The time has come to trash the polls. It's not enough to simply ignore the results, although that process has begun with the gradual discrediting of preelection and exit polls in many American contests over the past two years. That has served to raise public consciousness and arouse public anger against the marketing mentality that has spread to all aspects of American life and has misused polling data to dictate an ever-narrowing range of policies that (and people who) govern us, the products we consume and the options we are offered for entertainment and information.

All these things -- products that reach consumer markets; candidates who reach the public ballots and the policies they espouse; the content of our newspapers and magazines; and the range of our television shows and books -- are now predetermined by marketing data derived in large part from opinion polls of different kinds.

Such preemptive marketing techniques are intrinsically damaging to the public interest because they limit our choices to a lowest common denominator. We never get to try a new variety of apple or hear a new policy proposal from a candidate because polls have shown that neither will sell enough to justify the investment required.

The damage is even more telling for another reason. What if the polls are wrong? The "science" of opinion polling has progressed to the point where the neutrality of questions and the questioning process can be standardized. But it remains a subjective process.

Computers have helped pollsters locate a limited sample that is representative of the whole. But it is impossible for the computer programmer to prejudge all the variables that might skew the sample. Most important, a grass-roots trend has begun among citizens who refuse to cooperate with opinion surveys (up to 30 percent of all those approached in some situations). There is often no way to determine motivation for the failure to cooperate, and therefore no way to discount it evenly should the pollsters seek to do so.

Now there is increasing evidence that subjects are going beyond non-cooperation and lying to poll-takers. This phenomenon, which may be far more widespread than is currently known, has surfaced in exit polls, one of the few types of opinion measurement for which an undisputable accuracy gauge -- the official vote count -- exists. How many other polls, used as marketing guides for policy pronouncements, television ratings or the flavor of beer, are fatally flawed by faulty questioning, faulty sampling, the failure to factor in non-cooperation or outright lying? There is no way to know.

One theory for the growth of non-cooperation and lying is that a growing segment of the public regards pollsters as "them" -- the enemy, the establishment, the problem and not the solution. Another is that those questioned view poll-takers as authority figures, and shy away from stating opinions or giving answers that stray from conventional wisdom or are not "politically correct." This has been especially true when there is a racial aspect to the issue -- as in the 1989 contests in Virginia and New York City and this year's North Carolina Senate race -- or when support for a particular candidate, such as John Silber in Massachusetts, was perceived as an almost anti-social gesture of protest.

Nevertheless, the result of this grass-roots rebellion has been to make poll-takers and those who use their results think twice about relying on such projections. And this is all to the good. On election night in Massachusetts, exit polls by the Roper Organization and the Boston Globe both showed William Weld with a small but statistically decisive margin of victory over the Boston University president.

Yet, with the lesson of the previous month's Democratic gubernatorial primary and its polling fiasco still in mind, the Boston media refrained from projecting a winner on the basis of the exit polls and waited until the actual tally showed a Weld victory (by a margin smaller than the exit polls had indicated and too narrow for any poll to have defined as statistically significant).

What is the lesson, then, for those who believe (as I do) that polls restrict vital options to a dangerous degree even when accurate and even more so now that vulnerability has increased and accuracy has decreased? The answer, I think, is to draw inspiration from the instinctive reactions of the public. Say no to poll-takers, on principle. Or, for the adventurous, lie like a bandit.

It's the least a politically correct citizen can do.

The writer teaches public affairs reporting as an associate professor of journalism at Boston University.