The Media

By TOM LITTLEWOOD

The perilous power of the polls

"NOBODY knows."

The erratic beginning of this presidential selection season is best memorialized perhaps by that opening sentence of a Chicago Tribune story on the Sunday before the Illinois primary, written by a shell-shocked reporter who was by then thoroughly daunted by the jarring ordeal of having told the voters repeatedly what they were going to do before they did it.

The journalistic imperative in all political reporting these days is to discern significance in what may be unclear or even, heaven forbid, insignificant. In the months before the first Democratic party voting, the media allowed themselves to be lulled into a semicomatose premature consensus that the race was over before it started. When this proved to be sensationally incorrect in New Hampshire, they then overreacted by digging Walter Mondale's grave.

One of the sources of this pressure to overinterpret the straws blowing in the political winds is the omnipresent opinion polls. The poll stories dominate campaign coverage. They drown out the other news of the campaign. Our attention is fastened on who seems to be ahead and by how much.

44/June 1984/Illinois Issues

Actually, until the final few weeks before an election, many Americans are not thinking seriously about who they want in any public office, including the White House. They're worrying about the knock in their car motor, about their sour stomach, about their bowling score. Meanwhile, the polls are measuring name familiarity, image, flashes of impression — and doing that in a rather capricious way.

Particularly in this early phase, the polls can foul up even without Mike Royko organizing a conspiracy. They do this by eliciting opinions that are no more than momentary reflexes.

Traveling with Gary Hart in March, Meg Greenfield of the Washington Post got "the impression of an electorate impersonating itself. Little kids, even, seem to have been listening to the 6 o'clock news and to spout back the jargon about how they are feeling and what they are thinking and what the country wants. 'Well, I like his ideas,' they will say, or whatever the latest thing is. . . . The public sounds as if it is thinking something it only yesterday learned it thought. And, of course, the candidates are self-creating characters too."

Here she is on to the new political "realities" of the video age: instant reactions quickly washed away.

Royko's campaign to mess up the network exit polls is ironic because they are at least measuring something real. If the networks could restrain themselves from telling the world about their findings before the election has been completed, the exit polls would have the virtue of measuring something that has happened. Not so the preelection polls printed in, among other places, Royko's newspaper. More often than not, they are measuring the made-to-order responses of fickle opinion holders in what Greenfield describes as a simulated campaign. They can do real harm by distorting the electoral system.

Let us take an example not from the presidential campaign but from an unsimulated campaign: the Democratic nomination for U.S. Senate in Illinois sought by four well-qualified candidates with similar positions on the issues.

A little over a week before the primary election, the Tribune published a poll story with the headline: "Senate Battle a 3-Way Race to Wire." The statewide telephone poll conducted by Market Facts of Chicago consisted of 389 registered Democrats who said they were likely primary voters. Paul Simon had 23 percent, Roland Burris 20 percent, Alex Seith 19 percent and Philip Rock only 6 percent, with the other 27 percent undecided. On top of such a large undecided group, the 6 percent margin of error made the poll meaningless. Nevertheless, the story went beyond that by breaking out the subsample of black voters to tell us that Burris had 45 percent, Seith 11 percent and Simon 9 percent of the black vote, a mathematically invalid conclusion drawn from such a limited survey.

Because the turnout in a primary is both small and unpredictable, it is impossible to tell which poll respondents will vote. Rock's base was known to be in some of the most deliverable districts. Anyone remotely knowledgeable about Democratic politics knew he would do better than 6 percent. But imagine how difficult it is to raise money and energize your workers when the Tribune says you're running that far behind.

There is an even worse effect: how the dubious intelligence of the polls carries over into the news column. On the same day, the Tribune ran a story by R. Bruce Dold, which said: "Although Simon has failed to generate much excitement or improve his standing in the polls, he plans to stick with his basic themes of peace, justice and jobs for the last days of the campaign."

As it turned out, Simon's campaign generated enough excitement to win by almost 200,000 votes with 36 percent. He did this by identifying his support all around the state, and then inspiring these people to vote, which is how elections are won. Rock, meanwhile, finished with not 6 percent but 20 percent.

A case can be made for publishing and broadcasting scientifically conducted polls in the final weeks of an important general election campaign when the public is presumably thinking about the choices. The earlier polls are nothing but a nuisance, though, having no redeeming value as news or anything else.

The proliferation of poll stories will continue nonetheless, if for no other reason than everyone else is doing it. An interesting example of how this phenomenon works occurred in Bloomington just before the primary. Able to resist the temptation no longer, but not about to become entangled in all the paraphernalia of a full-dress poll, the Pantograph printed different phone numbers representing the Democratic presidential candidates. Readers were invited to dial the number of their favorite, the volume of calls to be automatically tallied. There was no way of restricting the call-ins to Democrats or to prevent zealots from dialing away to a fare-thee-well. The poll was not intended to be scientific, the Pantograph explained, but only to "give readers an outlet for their opinions" and to give what the paper described as "a fairly accurate picture of sentiments in central Illinois."

In the months ahead, the fall campaign for the Senate will present the Illinois media with many opportunities for good, interesting "non-horse-race" stories. It is a dramatic confrontation, with the protege of the late Paul Douglas going after the man who, as a young business executive, removed the aged lion from the Senate 18 years ago. It is a campaign with real issues. One of the important tests for the Illinois press will be how well Mideast issues and their part in the campaign are reported. These are differences that arouse the passions of a few voters, most of whom are throughly acquainted with the background of the region. For the vast majority of the electorate, however, it is a complex of issues that will require careful examination of the many interconnections.

June 1984/Illinois Issues/45