Pollsters and their hang-ups

11/7/2000

It's 9 p.m. The telephone rings. It's somebody trying to sell you something. You're enraged. “At 9 o'clock at night?!” you exclaim, immediately hanging up. “I don't think so.”

While more and more Americans are giving an unceremonious kiss-off to telephone solicitors, there is some thought that the trend may be having a complicating effect on political horse-race polls taken by telephone.

According to the Washington Post, only 35 per cent of people reached via phone will answer pollsters' questions, down from 65 per cent 15 years ago. With all the refusals, answering machines, and disconnected or nonworking numbers, it takes almost 6,000 calls to generate 400 usable responses.

From there, the raw numbers are massaged, or “weighted” to get what the pollster feels is a statistically valid sample of the population and who among them will actually turn out to vote. The numbers also are adjusted for a variety of factors, including gender, race, religion, age, and income. Sometimes, the weighted and adjusted results reverse the findings of the raw numbers. In any case, the outcome is an educated guess.

“Twenty per cent of this business is art, 80 per cent is science,” is the way national pollster John Zogby describes it.

On this peculiar brand of modern alchemy rest the hopes and aspirations of virtually every politician running for public office, from school board to president of the United States, plus news organizations plumbing the future. Who can resist the desire to know what will happen before it actually does?

The difficulty pollsters have in finding people to poll does raise the question of whether an indifferent electorate - one weary of being called at all hours of the day and night - leads the political seers to inject a little more art than science into the numbers than they might have in another era. The fun of it all is that we never know until after the only poll that really counts, the one the voters are conducting today.