ShareThisBy RYAN KIM
San Francisco Chronicle
Tuesday, June 05, 2007
In its biggest challenge in 30 years, the survey research and public opinion industry is cranking up its efforts to account for the growing number of respondents who use cell phones exclusively these days.
About 13 percent of all households in the country have "cut the telephone cord" in favor of cell phones, according to federal figures released last month. That puts this group out of reach of traditional surveys that rely heavily on calls to standard landline phones.
To remedy the situation, surveyors are trying to reach this demographic segment by turning to cell phone surveys and online polls and reworking the survey parameters that have served them for years. There is even talk of returning to more traditional methods like mailing questionnaires and visiting respondents door to door.
"This is probably the biggest change for the industry since when I came into the Field Poll in 1978, when the industry was changing over from door-to-door surveys to telephone surveys," said Mark DiCamillo, director of the Field Poll, California's oldest and largest continuous independent poll. "This is a new challenge for us. We're being forced to come up with alternative methods to now include this large segment of the public in our surveys."
The most dramatic move has been the adoption of cell phone surveys. Arbitron, which rates radio stations, began contacting people with cell phones to fill out its listening diaries earlier this year. The Field Poll recently started a new methodology that includes cell phone numbers. And UCLA's annual California Public Health Survey this year will now reach out to 800 cell phone users.
The shift to wireless-only households is worrisome for the survey industry because researchers fear that this group, if uncounted, could invalidate or undermine their data.
Researchers have a rough idea who's in this group from early demographic and behavioral data. Basically, they are mostly young, less financially well-off and single. It also includes a larger proportion of minorities and tends to be more progressive in thinking than the general populace.
According to the National Health Interview Survey, released last month by the Centers for Disease Control and Prevention, 29 percent of respondents age 25 to 29 used only cell phones, and 25 percent of people 18 to 24. More than half lived in a household with unrelated adults with no children, and this group of cell phone-only households is growing, rising from 4.2 percent of all households in 2003 to 12.8 percent last year.
"We see no sign that this trend for increased prevalence of wireless-only is going down or leveling off," said Stephen Blumberg, author of the CDC report and senior scientist at the National Center for Health Statistics. "If the trend continues, we could see a third or half of households being wireless-only in three to five years."
Researchers have already noticed some slight voting tendencies in this group toward Democratic candidates and liberal causes that could intensify in the coming years -- perhaps as early as this coming presidential primary season.
Failure to account for the cell-only group could affect the forecasts of pollsters.
A study by the Pew Research Center found that slightly more cell phone users voted for John Kerry than George Bush in the 2004 election. A separate study last year found that 53 percent of cell phone users lean toward or identify with the Democratic Party compared with 30 percent that favor the Republican Party.
In both cases, the results had yet to affect the overall reliability of the polls that didn't include cell phone numbers because response of the cell phone-only group wasn't radically different from other respondents in their same age bracket. Another factor has been the lower voter turnout for cell phone-only users.
Still, there is a growing concern the cell-only group may be harder to predict soon, said J. Ann Selzer, president of Selzer & Co., which conducts the Iowa Poll for the Des Moines Register.
"The worry is that there is something that distinguishes this group, and by not including them we're missing a dynamic of the caucus," said Selzer, who is not conducting cell phone interviews. "I have some data that says they are not that different enough to skew my numbers today, but I still worry about it every day."
The adjustments by some companies to reach this group _ particularly the move to cell phone surveys _ is stirring new issues and problems for the survey industry. For instance, many surveyors are opting to reimburse respondents at least $5 to pay for used wireless minutes, a practice that encroaches on traditional notions of avoiding payments to participants.
Researchers are also finding that cell phone users are less cooperative because they believe these calls are intruding on their privacy. Mike Brick, vice president and research firm director for Westat, which conducts the UCLA health study, said a pilot cell phone test last year found that cell phone-only response rates were a little above 20 percent, about 12 points lower than calls to landlines.
"Some cell phones are not always available. You have prepaid ones that aren't always paid, and some people _ they see your number and they block it," Brick said.
