POLLS: A YEAR TO BE WARY

ONCE in a while, all pollsters should take the kind of beating we took in the primaries, just to maintain equilibrium," says Don Muchmore, board chairman of Opinion Research of California, one of the many polling firms that came a cropper in one or more of this year's presidential primaries.

Looking toward November, the pollsters are unanimous in showing Lyndon Johnson far ahead of Barry Goldwater. But they are nonetheless nervous, partly because of their primary experiences and partly because they just don't like what they see in their statistics. Explains Dr. Peter Rossi, director of the University of Chicago's National Opinion Research Center: "In an election like this, you have a high proportion of the electorate undecided, a high proportion who are normally Republicans and saying they'll vote for Lyndon Johnson, but when they get into the voting booth, they may not be able to do it." Agrees Pollster Oliver Quayle: "Patterns now are not nearly normal. The situation is too volatile to be reliable."

The Techniques. As protection against a ruinous misjudgment this year, the individual pollster has only the techniques he has developed or borrowed from his colleagues over the years. They vary considerably from pollster to pollster.

California Pollster Mervin Field is following a trend toward "randomization" in selecting interviewees. He divides California into six regions, on which he collects basic social and economic data. Within each region he assigns interviewers to a few blocks in each county, instructs them to canvass, for example, every third house from the corner, ignoring race, religion, age and income. He insists only that an equal number of men and women be polled. Such random selection, the theory goes, will ensure a good cross section. Declares Field: "You either put your faith in probabilities or not."

This technique contrasts with the long-held theory that interviewers should reach a quota of persons in various categories--poor people, rich people, Republicans, Democrats, etc.--in the same ratio that they exist in the region studied. Lou Harris, Jack Kennedy's favorite pollster, uses randomization, but employs computers to spot-check the reliability of his sampling. If he suspects that his polls did not accurately reflect certain groups, he runs cards, on which the basic characteristics of key election precincts are punched, through a computer until it turns up a precinct that coincides with the types of voters he is worried about. Then he compares that precinct's actual vote with what his polls showed and corrects his sampling for the future.

The nation's most famous pollster, George Gallup, employs randomization, requires a fifty-fifty balance between men and women. His interviewers follow assigned patterns in selecting persons to question. They may be told to seek out the youngest voting-age person in each household on the probability that this will reach a balance of age groups. They skip some corner houses on the theory that corner property is higher-priced and its occupants are likely to be more affluent than their neighbors.

Probing the Past. Opinion analysts concede that they have not yet licked one problem likely to make them look bad on Election Day: their inability to assess the probable voter turnout. Explains Richard W. Oudersluys, president of the Detroit-based Market-Opinion Research Co.: "To take the figures from a poll and make them 100% accurate, we would have to have 100% accurate information on voter turnout.

Nobody wants to say he doesn't vote. It's not patriotic."

The best the pollsters can do is to press interviewees on their past voting habits. Oliver Quayle automatically eliminates anyone who admits not having voted in the past three national elections. When a person claims that he voted recently or is registered, Pollster John Kraft instructs his interviewers to seek out specific details, such as the place of registration, and drop any respondent who is vague. Yet such techniques cannot measure the effectiveness of a good political organization in getting out the vote. "Organization can make projections completely unreliable," declares Field. He notes that Goldwater could come close to election this year by holding on to Richard Nixon's Republican vote and, through a good get-out-the-vote organization, adding one-fifth of the 9,000,000 registered Republicans who failed to vote in 1960.

Do people lie to the pollsters? The question is particularly important in a year of such sensitive issues as race relations and "extremism"; yet the pollsters insist that they meet with a remarkable degree of honesty. Kraft contends that a good guarantee against deception is the type of interviewers employed. "The average person is faced with a quite pleasant, well-groomed, middle-aged lady not selling anything," he says. "It's very hard to look at that poor pleasant-faced lady and lie to her."

Yet the pollsters do take elaborate steps to prevent deception. The most common tool is the secret ballot or questionnaire that the respondent fills out himself and inserts in a box. Gallup uses this in about half of his interviews, thus can compare the secret and nonsecret results. Nearly all of the polls ensure anonymity by identifying all interview reports by only a code number once it is submitted.

Another--but more costly--way to determine truthfulness is the depth interview, in which several questions relating to the same point but phrased differently seek out inconsistencies. California Opinion Research uses up to 40 questions to assess the honesty of a respondent on a key point. It tries to measure the white backlash by a series of "rather or rather not" questions, such as: "Would you rather or rather not stand in a grocery line with a Negro? You may, of course, answer that it doesn't matter." Contends Muchmore: "By a series of questions on that theme, you can pinpoint what is going on, and you can pick up fast any backlash."

The 30-to 40-minute depth interview can turn up all kinds of information that more and more political candidates seem to find indispensable in planning their campaigns. Market-Opinion employs what it calls an "eight-part semantic differential questionnaire"--pollster doubletalk for a technique in which a person is offered eight adjectives and asked to circle the one that most closely reflects his attitude toward an issue or a candidate. On a candidate, for example, the words might range from "kind" to "cruel," and the answers can tell a candidate where his public image is weak, where his opponent is vulnerable. The California Poll allows respondents to select any of 23 traits to describe a candidate.

The Costs. The pollsters' product does not come cheap. Prices range from $3 to $7 an interview, depending upon their scheduled length. A Congressman may buy a 500-interview survey of his district for about $2,500. The Republican National Committee employs Princeton's Opinion Research Corp. for much of its polling, pays about $6,000 for a 2,000-sample study of a state as large as Massachusetts. Gallup requires 1,500 interviews for a national survey. For a nationwide depth study, a party may have to pay as much as $30,000. Jack Kennedy reportedly spent $1,000,000 on Lou Harris' polls in 1959-60.

Are the polls reliable enough to be worth such costs? In probing general attitudes toward candidates and issues, they undoubtedly come close enough to be of value to campaign strategists. When it comes to calling elections, most of the pollsters insist that they do not make predictions, merely measure the popularity of candidates at a given point in time. In the post-mortems they are, of course, the first to boast when they hit one right. But that seems fair enough, since they take a beating when they are wrong. And that is what has them worried this year, which may well be, as Chicago's Rossi warns, "one for the pollsters to be wary."

This file is automatically generated by a robot program, so reader's discretion is required.