters will be the last, last, last people on the phones."