A voter's guide to reading the political polls carefully


WASHINGTON -- It is time for another consumer warning on the use and misuse of public opinion polls in American politics today. Here are some things to consider when you read the latest numbers:

1. The dirty little secret of the polling business is that a large minority -- or perhaps even a majority -- of those who are questioned often know little or nothing about the people or issues on which opinion is being measured. The answers often depend on the way questions are phrased.

Recent polls show that 60 percent of Americans believe there should be an independent counsel appointed to investigate White House fund-raising for the 1996 campaign.

But poll-takers are aware that any question with the word "independent" in it is certain to evoke a positive response at a time when so many voters are dismayed by the partisanship on both sides in Washington.

It is a good bet that if the question were worded so it pointed out that such an investigation might take two years or longer and cost several million dollars, popular opinion would be different.

Another example right now is the high standing of Gov. George W. Bush of Texas among potential candidates for the Republican presidential nomination in 2000. Political professionals know that most voters wouldn't know George W. Bush if they fell over him in the supermarket. But the name does sound familiar.

This lack of real knowledge by voters can extend right through an election campaign.

Some people who cast ballots for Sen. Eugene J. McCarthy in his 1968 challenge to President Lyndon B. Johnson thought they were voting for the Wisconsin Republican, Joseph McCarthy.


2. Most polls don't measure the depth of feeling among voters -- that is, whether they feel strongly about a particular candidate or issue or, alternatively, are just offering a response.

Poll-takers do have techniques for gauging how strongly people feel about an issue, but they are not used in many of the surveys routinely reported.

Thus, for instance, in presidential campaigns, a third-party candidate -- a George Wallace, John B. Anderson or Ross Perot -- may seem particularly strong in the summer and early fall, then lose support as the voters focus on the "serious" alternatives on the ballot.

Some poll respondents also lie, particularly when race is involved. Poll-takers know from long experience that the white support for a black candidate is usually overstated because there are some voters who won't vote for a black person but don't want to appear prejudiced against blacks.

3. Polls often force respondents to make premature choices -- that is, to identify a favorite candidate when they haven't really paid much attention and may well change their minds. Again, poll-takers take steps to determine how "hard" the support may be for a candidate, and that information is particularly valuable in polls commissioned by candidates for their own use. But the depth of commitment isn't obvious in the matchup numbers you see in the newspapers or on your television screens.

Another variable difficult to measure is the projected turnout. Poll-takers use "screens" of different standards to determine who are the most likely voters; the rule is that the tighter the screen, the more accurate the poll is.

Poll influence

4. Despite some poll-takers' protestations to the contrary, polls often influence campaigns. In New Jersey's gubernatorial campaign, Democratic challenger James McGreevey was getting a cool reception when he sought money from the national party until two recent surveys showed him running only a few points behind the Republican incumbent, Gov. Christine Todd Whitman.

Candidates often manipulate polls. When they know that a major poll is about to be taken in their state, they invest heavily in television advertising to "drive up the numbers." Then they use their enhanced position to raise money for more advertising to drive up the numbers in the next round of polls.

None of this suggests that polls are not extremely valuable in measuring the approval of a political figure or the trends in a campaign. And these days most of them are remarkably accurate. But even the best polls need careful scrutiny. Things are seldom as simple as they appear in an opinion poll.

Jack W. Germond and Jules Witcover report from The Sun's Washington bureau.

Pub Date: 10/22/97

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.