IN THE SPACE OF 50 YEARS, the public opinion poll has evolved from an occasional curiosity to a carefully heeded political force. "Bad poll numbers" can be disastrous. Nobody wants to champion unpopular schemes or seek public office when drawing single-digit name recognition in the latest survey. Even skeptical officials routinely monitor approval ratings lest tumbling figures embolden rivals. Poll-supplied numbers often shape momentous decisions. The Reagan White House spent $1.000,000 a year on polling. Pres. Bill Clinton's legislative agenda on spending the budget surplus, funding Social Security, and other key issues was shaped almost entirely by poll findings. When he discovered that his plan to make parents legally, responsible for their children's crimes drew dismal numbers, it was quickly abandoned.
Is such homage to the polls wise? Pollsters certainly think so. They celebrate their accomplishments as promoting democracy, especially defending the "ordinary citizen" against the powerful. Telephone surveys, they maintain, allow once-unheard voices to force government responsiveness. Ordinary citizens seemingly echo pollsters' claims. One 1999 survey found that 80% believed the nation Would benefit if leaders heeded poll results. Ninety percent expressed greater confidence in their own reasoning over what leaders believed. Two-thirds agreed with the pollsters that polls served the public interest.
We profoundly disagree--surveys cannot provide useful advice. It is not that numbers lie, analysts are dishonest, or people are secretive over the telephone, although these misrepresentations do occur. The surveys' shortcomings are more serious and. critically, these deficiencies are not curable under present-day conditions, if at all. What garden-variety advice solicitations uncover is generally politically irrelevant, even if respondents are absolutely honest and the highest technical standards are satisfied. The wrong opinions are being collected, and leaders following this guidance only invite trouble. Intuition or personal experiences provide a better course of action.
Let's begin by comparing polling with elections. An obvious point is that elections impose strict participatory standards. Laws about voting stipulate necessary age, citizenship, and residency qualifications plus standards of mental competence and criminal background for potential voters. Detailed registration requirements guard against fraud. "Everybody welcome, no questions asked" sign on Election Day! Abolishing standards would insult democracy. Who would accept an outcome if illegal immigrants or foreign tourists cast millions of ballots?
The modern poll, by contrast, resembles a "you cannot be turned down" credit card offer. Who knows if the agreeable voice at the other end is really a citizen, over 18, or mentally competent? The interviewer can hardly verify fraudulent claims by checking proof of age, citizenship documents, or any other requirement prior to soliciting opinions. The opposite is possible as well. No doubt, millions of absolutely qualified voters are "disqualified" by interviewers if their English is limited or if they are uneasy about revealing guarded views to strangers.
The vote. unlike the survey response, is private, and this protection is legally guaranteed. The scientific pollster is absolutely correct that small samples can represent entire populations. This is beside the point, though. Random samples of people owning phones are not legally stipulated governing majorities. Governance is only by those legally permitted to participate, however we may draw the lines. It is bizarre to insist that any 1,000 people selected by a random number generator should authoritatively advise government.
The electoral process is also eminently accountable and transparent. Everything about it, from ballot layout to polling place location, must pass official approval. These critical details may be flawed--recall the famous Palm Beach, Fla. …