Published on Sunday, February 6, 2000, in the San Jose Mercury News
Increasing Number Of Hang-Ups Has Telephone Pollsters Worried
by Lorrie Lessner
WASHINGTON -- The phone rings with a kind of vengeance that's especially irritating during dinner. This time it's not a telemarketer peddling mutual funds, but a political pollster asking about this year's crop of candidates.
Increasingly, though, the public's response is the same.
More and more Americans are treating pollsters the same way they treat callers who try to sell them something. They use answering machines and caller ID to dodge the calls, and when all else fails, they slam down their receivers.
At a time when America's newspapers, computers and airwaves are filled with two or three political polls a day, the trend is beginning to alarm statisticians and experts. High refusal rates, as they're called, raise new questions about the validity of surveys that use responses from 1,000 or fewer voters to conclude, for instance, that George W. Bush is ahead of Al Gore.
The first question is: Who's hanging up and who's answering? Are higher income, better educated voters -- the preferred targets of telemarketers -- more likely to slam down the phone? What about voters for whom English is a second language? Older vs. younger citizens? Any biases could skew pollsters' samples.
No one knows to what extent the lack of participants threatens a poll's credibility and usefulness, if at all, said Warren Mitofsky, who started the CBS News/New York Times poll.
But members of the polling industry are troubled nonetheless, and some convened an international conference last fall to discuss the problem.
``The trust factor is waning,'' said Jane Sheppard of the Council for Marketing and Opinion Research. ``When the caller says, `Hello my name is . . .,' right away people think they are being hounded by someone who is going to sell them something. They're not giving it enough time.''
Political pollsters contend they can take a reasonably accurate measure of America's attitudes. It's just more costly, because they have to make more calls to reach a representative sample. Ed Goeas, president of a top GOP firm called the Tarrance Group, said his firm needs to make 15,000 calls over three nights to reach 1,000 people.
``. . . I'm not overly concerned,'' he said. ``If you can tell people quickly that you are calling for political opinion, you find out they're very willing to tell you what they think.''
Conservative columnist Arianna Huffington is waging a crusade to change that. She maintains that candidates use polls as a way of sticking their fingers in the wind to see what the public wants to hear, and is urging people to hang up on pollsters. So far, she said, nearly 10,000 people have signed pledges on her Web site (www.ariannaonline.com) to do just that.
``Politicians don't lead anymore,'' Huffington said. ``But if enough people hang up on pollsters, poll results will be useless and leaders will be forced to lead.''
Goeas and Neil Newhouse, a Republican pollster with Public Opinion Strategies, say a legitimate political pollster's goal is to test a policy message, not dictate it or even develop it.
``I have never told a candidate to change a position,'' Goeas said. ``I have told a candidate, `Out of these 17 issues in the polls, these are your three strongest to talk about, not those four.' ''
As a campaign's intelligence officers, pollsters must gauge how favorably their candidate and challenger are perceived by the public. They conduct polls because it's quicker and easier than calling all 200 million Americans of voting age.
Yet it's an inexact science at best.
But as the industry is getting worried about finding people to participate in polls, a national survey of 800 registered voters shows that nearly three-fourths of the respondents had been called before by a media or political pollster or someone doing a product survey.
``It surprised the hell out of me,'' said Newhouse, who conducted the survey.
To Robert Groves, senior research scientist at the University of Michigan, the most significant note in polling history occurred in the last 15 years, when independent pollsters were replaced by partisan ones.
``Information is now being controlled by a Democratic pollster or a Republican pollster,'' Groves said, ``and each has a desired outcome in mind for the poll.''
### © 2000 Mercury Center.
© 2000 Mercury Center.