EVEN BEFORE George W. Bush stomped Al Gore in the second presidential debate, there were signs that Bush's campaign was gaining ground. One big sign, actually. A Gallup poll commissioned by CNN and USA Today showed Bush ahead nationally by 8 points. A poll by Gallup released three days before had indicated that Bush was losing to Gore by 11 points. In 72 hours Bush had surged 19 points. That's a gain of more than 1 percentage point every four hours, including the hours when most of America is asleep. No presidential candidate has ever risen so quickly in the polls. Bush partisans cheered.

Pollsters who don't work for Gallup groaned: "It doesn't happen that way"; "Get real"; "That poll is a joke and a disaster." Those were comments from Republican pollsters. Even the Bush campaign made it clear that the Gallup poll must be wrong.

And it was. The samples of voters Gallup used for its polls apparently were skewed. In the first poll, too many Democrats were surveyed; in the second, too many Republicans. There was also the problem of data gathered on weekends, as some of the Gallup numbers were. It has been long known that polls taken on Friday and Saturday tend to yield odd results. (Consider who would willingly engage in an extended colloquy with a script-reading poll company employee on a Friday night. Not your average voter.) In the end, a spokesman for Gallup admitted to the Washington Post that the company's techniques may have "slightly" inflated the results.

If Bush isn't up 8 points, where is he? In an unusually close race with Gore, is the short answer. Who's winning depends on how you read the numbers. At the end of last week, according to an analysis by National Journal, Gore was leading Bush in states with a combined total of 281 electoral votes, 11 more than required to win. Bush was up in states with a total of 227. But if you discount the polls in which either candidate is leading only within the margin of error, things look different. By this measure, Bush has locked up 127 electoral votes. Gore has 92.

Consider the most important states individually and you get the same split picture. Gore is now winning in a huge hunk of the vital eastern heartland -- Michigan, Wisconsin, Minnesota, Missouri, and Illinois -- as well as in Pennsylvania. Or he may not be. In each of these much-written-about swing states, Gore's lead falls within the margin of error. The same is true of Bush in Ohio, Tennessee, and West Virginia, among other states.

A race this close makes for a confusing electoral map. "If the grand and august national political media only knew how much they were jerked around by the margin of error, they'd be shocked," says GOP consultant Mike Murphy, who with obvious pleasure describes polls as "the astrology of the modern campaign." The Bush campaign, however, believes it sees movement in the numbers. This is partly because there has been some. Three weeks ago, an unreleased poll commissioned by the campaign showed Bush down by 13 points in Pennsylvania. Last week, the same internal poll (this one released to the press) indicated that Bush was within 2 points of Gore there.

Strategists in Austin, who have access to far more detailed data than news organizations are willing to pay for, say they have other polls that show similar, though mostly less dramatic, movement. But ultimately even the most basic national surveys may be enough to call the race. If Bush winds up leading Gore nationally by 3 points or more, the Electoral College probably won't matter. At that point, the swing states swing. No candidate is going to be 3 points up on Election Day and lose the election.

The Bush campaign believes it is already there. In a tight race, debates matter, and since the candidates' first encounter in Boston, there has been a 4- or 5-point movement toward Bush in national polls. Matthew Dowd, a longtime political consultant who analyzes polls for the Bush campaign in Austin, points out that Bush now stands at where he was in April. Once Gore's post-convention glow wore off, Dowd argues, Bush returned to his natural level of support.

And unless something dramatic happens, the Bush campaign seems to fully expect him to stay there. Although their product is constantly cited to foretell the future, most pollsters are curiously hesitant to make predictions. Matthew Dowd isn't. The polls taken after the third and final debate, Dowd says, will be a "signal to where this race is going into Election Day. There will be a pattern." If Bush is up at the end of the week, he'll likely win. If he's not, he likely won't. We'll know soon.

Tucker Carlson is a staff writer at THE WEEKLY STANDARD.

Next Page