The plague of polling

June 25, 1996|By Daniel S. Greenberg

WASHINGTON -- Now is the time for all who value the good health of the American political system to stand against the quadrennial plague of presidential election polling.

The polling industry and its political clients won't like it, nor will the news media, which treat the voodoo output of polling like sacred text. But the body politic will be better off if citizens thwart the pollsters. It's not hard to do.

Encounters between voters and pollsters are still relatively rare, but are becoming more common as the infestation of polling moves deeper into the electoral system. Thus, it is wise to look behind the claims of polling as a science and recognize that these supposedly rigorous inquiries into the opinions of the electorate are inexact, often wrong and damaging to the electoral process.

No useful purpose

First, it is useful, as well as amusing, to contemplate the frequent mismatch between poll-based forecasts and election-day outcomes.

The broad extent of the GOP sweep of the 1994 congressional elections caught the pollsters off guard. The 1992 parliamentary election in Britain may have set a mark for pollster ineptitude. "Not a single opinion poll published on election day came even close to predicting a Tory victory," the Economist observed. The heavily polled Nicaraguan presidential election in 1990 inspired a New York Times column that asked, "Why were the polls in Nicaragua so far off?"

Yes, the major polls were accurate in forecasting the outcome of the Bush-Clinton election, but the overall record of polling is so spotty that the achievement can be credited to the principle that a stopped clock is right twice a day.

The problems of polling, however, extend far beyond the customary game of forecasts, votes and pollster alibis.

One trouble with polling is that the results are easily confused with political reality, producing bandwagon effects, heartening the leaders and disheartening the laggards. Polls can make it seem like it's all over long before election day, though the historic role of campaigns has been to educate the voters about candidates and issues so they can arrive at a decision on election day.

Yet another liability of polling is its use by candidates to shape a saleable personality based on findings about the voters' anxieties and fears. Finger-to-the-wind politics represents a triumph of the polling technique, as essential to modern politicians as radar is to aviation.

'Push polling'

But even more worrisome is the increase in the camouflaged use of polling as a partisan tool to warp public opinion, rather than merely to measure it. Known as "push polling," this technique subtly flavors the questions with innuendoes, asking, for example, whether a particular candidate's reported moral turpitude should be considered a disqualification for public office. In the guise of objective inquiry, political poison is thus spread about. The growing popularity of push polling has created alarm in the ranks of pollsters who subscribe to ethical standards of behavior. But there's no sure way of stopping it, despite proclamations of codes of behavior for the polling industry.

There's no need for citizens to stand by helplessly while the pollsters continue their rampage. Though polling has fastened itself on the political process and is treated with reverence by the news media, it is a commercial business and has no constitutional claims on participation in the electoral process.

To succeed, polling needs cooperation from the polled -- and that's where the citizenry can assert its rights. In a nationwide poll on the presidential election, about 1,200 randomly selected persons will be questioned to obtain a statistical portrait of the voting public -- which numbered 103 million in 1992.

Refusals to reply raise worrisome doubts about election-day intentions. Among those who reply, each of the 1,200 polled stands in for approximately 85,000 voters, giving the pollsters recurring nightmares about the large errors that can result from small events -- such as a fib about voting intentions. Get it?

Daniel S. Greenberg is editor and publisher of Science & Government Report.

Pub Date: 6/25/96

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.