Cell phones throw off poll takers
By Abdon M. Pallasch Political Reporter firstname.lastname@example.org June 26, 2011 6:04PM
Democratic presidential hopeful, Sen. Barack Obama D-Ill., addresses the crowd at a 'Meet The Candidate' event Friday, Dec. 21 2007 in Washington, Iowa. (AP Photo/M. Spencer Green)
Updated: September 30, 2011 12:24AM
How do you accurately poll voters in an age when increasing numbers of them own only a cell phone?
“I worry about this more than anything else — you can have a California number and still be living in Iowa,” said Ann Selzer, the only pollster to predict Barack Obama would win the Iowa Caucuses by 7 percentage points. (He won by eight.)
This will complicate tracking not only the horse-race among Republican presidential candidates trudging through Iowa county fairs this summer, but also will challenge politicians of both parties trying to navigate Illinois’ newly redrawn congressional districts.
Two weeks ago, a poll of voters in the new Joliet-Aurora-based 11th Congressional District found former Rep. Bill Foster leading potential candidate John Atkinson 36 percent to 6 percent. The poll helped persuade Atkinson to drop his potential candidacy.
About 40 percent of Hispanic households have no landline — just a cell phone. About a quarter of white families have only a cell phone.
Polling only landlines means you get lots of older, white women. You disproportionately miss young voters and minorities.
That’s part of the reason all the polls before last year’s election for governor of Illinois had Republican Bill Brady beating Democrat Pat Quinn by five percentage points or more, said Quinn’s pollster, Mark Mellman.
“It’s the failure to reach cell phone people and the focus on people easier to reach,” Mellman said. “Every poll that I saw had Gov. Quinn losing. We had him winning very narrowly.”
Mellman was simultaneously assuring Quinn and U.S. Senate majority leader Harry Reid of Nevada that they were going to win even though the public polls showed them losing.
One of the ironies of the current state of polling is that candidates often splurge for the more expensive and more accurate telephone interview polls while cash-strapped news organization often opt for the cheaper “robocall” polls.
With robocall polls, it’s not always clear who in the house is answering the phone.
All polls put the greatest weight on respondents who identify themselves as “likely voters.”
The key, Mellman said, is to remember that some “unlikely voters” also will make it to the polls.
While likely voters favored Tea Party favorite Sharron Angle over Reid, Mellman predicted about a fifth of the electorate would be unexpected voters, who put Reid over the top.
Likewise in Illinois, more casual voters who may not have cast ballots in every previous election helped give Quinn his razor-thin victory over Brady, as Mellman forecast.
In Iowa 3 1/2 years ago, Selzer awoke to a storm of controversy when her poll for the Des Moines Register showed newcomer Obama seven points ahead of the more seasoned John Edwards and Hillary Clinton campaigns, which led most of the polls.
“There was kind of this explosion,” Selzer said. “. . . and an e-mail blitz from the Clinton campaign and the Edwards campaign saying I should be fired and the poll should be ignored.”
Selzer’s controversial poll forecast an unheard of 59 percent of new attendees at the caucuses — voters inspired by Obama who had never caucused before inspired by Obama to make their maiden voyage to caucus.
“I had a call from one Democratic precinct captain who said, ‘I’ve always trusted your polls until now. But I’ve knocked on 99 doors and I don’t find this lurking Obama support you’re talking about,’ ” Selzer said. “I said, ‘Tell me about the doors you have knocked on.’ He said, ‘Registered Democrats and former caucus attenders.’ ”
I said, ‘Well, you’re not going to find the lurking Obama support there.’ [Other pollsters] were cutting that corner — they weren’t talking to independents.”
The challenge for all pollsters is to track down cell phone users. Selzer noted that some people port their old landline numbers to their cell phones so she gets some that way.
Gallup now includes 40 percent cell phones in all their national polls.
But in smaller races for, say, a congressional district, pollsters mostly have to guess at cell phone users’ views by extrapolating from younger, minority voters they reach at home.
Years ago pollsters used to have to knock on doors, and they may have to start doing that again.
Cheaper robocall polls always miss cell phone users because laws prohibit using computers to autodial cell phone numbers.
With Obama facing no primary challenge this time around, Selzer will be polling only Republicans and independents.
She’s not too worried about Democrats infiltrating Republican caucuses to push weak Republicans. “It would be hard to think they’re going to go caucus with [Minnesota Rep.] Michele Bachmann,” Selzer said of the Iowa-born Tea Party favorite. “If Sarah Palin were in, they’d definitely go in and do that.”
Many of the more conservative candidates think they can repeat Mike Huckabee’s success last time, but Selzer rejects the characterization of Iowa’s Republican caucuses as “Camp Christian” and predicts a moderate could win.
According to the Des Moines Register’s tracker, the candidates who’ve held the most events in Iowa so far are former U.S Sen. Rick Santorum of Pennsylvania (55 events); former Minnesota Gov. Tim Pawlenty (53), and former House Speaker Newt Gingrich of Georgia (43.)
Selzer and Mellman were just a few of many of the nation’s top pollsters giving their assessment of the state of polling at an American University conference in Washington, D.C., last weekend.