MOST VOTERS don't care a fig about polls on the presidential race, and it is obvious why, since the only poll that really matters is the one taken on Election Day.
Even so, polling on the year's biggest race is a big and growing business.
Matthew Dowd, the president's pollster, says there are twice as many publicly reported polls on the race this year as there were on Bush vs. Gore in 2000 and four times as many as there were when Ronald Reagan ran for re-election in 1984 against Walter Mondale.
The people who tend to pay a lot of attention to poll results are found in (where else?) politics and in the press. Candidates for office have been known to tack this way or that in response to what their pollsters think they have discerned about the electorate. Those covering campaigns know that and know, too, that a poll can furnish a peg for a story that is useful when better pegs are lacking.
Because we have an incumbent running for re-election, the polls drawing the most notice this year are those reporting how many Americans approve or disapprove of the way President Bush is doing his job.
Every poll on that question taken so far this month--six months from the election--reports approval ratings ranging from 49 percent (Rasmussen) down to 42 percent (Newsweek).
The question is what those under-50 numbers portend.
One answer, embraced (as you would expect) by the Kerry campaign, is that Bush is headed for defeat.
That answer draws on the past. Starting with Harry Truman in 1948, the four incumbents winning on Election Day had approval ratings of at least 52 percent in the Gallup Poll taken in May, while the three incumbents who lost had ratings no higher than 47 percent.
Dowd (as you also would expect) has a different answer. He agrees that "approvals" predict the ballot outcomes. But he says it isn't the May rating but the one in October that matters.
Dowd thinks the president will win if his last approval rating is above 50 percent and surely will lose if it is below 40 percent. He thinks a downward shift below 40 percent is unlikely because of the solid level of support Bush has among Republican voters.
Though Bush's approval ratings have declined in recent weeks, John Kerry hasn't been able to move ahead in the horse-race polls.
Andrew Kohut of the Pew Research Center for the People and the Press says voters aren't focusing on Kerry but on Bush, who as a war president dominates the news (much of which has been negative for him).
"There is no reason to expect a one-to-one relationship between public disaffection with the incumbent and an immediate surge in public support for his challenger," Kohut writes.
But a later surge could occur. Again, there is history:
Jimmy Carter, notwithstanding his dismal approval rating in May 1980, still led Reagan at that same point in the race by 12 percentage points. Only in the last two weeks of the campaign did Reagan begin to gain in the head-to-head polls--a sign that he had become a viable alternative.
Almost everything the Kerry campaign now is doing is designed to ensure that the candidate acquires that status.
Both campaigns think the race will be close. Yet "the hot new theory that has the chattering class all abuzz," writes Roger Simon of U.S. News & World Report, isn't a close race but a blow-out, by Kerry or by Bush.
If the latter happens, the political pros will have to ask what it was about this election year that made the May approval ratings no basis for predicting the final outcome.
Terry Eastland is publisher of The Weekly Standard. This column originally appeared in the Dallas Morning News.