Pollsters Need Cell Phone Users More Than Ever

Your cell phone rings, an unfamiliar number shows up and heck if you’re going to answer a call from a stranger — especially if you're busy filling the cart at the grocery store.

Or driving to the movies. Or trying to get the kids to finish dinner.

That reality — a mobile nation increasingly reliant on cell phones and less likely to answer calls from unrecognized numbers — is upending the way polling firms survey Americans on everything from political contests to gun ownership.

A man on his cell phone in Los Angeles last month. i

Pollsters are failing to consistently reach cell phone-only users. Robyn Beck/AFP/Getty Images hide caption

itoggle caption Robyn Beck/AFP/Getty Images
A man on his cell phone in Los Angeles last month.

Pollsters are failing to consistently reach cell phone-only users.

Robyn Beck/AFP/Getty Images

Gone are the good old days (as recently as the 2008 presidential campaign), when reaching Mr. and Mrs. John Q. Public at home on a landline could still produce a credible survey.

And failing to consistently reach cell phone-only users, who are disproportionately younger and more diverse than those reliant on landlines, is now officially producing bias in surveys, says Scott Keeter of the Pew Research Center. "The issue of capturing the opinions of cell phone users in surveys dates back to at least 2003, Keeter says, when only about 7 percent of Americans were estimated to be using cell phones exclusively.

"Now we're at three times that number," he says.

In a potentially pivotal year when control of Congress is at stake in the midterm elections, getting an accurate read on public opinion is crucial for the candidates, the news media and the public — and that means getting to those who have ditched their landlines is far more important than it was even just two years ago.

Skewed Results

In fact, Pew researchers have found that recent surveys of party preference in the upcoming midterm elections skewed Republican 47-41 percent over Democrats when only landline responses were recorded.

When cell phone responses were combined with landline samples, the parties were tied at 44 percent each in the generic preference poll.

It's a reality that Des Moines, Iowa-based pollster J. Ann Selzer says reputable surveyors no longer have the luxury of ignoring.

"Before, if you were doing political work, you were missing people but not enough to meaningfully change your findings," Selzer says. "But that's changed, and the number of cell phone-only users is only increasing."

That increase means that the polling science of weighting results — extrapolating results from official data known to be true — to account for hard-to-reach cell phone users is no longer defensible.

"You just can't weight away the problem," says Kristen Soltis of the Winston Group, a research organization based in Washington, D.C. "Not when you have about a fourth of households without landlines, with the growing number of African-Americans, Hispanics, and young folks using cell phones."

Cell Phone Imperative

If left unaddressed, the bias inherent in failing to reach a sufficient number of cell phone users will undermine the credibility of pollsters, says Keeter, an author of Pew's recent report, Assessing the Cell Phone Challenge.

A landline, push-button phone, circa 1964. i

Remember this? AT&T Archives and History Center/AP hide caption

itoggle caption AT&T Archives and History Center/AP
A landline, push-button phone, circa 1964.

Remember this?

AT&T Archives and History Center/AP

And it will allow Americans' understanding of themselves and the world around them to be disproportionately shaped by a population that's older, whiter and more conservative.

Pew found that adults age 50 and older are "significantly overrepresented" in landline surveys, comprising 66 percent of an average sample, when, as a percentage of the overall population, they should be 40 percent.

"The cell phone-only population tends to be slightly more liberal on all kinds of issues, from support for health care reform, to presidential approval and party affiliation," Keeter says.

Expensive Solution

So, how to fix a problem that, on its face, appears almost intractable? How to reach those elusive cell phone users?

The solution is expensive, Soltis says.

Federal telecommunications law prohibits surveyors from using automated calls to reach cell phone numbers. The 1991 law's intent was to protect cell phone users from paying for unwanted calls.

"You have to manually dial cell phone numbers," Soltis says. That, and the time it takes to screen out cell phone users who may be underage or living in a geographic area different than their number indicates, adds to the cost. Keeter says that upward of 40 percent of the people Pew surveyors reach by cell phone are underage. Surveyors also have to make sure people reached on cell phones are in a position where it's safe to talk — that usually means ensuring they're not behind the wheel of a moving vehicle.

Capturing cell phone users can dramatically increase the cost of surveys, pollsters say. For her polls, Selzer conducts two separate surveys — one landline and one cell phone. The cost of the cell phone portion "is easily double, sometimes triple, depending on the market," she says. The extra expenditures come on top of other costs that have also been on the rise in recent years — most notably the time and money it takes to get people on the phone in an age when many have caller ID and just don't pick up (pollsters swear that the "response rate" problem does not affect the accuracy of their results).

That cost has led to the emergence of cheaper options that may produce limited results, including landline-only "robo-call" polls where computerized voices ask the questions (not live operators) and controversial online surveys. And there have been cutbacks in traditional phone surveys.

"The cost is a significant burden on all of us," Keeter says.

The silver lining? "One good thing about the cell phone revolution is that there is almost universal telephone coverage in the United States," he says.

But the issue of reaching cell phone users will play out in a big way in the coming months in the ramp up to the midterm elections.

"For better or worse, elections are the benchmarks used to give polling credibility," Keeter says. "Polling by phone is increasingly difficult and expensive to do well, but it's still feasible and likely will remain so for many years to come."

Some are looking for easier and cheaper ways to take surveys. And today's challenges represent an interesting parallel to the 1980s, when pollsters who had relied almost solely on face-to-face interviews began switching over to less costly telephone surveys. But telephone surveys aren't going to become extinct any time soon, he says.

Comments

 

Please keep your community civil. All comments must follow the NPR.org Community rules and terms of use, and will be moderated prior to posting. NPR reserves the right to use the comments we receive, in whole or in part, and to use the commenter's name and location, in any medium. See also the Terms of Use, Privacy Policy and Community FAQ.