Download The Influence and Impact of Political Polls and more Study notes Probability and Statistics in PDF only on Docsity!
Published in Washington, D.C. 5am -- October 19, 1998 www.washtimes.com
Polls take
commanding lead
in politics, business
By August Gribbin THE WASHINGTON TIMES
Part one of two here are folks out there who want to probe your mind. They want to know what you think, how you feel, how you'll vote and why. Even if you don't tell them, they have ways of finding out -- or at least of surmising. They abound in business and industry, where their marketing surveys ultimately determine what you will eat, wear, read, drive and view on television. Importantly, their work can be used to slyly influence your perceptions of the world. Many scholars say pollsters now are pivotal in the democratic process. Some say they're wrecking it. Yet few people outside the polling industry clearly understand what these hidden persuaders really do, or how their work can inform, manipulate and, at worst, deceive. It's generally conceded that polling is indispensable in smart campaigning. That's partly because changes in the size and lifestyles of the electorate prevent candidates from learning about their constituents firsthand. Similarly, officeholders need poll data to support decisions and for planning ways to implement them. But those who find polls destructive say poll results too often dictate the policies of officeholders and candidates, who bend with the winds of opinion, letting surveys serve for conscience. -- Continued from Front Page --
There's another reason why some see polls as a threat, and it also shows how polls can influence public perception. In the 1980 presidential election, early exit poll results showed Ronald Reagan a decisive winner. The results were released three hours before polling places
closed in the West. Many Western voters concluded voting was pointless and didn't bother to cast ballots. The critics notwithstanding, political pollsters have emerged as the seers and savants of officeholders, office- seekers and political pundits. They have armed clients with the ability to spot -- and dodge -- issues on which they and their constituents may disagree, while identifying emotion- churning topics that will attract support. And in general, states Victor Cohn, "Surveys are conducted and often distorted by groups trying to sell, rather than report, some view." Mr. Cohn, a fellow at the American Statistical Association, has authored "News & Numbers," a book about reporting statistical information. He notes, "All of us -- very much including journalists -- need to temper our acceptance of surveys and polls with some knowledge of their structure and limitations."
Finding a random sample
In fact, legitimate survey researchers use the same basic tools -- questionnaires, trained interviewers and the mystifying process called "scientific sampling." That's a proven method of tapping the opinions of hundreds of millions of people by getting answers from only a small, representative and randomly chosen group, or "sample." Since nearly everyone these days has a telephone, public opinion and political pollsters conduct interviews by phone, using "random-digit dialing." Computers generate a random sample of phone numbers from a list of all working numbers, including unlisted ones. Kathleen Frankovic, poll director for CBS' television news division, says phone interviewing is cheap, quick and dependable. "We can control the phone interviewers and the process better than we could control interviewers we send out on their own, making door-to-door calls," she says. CBS produces its surveys jointly with the New York Times. The two organizations share expenses and expertise, maintain phone banks, train interviewers, organize surveys and the rest. Miss Frankovic and Michael Kagay, the Times' news surveys editor, report to news executives. They attend news meetings and work with the reporters, producers and editors assigned to prepare survey stories. The polling editor or director develops survey questionnaires working with reporters or TV producers. That's because preparing the right questions in appropriate, unbiased language and in the correct order is essential.
Crafting unbiased questions
For example, a poll might ask, "Would you vote for someone who favors the slaughter of unborn infants?" The
Since few people can be expected to understand the issue, which they may not have heard about, their answers would be of little value. Resulting headlines indicating "U.S. Citizens Oppose [or Favor] Russia Gas Line" would give the misleading impression that the question is a "hot button" issue. Polls "alter people's perception and have a residual effect on public opinion," says LaSalle University political scientist Ed Turzanski. If they didn't, he argues, politicians would not use "push polls." Push polls ask carefully honed questions that guide (or "push") interviewees to a predetermined conclusion. For instance, they might imply that political candidate Jones is an alcoholic or that a health care bill of rights would send medical costs soaring. Such polls actually are pseudo- surveys -- propaganda devices. A prime example of a push poll question, given by researchers Norman Bradburn and Seymour Sudman in their book "Polls and Surveys" is: "Are you in favor of allowing construction union czars the power to shut down an entire construction site because of a dispute with a single contractor, thus forcing even more workers to knuckle under to union agencies?"
Polling partisanship
Although many political survey research firms work for businesses and lobbyists and turn out newsletters as marketing tools, they generally don't work for the news media. Also, they limit their work to members of just one political party. They are partisans. And their work is extensive. Says pollster Bruce Blakeman of Wirthlin Worldwide: "In most campaigns today, the pollster is a key strategist. He sits at the table and helps decide what kind of advertising to use, what areas the candidate should campaign in. He is one of the four or six people who make the big tactical and strategic decisions." Political pollsters routinely produce tracking and exit polls. They also run surveys and conduct "focus group" sessions. At these, a trained "facilitator" engages a small group of randomly selected persons in conversation about a topic as the group's reactions are secretly monitored. Pollsters determine how familiar a candidate's name is, the extent of voters' devotion to the party, and the kind of image the candidate --and his opponent -- projects. They "test" the candidate's proposed "messages," or advertising ideas. And they determine which voters the candidate must win over on Election Day, what messages will persuade those voters and when and where the messages should be given. Pollsters test speeches, too. They provide members of a
selected audience with hand-held dials and show them a video of the candidate speaking. Viewers turn the dial to indicate the ideas they like and those they dislike. The result is a computer-generated, second-by-second graph overlaid on videotape that shows the speech's word- by-word effect on the audience. It reveals "power phrases" that stir the audience and can be used over and over in subsequent speeches. One example: President George Bush's "points of light," a phrase that captivated test audiences and ultimately stimulated development of citizen volunteer programs across the nation.
Revealing voters' emotions
Using specially trained interviewers asking meticulously devised questions during one-on-one sessions, the Wirthlin organization gets at the "values and inner emotions that drive voters' decision-making," pollster Blakeman says. He won't explain the precise techniques interviewers employ, but he claims they can uncover "the issues voters are using to discriminate between candidates," and learn why they are important. The pollsters then chart how people in different areas feel and build a "communications strategy" for "what to say and what to avoid saying in the campaign." Depending on the candidate's inclinations, the proprietary information pollsters provide can be used to deal with the electorate's concerns or to avoid them. An oft-cited textbook example of issue avoidance came in the Bush-Dukakis campaign 10 years ago. Polls revealed the public was interested in education, health care and the environment. Instead of focusing on those issues, Mr. Bush challenged then-Gov. Dukakis' liberalism, his policy of furloughing prisoners and his opinion on requiring students to the salute the flag. The governor responded in kind. The result: a campaign about "flags and furloughs" that one columnist called an "issueless, charisma-barren campaign ... ideal for the special art of public-opinion manipulation." Not all pollsters use all techniques, of course. Their clients may be unable to afford them. Polling costs can range from $7,500 for a 15-minute survey of a small congressional district to $30,000 and up, up, up for a 40- or 50-question poll of 800 persons. Costs depend on the kind of questions asked, the number of subgroups included, and a host of other variables. Beyond that, not all pollsters do polling well. Some choose the sample group poorly, write inept or biased questions, use shoddily trained or untrained interviewers, conduct polls when people are unlikely to
