Cell-only households complicating life for pollsters on everything from health surveys to political questions

Cell phones are giving pollsters big headaches. And the pain has nothing to do with the brain-damage scares you read about on the Internet.

This problem relates to the fact that our growing penchant for talking on the run is making our opinions and our behavior far more difficult to track.

In 2004, just 5 percent of American households had given up landlines in favor of cell phones; last year, the count was 17.5 percent and growing.

Minnesota reflects the nation in phone preferences with 17.4 percent of households wireless only, according to new estimates by researchers at the University of Minnesota's School of Public Health. That's the 15th highest in the nation. The highest state is Oklahoma with 26 percent. The lowest is Vermont with 5 percent.

The state-by-state estimates were calculated by researchers at the Public Health School's State Health Access Data Assistance Center (SHADAC). They are the first state breakdowns of national survey information used by the Centers for Disease Control.

Public health officials worry about this trend because they use telephone surveys to track information that helps shape public education and health policies: how many of us lack health insurance, for example, how many smoke, whether our kids are vaccinated and how many of them will go to school with special health needs.

When Minnesota legislators contemplate changes to medical assistance programs, they turn to such survey data for estimates of how many people would be affected by a given change and how much it would cost, said Michael Davern, SHADAC's research director.

In other typical applications, the surveys are used to evaluate the long-term impact of smoking cessation programs and other public health initiatives, he said. If a program isn't making a difference, it's probably not worth the money and effort.

Modeled estimates, 2007

In other typical applications, the surveys are used to evaluate the long-term impact of smoking cessation programs and other public health initiatives, he said. If a program isn't making a difference, it's probably not worth the money and effort.

Inaccurate results could lead to costly inefficiencies in a variety of programs that are geared toward preventing chronic diseases and giving people help where they need it.

Of course, the cell-only trend is a big worry for political pollsters, too. Thanks to statistical weighting techniques, they've managed to adjust their results for the cell-only voters and continue to give usually accurate snapshots of where we stand on candidates and issues.

But as the cell-only crowd grows, so do concerns about bias in the polls, said Lawrence Jacobs of the MPR/Humphrey Institute Poll in Minnesota. He also directs the Center for Study of Politics and Governance at the University of Minnesota's Humphrey Institute.

"I'm not going to sugar-coat this," Jacobs said. "The issue of cell phone use is obviously something of real concern to survey researchers. … But we're not asleep at the switch here. There is a lot of conversation about it, and during the election we were constantly following it."

Cell-only profile
If you leave cell-only households out of a survey, you miss a good share of the young adults, renters, the poor and minorities.

When it comes to health measures, the SHADAC report says that adults in wireless-only households are less likely to be insured and have a regular place they can go for medical care. Meanwhile, they are more likely to smoke and to drink heavily.

Politically, they differ significantly from people who favor the traditional landline, according to a study conducted by the Pew Survey Research Center.

In recent Pew surveys, only about 10 percent of respondents in landline samples were under age 30, which is roughly half of what the U.S. Census suggests it should be.

Just 58 percent of the cell-only people were registered voters, compared with 76 percent of those who used standard landline phones and 74 percent who used a combination of the two phone types.

Candidate preferences differed dramatically, too. Sixty-one percent of the Cell Onlies leaned toward now-President Obama in last year's election, compared with 46 percent of the landline only and 48 percent of the combo phone users.

Why polling cell users is complicated
The differences have driven Pew to revamp its survey processes in order to reach the cell-only households. But that isn't easy or cheap by any means.

Here are a few of the reasons why cell phones are making the work of the pollster more complicated:

• Some households now could be called more than once in a survey. Take mine. We have a landline. My husband and I each have a cell phone. And I have purchased a phone number to use with Skype on my laptop. Indeed, the whole basis for measurement is changed in that landline phones are typically associated with a household while cell phones are associated with a person. "There is currently no agreed upon method to resolve this unit of measurement issue," said the SHADAC report.

• Some states treat surveys as a form of business solicitation and bar survey organizations from calling cell phones. Further, survey organizations typically use computers to generate and dial the phone numbers, connecting an interviewer only after someone picks up the phone. But the Telephone Consumer Protection Act prohibits use of a computer to dial cell phone numbers. Bottom line is that "manual dialing adds to the cost of data collection," the SHADAC report said.

• Think about the safety factor. It's enough of a threat when people discuss dinner plans while driving. We definitely don't need them answering lengthy poll questionnaires. So the interviewer would have to ask safety screening questions.  Are you driving? Riding your bicycle? Pushing a stroller through busy intersections?

• Because we can use cell phones everywhere from cars to crowded bars, pollsters also worry whether they would get the same quality of answers. Would you frankly answer questions about your health and your political preferences while sitting with everyone else in your doctor's waiting room?

• There also are sticky questions about who pays for the call. Many cell service providers charge for the minutes you use on inbound calls. Thus participation rates could suffer.

• Until all of our phones are tethered to the GPS system, no one knows for sure where we are when we take a cell phone call. Thus, in a survey of Minnesota voting preferences, a pollster could dial a number in the 612 area code and get someone who moved to Wisconsin last year.

Sharon Schmickle writes about national and foreign affairs and science. She can be reached at sschmickle [at] minnpost [dot] com.