Political Pollsters Struggle to Get the Right (Cell) Number
As they gauge voter sentiment in this tight presidential race, pollsters face a big challenge: more and more voters hang up on them.
So it sounds odd that some pollsters have decided to hang up on more voters.
Yet that is one way survey researchers have adapted to the communications revolution that has upended old methods of measuring which political party is ahead. In the polarized battle between President Obama and Mitt Romney, arcane shifts in polling techniques can have important consequences for the results — and public perceptions of the contest.
Bill McInturff, a Republican pollster, and Peter Hart, his Democratic counterpart, who conduct the NBC News/Wall Street Journal poll, proved the point in their latest poll, conducted July 18-22, when they increased the proportion of respondents who rely exclusively on cellphones to 30 percent from 25 percent. To home in on them, the pollsters ended calls answered on cellphones if the respondents said they also had landlines.
Their findings affirmed arguments that “cell only” Americans have significantly different, and more Democratic, political views than those with landlines. Over all, the poll showed Mr. Obama leading Mr. Romney by 49 percent to 43 percent — providing a confidence-boosting talking point for Democrats and provoking sharp criticism from Republicans.
Scott Rasmussen, who owns an independent polling firm, approaches the “cell only” problem differently, as he must by law. His Rasmussen Reports conducts surveys through automated dialing, which under Federal Communications Commission rules is permitted for landlines but not cellphones.
So in Mr. Rasmussen’s polls, online interviews account for 15 percent to 20 percent of each survey, which he figures helps him reach the same kinds of voters, especially younger ones, in the “cell only” category. The result he reported the morning of July 25, a few hours after the NBC News/Wall Street Journal poll was released, was strikingly different: Mr. Romney had 47 percent, and Mr. Obama 44 percent.
“Nobody has the answers,” Mr. Rasmussen said of different approaches to the issue. “We’re all experimenting with the same thing. How do you reach people in a way they communicate?”
Political pollsters have long struggled to keep pace with changes in how Americans communicate with one another — or don’t. One of their central challenges is the decline in “response rates” among voters.
Another is ensuring that their means of reaching people produce accurate reflections of voter sentiment. The rapid rise of cellphones as the sole means of telephone communication for many Americans — now about one-third of the population, government researchers say — has produced a distinct subgroup within the electorate.
That group is not only younger but also “attitudinally different from other people” of all ages, said Mark S. Mellman, a pollster for Democratic candidates. Among their characteristics: they are disproportionally urban, African-American, on either the high or low ends of the economic ladder — and Democratic.
“It used to be if you got the number of young people right, you got the poll right,” Mr. Mellman said. “That’s no longer true.”
But accounting for those “cell only” voters is not so easy. Regulations limit how pollsters can dial cellphone numbers, making those calls roughly twice as expensive as landline calls — a cost some pollsters and their customers are not willing to bear.
Moreover, there is no consensus on the right method for handling such polling. The ABC News/Washington Post poll, like the NBC News/Wall Street Journal poll, terminates calls if cellphone respondents say they also have landlines.
The New York Times/CBS News Poll, like the Gallup Organization and Pew Research Center, does not. Instead, Times/CBS pollsters complete interviews with all willing cellphone respondents and “weight” the views of those without landlines to make them reflect one-third of the survey’s results.
Others remain skeptical that one-third is even the right target for “cell only” voters.