Washington -- With the conventions over, the pollsters are out in full force, contaminating our sacred electoral process with their schlock-scientific prophecies of how some 90 million Americans will vote on November 3.
Must this scourge be tolerated? There's no legal way to banish it. But, as a practical matter, the polling trade lives in dread of little acts of public non-cooperation that can easily throw its frail statistics embarrassingly out of whack. The public is not helpless against the plague of polling.
Though professing merely to gauge public attitudes and intentions at a specific moment, the pollsters are the indispensable instrument of finger-to-the-wind politics.
Somewhere back in history, election campaigns were for the purpose of revealing the candidates so that voters could evaluate their qualities. Polling, however, turns the process around, revealing the voters so that the candidates can tailor their positions for political salability. George Bush's chameleon history on the abortion issue is a testimonial to the political value of polling.
Polling also defiles politics in other ways. It interferes with the evolution of popular political judgment by depicting as tangible what, in fact, may be extremely fluid. The election campaign is supposed to help voters make up their minds. Poll reports -- even with their cautionary statements -- convey the impression that minds have already been made up. A big lead in the polls can create a bandwagon effect that can translate into campaign contributions to help sustain the big lead. It can also give a valuable aura of inevitability to a candidacy.
Like a noisome party crasher whom no one will oust, polling long ago insinuated itself into politics. It is now so well established in the electoral process that few members of the public seem to recognize that polling is a profit-seeking enterprise feasting on the serious business of choosing leaders.
With their standardized mumbo-jumbo about "scientific samples" and "margins of error," the polling trade conveys an impression of antiseptic, scientific inquiry. Sometimes it scores high levels of accuracy. Often, however, election day is followed by a barrage of explanations for forecasts wildly off the mark. The variability mocks the claim of scientific precision. But even a bull's-eye does not absolve the polls of trespass on the political process.
Voters who want to hit back should understand the statistical dilemma of polling. To be affordable, polling seeks to question the smallest number of persons statistically needed to represent the full electorate. About 1,200 to 1,500 will do, allowing for the customary small margin of error.
But many doubts creep in if the target hangs up -- as increasing numbers are doing, to the great distress of the polling trade. Are the silent subjects evenly distributed on the issues and candidates? Or are those of a particular mind more inclined to silence? These are menacing uncertainties for pollsters, since each person called in a nationwide poll is the stand-in for 60,000 to 75,000 voters. A dozen opaque responses in a crucial state can make the difference between getting it right or mouthing a lame excuse for an erroneous call.
Even more dangerous for the pollster than silence are replies at variance with the respondent's heartfelt beliefs -- fibs, some would call them. The polling industry smells its own catastrophe in this practice, and rails against it as a form of treachery.
When the rinsed-off Nazi David Duke fared better than the polls predicted in the Louisiana gubernatorial race, the pollsters concluded that many voters were too embarrassed to admit support for Mr. Duke, but then voted for him. What the pollsters tended to gloss over in this episode is that their techniques are defenseless against non-cooperation.
When the pollster calls, the proper response to inquiries about political preferences is "None of your business, thank you." The rationale for political campaigns is that they help mold political preferences. The function of elections is to demonstrate those preferences. Polling detracts from the process. Fortunately for good citizens, a homemade remedy is easily available.
How often can you solve a serious problem by simply hanging up the phone?
Daniel S. Greenberg is syndicated columnist specializing in scientific and medical issues.