Excerpted from A Nation Of Numbers by Paul A. Scipione

Whether you like the idea of political polling or not, this most visible component of the research business is here to stay. During the 1996 presidential campaign, it is estimated that the three major candidates – Bill Clinton (Democrat), Bob Dole (Republican), and H. Ross Perot (Independent) spent more than $15 million on various quantitative and qualitative research studies. Add in all the senatorial and house candidates and that figure increases to nearly $40 million. Clinton, in particular, makes heavy use of polling, not only at election time, but to help his administration make day-to-day tactical decisions after elections:


President Clinton not only wants to feel your pain, he wants to feel your pulse. To keep in touch, he uses polls like no other president before him … Polling data were at times collected almost every night to help the president win his battle with Republicans over this year’s budget … With every major policy initiative and political maneuver, Mr. Clinton must know what the public is thinking … “He uses polls to help make an argument in a way that people understand,” says Mandy Grunwald, a media consultant … Critics contend the president doesn’t rely on polls just to promote his policies but to decide them. “This massive polling proves Bill Clinton long ago gave up on reinventing government and instead is focused on reinventing himself for reelection,” says Nelson Warfield, spokesman for Republican presidential rival Bob Dole.

Presidents and presidential candidates … have long used polls, of course. Says Democratic pollster Mark Mellman: “The key in modern politics is to be able to relate to people, but it is difficult for presidents because they are removed. But no president or presidential candidate has spent more on polling than Mr. Clinton. The record $2.2 million that his campaign spent for polls in 1992 is almost certain to be topped this year.” Already, Mr. Clinton’s campaign has doled out $751,996 to New York polling firm Penn & Schoen, $113,647 to pollster Stan Greenberg and another $193,121 to [strategy advisor] Dick Morris. 2

Apparently Clinton didn’t like some of the numbers or the marketing strategies that were based on those numbers, because shortly before the 1996 presidential election, the president replaced his longtime personal pollster Stanley Greenberg with the doctrinaire Democratic polling firm of Penn and Schoen.3

The use of polling as a strategic political weapon in a democracy must work, because Clinton (45.6 million votes) beat both Dole (37.8 million) and Perot (7.9 million) to win reelection. The total cost of polling for the Democrats worked out to around 20 cents for each voter they attracted, although voter turnout in 1996 was among the lowest in American history. Perhaps many Americans are now under the illusion that responding to a telephone or mall poll actually counts more than going out to vote on the first Tuesday in November because fewer than 48 percent of registered voters cast a ballot in 1996, the lowest proportion since 1924.

There are many, though, who argue that political pollsters have become the handmaidens of their client-masters and that the results of their polls are either deliberately manipulated and/or misreported in order to help the client candidate attract more votes. If true, this would give new meaning to the poll as a “strategic political weapon,” as well as poking holes in the “myth” that pollsters are neutral, apolitical beings. This controversy came to a head in 1996 over the wide divergence in poll results. In an op-ed piece in The Wall Street Journal entitled “The Pollsters’ Waterloo,” Everett Ladd, a professor of political science at the University of Connecticut, where he also directs the Roper Center for Public Opinion Research, wrote:

Election polling had a terrible year in 1996. Indeed, its overall performance was so flawed that the entire enterprise should be reviewed by a blue-ribbon panel of experts – from academia, commercial polling firms and the news media – who should recommend ways to improve the accuracy of polling and of news reports about the surveys’ findings … Of course it would be naïve to think that, absent such a panel, polling will be doomed. But it would be equally misguided to assume that our profession will ever regain public esteem without such a major reassessment.4

The editors of the Newark (N.J.) Star-Ledger made the same complaints, but went even further, in a post-1996 election editorial piece entitled “Polls That Lie,” ultimately concluding that perhaps we have become too much a nation of numbers:

The polls got away with one [lie] in the election. They were up to 10 points off, but it was 10 points in the right direction. If they had erred by this amount in the other direction, some polls would have picked the wrong winner. As it was, they did something almost as seriously wrong, predicting a landslide that didn’t exist …  Did all this affect the election results? Probably not. But it affected the way we felt about the election and it retarded the democratic process. Repetitious daily poll results, each predicting a landslide, may well have been part of the reason for the disgracefully light turnout. They also were one reason why so many people felt no understanding of what issues were at stake in the election. Who can worry about issues when we are being told that it’s all over except the voting?

Some would argue that polling is still being used as a political weapon. Arguably the biggest and most expensive MR study in the world – the decennial U.S. Census – is used by state legislatures to redraw the boundary lines on U.S. congressional districts. There has been an unprecedented level of controversy (more than 100 lawsuits to date) over the issue of whether local probability sampling should be used to supplement actual census counts.

Since supplementary sampling is presumed to result in higher counts of blacks, Hispanics and certain other minorities, Democrats favor sampling because they feel virtually all minority voters will vote for Democrat candidates. Republicans do not favor supplementary sampling, at least as it might apply to the population counts used for redistricting. Many Republicans say their opposition is because sampling would violate Article 1, Section 1 of the U.S. Constitution, which calls for a complete enumeration (count) rather than projection or approximation (sampling). At the time of this writing, the issue is yet to be resolved.

The book is available for purchase from Paramount Books.