Are people who only use cell phones hurting the accuracy of polling results? The number of people who only use cell phones now -- known as
"cord-cutters" in the telephone industry -- has risen dramatically since 2000. And polling stations can only randomly call traditional land-line
telephones for people's opinions.
Cord-Cutters Spell Trouble For Pollsters
Knight Ridder News
"PHILADELPHIA - Jennifer Malivuk says she'll probably vote next week.
A Pittsburgh native, the University of Pennsylvania senior is undecided but is leaning toward President Bush. Pollsters would surely be interested in
the opinion of someone like Malivuk, an undecided voter in a swing state. But there is no chance they will ever talk to her. She's a cord-cutter.
That's the telephone-industry term for as many as 6 percent of Americans who have a cellular phone but no traditional phone with a cord. Pollsters
get their voter samples by using computers to dial random telephone numbers. But it's illegal to call cell phones with these auto-dialers because it
costs people money to answer a cell phone. So pollsters call "regular" phones, and the politics of cell-only folks remain unknown.
That is just one of many uncertainties in polling, a field that practitioners concede has elements of art as well as science. And after two national
polls differed by 13 points last month, polling is getting more criticism than usual in a nation where some are still bitter over the disputed 2000
"Hostility to pollsters crosses party lines," quipped Michael Hagen, head of a new poll sponsored by Temple University and the Philadelphia
Hagen, a political-science professor at Temple, said much of the criticism was misguided. The nonprofit National Council on Public Polls found that in
2000, for example, the election-eve predictions of 10 major polls were among the most accurate since 1936. Yet he acknowledged that pollsters can
disagree, both in their results and their method.
Charles Golvin, a principal analyst at the research firm Forrester Research Inc. in Cambridge, Mass., said cell-only people were more likely to be
from urban areas than rural, where cellular coverage can be spotty.
That might suggest they are more likely to vote for Democrats, but much more study is needed, said Cliff Zukin, professor of public policy at Rutgers
University. The cell-only issue might not matter too much in polls because these people are young, and therefore less likely to vote, he said.
"Polls might miss them, but also they might miss the election," said Zukin, president-elect of the American Association for Public Opinion Research.
A broader reason for polling disagreements is that people are fickle. They change their minds about whether they are going to vote, and they give
different responses based on when a poll is taken, how the questions are worded, and in what order they are asked.
One factor that may appease poll critics in the waning days of the campaign: The fewer days until the election, the more accurate polls tend to be.
That's because by that point, people who say they are likely to vote probably will."