clock menu more-arrow no yes

Filed under:

PRIMARY EXPOSED OPINION-POLL FLAWS

This election year looks like a tough one for the pollsters.

The hazards of using candidate preference polls as part of the coverage of a volatile campaign were never better illustrated than in last week's primary.Both the Deseret News-KSL poll and the Tribune-KUTV poll were correct on the direction of most of the major races but not on two of the most important. Both had Joe Cannon winning the Republican Senate race, the News by 2 percentage points, the Tribune by 4.2. The Tribune correctly indicated Stewart Hanson's win in the Democratic gubernatorial contest; the News had Pat Shea winning by 9 points.

The margins furthermore were wildly off the mark in most contests. The News poll, for instance, had Scott Burns winning the Republican primary for attorney general by .06 of a percentage point, the Tribune by 1 point. His walkaway win: 18 points.

- ELSEWHERE AS WELL polls continue to vary wildly. Five major national polls on the Clinton-Bush race are hard to reconcile. One shows Clinton ahead by 15 percentage points, another by only 5. In France, more than a third of the voters have not made up their minds on the vote on European unity, giving pollsters a fit in trying to sort the undecideds. In Britain last spring, the polls erroneously predicted victory for the Labor Party, leading the London Times to note that newspapers have been blamed for taking polls too seriously as predictive devices.

In the Utah polls, there was a good explanation for the discrepancies, of course. The pollsters noted the exceptional volatility of the race this year. Largely this was because of what News pollster Dan Jones called "voter movement." Huge numbers of voters were undecided as late as election eve, up to two thirds of those questioned.

The pollsters also were at pains to point out that preference polls are not predictions but only an indication of how voters would act at that time.

But for all of these protestations, readers are almost certainly going to take the tabular figures at face value more often than not, and that's something the pollsters have to live with.

- IF THE POLL CAN FLUCTUATE so much and can mislead the unwary, why should media even bother with them?

Whether newspapers ought to use candidate preference polls at all used to be a matter of considerable debate in journalism and politics. Today's editors are part of the computer generation that grew up with polls and don't have much first-hand acquaintance with the notable failures of past polls that turned off their predecessors.

Now it is a given that polls are part of the information voters are entitled to. This year's polls were valuable for no other reason than they did show the vast areas of uncertainty and misunderstanding symptomatic of the epidemic negativism of the electorate this year.

- POLLING IS STILL MORE an art than a science. The fundamental method hasn't varied since George Gallup introduced random probability sampling in 1936 and correctly predicted that year's presidential election. The same year the Literary Digest magazine predicted Alf Landon would whip Franklin D. Roosevelt even though it sent out ballots by the millions.

The techniques came in for a searing look after the 1948 polling debacle. Primarily, the big polling organizations stopped surveying too soon. Thomas E. Dewey was so far ahead of Harry Truman in the Roper Poll, then one of the three major national surveys, that it stopped surveying two months before the election. Roper found Truman would pull only 37.1 percent of the vote. He got 49.6. George Gallup and Archibald Crossley did a little better but still were wrong.

I well remember Elmo Roper, then a nearly mythic polling guru, speaking before my journalism class in New York in the fall of '48. With grand assurance he waved his cigar as he intoned that "it would be a major political miracle" if Truman won. He was right about that. He was dead wrong in ignoring Truman's whistle-stopping the Midwest with his "give 'em hell Harry" lashing of that "do nothing, good for nothing, worst 80th Congress."

ELECTION COVERAGE

Why there is so much voter confusion over what a party primary is and where the candidates stood on the issues confuses me. Certainly in very few elections have the issues been covered as well as they were these past few weeks.

We had a plethora of television debates between the major candidates and copious profiles of the candidates in the press. The Deseret News performed a superb service with those charts that compared candidate stands on various issues; the Tribune picked up on the same technique late in the campaign.

Brian Mullahy told KSL viewers during the election coverage Tuesday night that the press had perhaps not done so well with the minor races because so "many high-profile races have made it hard for the media to focus on the lesser races and hard for the voters to sort out the glut of information."

Our understanding of why people voted as they did would have been improved had the pollsters conducted exit interviews with voters leaving the polling places. Exit polls are much maligned, but typically ask people not only how they voted but why, thus yielding much-valued data for analysis by the election pros and political scientists. As it was, election night analysis was mostly speculation.

- PLUSES AND A MINUS: I liked: As usual, Channel 11's election night reports, in which students who have worked hard to get information on issues presented their analyses under the gentle and knowledgeable questioning of their mentor, Professor David Magleby . . . Channel 2's sensibility in making a repair of a botched incident. KUTV had cut away from a fascinating Jane Pauley interview with Hillary Clinton to update some voting figures; later it aired the excised program in its entirety . . . Channel 4's use of Randy Horiuchi, a Democrat, and Budd Scruggs, a Republican, as active analysts - they even quizzed some of the candidates from the studio . . . Channel 5's able insights by pollster Jones. The station even left it to him when to dub the election winners from the incomplete returns . . . Channel 5's 9 o'clock full-hour election coverage and comment, when other news stations were playing network shows . . . Ron Yengich's off-the-wall column of pre-election picks in the current edition of the weekly Private Eye; he was just as accurate as the polls, picking Bennett and Hanson and missing only on the Graham-Daniels race for attorney general. I didn't like the Tribune's leading the front page with an AP poll on election eve under the headline "Parents Want School Choice," a survey so flawed that even the story about it hedged.