Read the polls and you'd think the White House race was all over. ``Not so!'' warns David Moore, a veteran pollster at the University of New Hampshire.
Professor Moore and other political experts recall earlier presidential races when the polls were widely proclaiming one thing, and the results were quite different.
In 1980, for example, some polls near election day said Ronald Reagan and Jimmy Carter were in an extremely tight race - too close to call. When it was over, Mr. Reagan had nearly 51 percent, Mr. Carter only 41. Why didn't the polls detect that?
This year, the leading polls are sending a different kind of signal. The latest Times Mirror survey has George Bush leading Michael Dukakis by 11 points. A CBS News poll puts it at 12 points. ABC/Washington Post says it's 13 points. And the KRC/Hotline tracking poll estimates the Bush margin at 11 points.
All that looks discouraging for Governor Dukakis, but Mr. Moore suggests it's too early for Democrats to fold their tents. Even President Reagan, a perennial optimist, issued a word of warning to Republican enthusiasts here in California this week:
``This election campaign is not over, and the only poll that means a thing is the one on election day,'' he told supporters in Orange County.
Three factors could keep everyone on the edge of his seat next Tuesday night as the votes are being counted.
One is the margin of error. The ABC/Post poll, for example, could be off by as much as four points. That's four points for each candidate. So instead of a 55-to-42 lead for Mr. Bush at this time, the race could actually be Bush 51, Mr. Dukakis 46. That's a whole different story.
Then there are the undecided voters. As the election nears, more voters are making up their minds. But throughout this campaign, the number of undecided voters has run twice as high as in other recent elections. Which way will they break in the final few days?
Finally, there's the softness of the vote. About one-fourth of the supporters for Bush and Dukakis have only lukewarm feelings about their choices. They admit that their opinions could change. A sudden, unexpected event in the next four days could make many of those voters switch sides.
Even with a 12-point lead for Bush, Moore points out that only 6 percent of the voters nationwide would have to change sides to make the race dead even.
All this discussion about Bush's large lead irritates some voters, who think the polls influence the outcome of the campaign. Indeed, a new ABC News poll finds that 45 percent of the voters this year think ``horse race'' surveys that show which candidate is ahead are bad for the country.
Thomas E. Mann, a political scientist at the Brookings Institution, notes that one scholar has compiled a list of about 140 national polls so far this year on the presidential campaign.
``The proliferation has gotten out of hand,'' he says. ``The problem is not the predictive power of polls. It's the fact that polls have become too much of the story of the election, rather than a mere indicator of how the race is going. The polls are driving coverage.''
The problem is particularly acute on the three major TV networks, Mr. Mann says. The evening broadcasts often start with the latest poll numbers, then everything that follows is qualified by what the polls say about the campaign.
``It is irritating,'' he says. ``On the other hand, print, cable TV, public television, and public radio have done a superb job of conveying to us what the candidates are saying, thinking, and doing. The substantive coverage on candidates and issues has never been greater than in this election year. But people get their most [vivid] impression from national TV news.''
Will the polls be wrong this year? Could there be a big surprise?
Although some experts think it's possible, Mann says it is the ``relative stability'' of the polling results that is most striking this fall. With the exception of a couple of polls with results that seemed at the fringes, ``this race has been about 10 points apart since late August. All the drama of the campaign has not really changed that.''
Mann discounts recent talk of a late surge for Dukakis. What appears to be happening is typical for the final days of a campaign: Dukakis's base is hardening as some Democrats rally around the ticket. But that probably won't change the outcome.
Does all this talk of an inevitable Bush victory color the public's perceptions of the candidates? Could it even alter the final outcome?
There is no agreement on that point. Mann, for example, doubts that it is shaping the public's views of the candidates. When Bush trailed by 17 points last summer, it was no deterrent to his fighting back and taking the lead.
But Moore suggests there may be an impact. He explains:
``If someone is presented as a winner and portrayed that way by the press, it does have a certain bandwagon effect. I don't know how much. ... But there have been numerous stories that reinforce the impression that Dukakis is a bad candidate, and that leads to the impression that this candidate is no good. That boosts Bush's lead - the polls show it - and the bandwagon.''