Of what value are public opinion polls?

USA Today (Society for the Advancement of Education), May, 2006 by Richard E. Vatz

THERE IS A SPECTER haunting America--it is the never-ending proliferation of public opinion polls. From the contracts that polling organizations have with major and minor news organizations to candidates' and others' private polling, the practice simply is too much with us. It literally is impossible to go a day without hearing or reading of a new poll finding regarding what Americans--or, with increasing pretension, the world--think of some idea or position.

More specific, these polls often bring audiences unimportant, alleged opinions regarding the status of political elections as well as general positive or negative predilections regarding an ungodly variety of topics. To choose some representatively offending polls is akin to selecting specific grains of sand in the Sahara, but let us look at a few examples.

In one prototypical polling report, it was explained in an Associated Press article that "Americans are far more likely than the Japanese to expect another world war in their lifetime, according to AP-Kyodo polling 60 years after World War II ended. Most people in both countries believe the first use of a nuclear weapon is never justified."

Even ignoring cultural and language differences--an overly large concession, incidentally--can one reasonably infer from this poll that American and Japanese people are walking around with such opinions'? If they are, what possible significance--or difference--would it have?

First, we should look at the questions of whether it indeed is true that "Americans are far more likely than the Japanese to expect another world war in their lifetime" and that "Most people in both countries believe the first use of a nuclear weapon is never justified."

Perhaps the first reaction should be: So what? Indeed, what are the consequences of a larger percentage of Americans expecting a world war, even assuming there is a common definition of what "world war" or "expecting" constitute. Are we not in a world war now? What would have to happen to transform the war against terrorism into a "world war" when the current conflict includes participants in dozens of countries?

The poll claims that "Six in 10 Americans said they think such a war is likely, while only one-third of the Japanese said so, according to polling done in both countries for the Associated Press and Kyodo, the Japanese news service."

Periodically. the American public is asked how likely it believes the imminence of terror attacks to be. What is the importance of this finding, even, again, assuming that polls accurately can measure this anticipation?

Then there is "the margin of sampling error," which gives the impression of precision in polling results. In opinion polls, this margin cannot be verified by any means, but when people are told that pollsters "know" that the "real" polling numbers and percentages are ascertainable, they think that polling must elicit a true picture of public opinion.

The poll in question also asks whether the use of atomic weapons 60 years ago was avoidable. Avoidable at what cost? What does "avoidable" mean? How many American lives were saved by the bombing of Hiroshima and Nagasaki? What is the ability of the average respondent to assess a cost-benefit ratio of a war waged 60 years ago? The article points out that "Two-thirds of Americans say the use of atomic bombs was unavoidable. Only 20 percent of Japanese felt that way and three-fourths said it was not necessary."

Are the findings surprising or informative? Of what value are they'? Are the respondents informed regarding the issues they are polled on?

In the never-ending onslaught of polls into the consciousness of Americans, we learn that there allegedly is significant opinion-holding by Americans on complex issues; yet, no one has taken the time to discover the extent of the respondents' ongoing concern about issues before being asked, or their knowledge base. The Associated Press reported that a poll conducted by the Pew Research Center for the People and the Press discovered that "Americans consider Supreme Court decisions about the rights of detained terrorists as important as its rulings on abortion." Pollsters must have amended 19th-century columnist Peter Finley Dunne's Mr. Dooley's dictum that the Supreme Court follows the election returns to "the Supreme Court follows the ups and downs of pollsters' claims of public opinion."

Because 62% of those polled affixed the poll-provided category of "very important" to "detained terrorists' rights"--and this was only one percent less than the percent which calls Supreme Court decisions on abortion "very important"--Andrew Kohut, director of the Pew Research Center, inferred that "This important question of the trade-off of civil liberties and protection is one the public takes very seriously."

It did not seem to occur to Kohut that the polling itself makes issues--including this one--salient. Moreover. it also appears not to occur to many pollsters that the complexities of definitions of different categories of detainees, various informational and time-dependent requirements regarding incipient terrorist acts, as well as a multitude of other considerations might change the direction and intensity of poll responses. Certainly, any conclusion regarding a stable view of Americans as to their concern for "civil liberties" of detained terrorists simply is invalid, or at least insufficiently examined here. Nor is it clear what "very important" even means in this context. The same Associated Press article also stated that "Recent polling by ABC News and the Washington Post found that six in 10 Americans say the United States is adequately protecting the rights of people detained in the campaign against terrorism, such as those held at Guantanamo." Go figure.