- The Washington Times - Friday, May 18, 2012

ANALYSIS/OPINION:

Public-opinion polling, like politics, prostitution and punditry, is an honorable enough profession, if properly understood and taken with enough salt. But usually it isn’t.

We’re awash in polls, most of them contradictory, each pretending to be an accurate barometer of what’s on the mind of the average voter. Alas, the average voter has the attention span of a fruit fly flitting from banana to mango to plum. This suits the average politician just fine because managing and manipulating attention spans is the key to what counts most, his survival.

The average politician understands what polls can tell him and what they can’t, enabling him to adjust his convictions and rearrange his principles, if necessary, and sometimes even government policy as public opinion shifts and changes with each news cycle.


Early in World War II, the Gallup Poll asked Americans what they missed most as the government imposed ever-stricter rationing of consumer goods.

Women’s nylon stockings were second only to tires and tubes on the list of most-missed goods; hairpins were high on the list, along with refrigerators, automobiles and washing machines. President Roosevelt couldn’t do much about the missing stockings, because scarce nylon was needed for parachutes, or refrigerators, because steel was needed for guns and tanks. But he loosened restrictions on glycerin so women could have lipstick, which improved female morale no end and the morale of men even more.

Gallup’s archives give an insightful glimpse of public attitudes of the past, some of them mercifully ignored by politicians of that earlier day. On the eve of World War II, 39 percent of the Americans Gallup polled said the health of their families would improve if they only had more money to spend on food, and if they had more money, they would spend it on meat, vegetables, milk and fruit. Sweets barely made list. When Gallup’s pollsters asked Americans in early 1945 what should be done with Japan when the war was over, 33 percent wanted to destroy Japan “as a political entity,” and 13 percent said simply, “Kill all Japs.” Polls have their uses, but in small, carefully controlled doses.

The average voter of 2012, on the other hand, is tempted to use a poll result to tell him who’s going to win a November election. Polls taken in May can’t do that. A poll is only a snapshot of public opinion, accurate enough but only of opinion at that moment. Cameras typically take snapshots at a shutter speed of just 1/125th of a second, a short time even for a fruit fly.

This week, polls were all over the place. Fox News reported a poll showing Barack Obama leading Mitt Romney by 46 percent to 39 percent. Rasmussen reported that its daily “tracking” poll had Mr. Romney leading the president by 46 percent to 45 percent. The necessary caveat is that voters were asked their preference “if the election were held today.”

If the election were held today, we would all be very surprised. The astute politician is more interested in a new Rasmussen poll that finds Mr. Romney more trusted to deal with the economy, and by a comfortable margin of 51 percent to 39 percent. This is the finding scary enough to frighten the paint off the walls at the White House. A perception like that is hard to reverse. Other recent poll results to frighten Democrats are that just 33 percent think the economy is getting better, just 37 percent give the president passing marks on the economy, 65 percent are angry at government policy, and 56 percent want Congress to repeal Obamacare.

The so-called horse-race poll, which only purports to measure which candidate is ahead, is thus worth very little in May. The more astute the politician, the more questions he has. How likely are the voters in the poll to actually take the trouble to vote? How passionately held is his opinion? Pollsters insist that what they’re selling is more science than art, and the accuracy of the best of them can be astonishing.

Predicting what humans will do is always risky business, but handicapping elections is more accurate than handicapping the horses (if only because horses aren’t handicappers). “It would be folly to argue that behavior can be predicted with perfect accuracy,” George Gallup, the godfather of polling, said many decades ago. “It can’t and never will be. But behavior can be predicted with a high degree of accuracy. The goal is to increase this accuracy.”

These are words from the wise to the wary.

Wesley Pruden is editor emeritus of The Washington Times.