Election Day, at last, is just around the corner. Your heart's in your mouth, and you don't know if you can stand the tension while you wait to see how things turn out on the evening of Nov. 2. More to the point, you're hoping you don't have to wait until the wee hours of Nov. 3-or maybe even until sometime in December-to see who the winners will be.

How I wish that WORLD could address all that anxiety by providing a flawless forecast of how it will all turn out: who will win the presidency, which party will control the Senate, what our foreign policy will be for the next four years, and who's likely to be appointed to the Supreme Court. I wish-but, of course, we can't.

At the same time, and with a big chance of ending up with a good bit of egg on my face, it's altogether possible that what happens on Nov. 2 won't be close at all. It would hardly be the first time that the nation came right up on an election full of doubt and wonderment, only to be greeted by results resembling a landslide.

One reason this happens is simply that the media like a horse race. There's an inherent conflict of interest here, and the media should show a little more reticence at playing this game. Here are a few examples of what happens:

Letting polls eclipse the electoral vote. By mid-October, most of the major national polls were showing a virtual dead heat between George Bush and John Kerry. A month earlier, in the afterglow of the Republican convention, several pollsters had reported Mr. Bush with a lead of more than 10 points-but that big margin predictably evaporated as the campaign wore on. Of 18 recognized national polls between mid-September and mid-October, 14 showed Mr. Bush with a lead (averaging 3.9 points), two showed Mr. Kerry with a lead (averaging 1 point), and two showed the race as absolutely even. And virtually everyone predicted the polls would get even closer by election day!

The media love all that. It keeps everybody breathless. It keeps them tuned in and buying newspapers.

But presidents aren't elected by way of national polls-even when those polls are accurate. Presidents are elected by way of the Electoral College. And the media consensus through that lens last week was that Mr. Bush enjoyed something like a 300-238 edge. That was the consensus-but it wasn't making headlines. Were the media afraid to give up the horse race? Did the need to keep it close affect their reporting?

Using focus groups and other not-necessarily-representative samples. The TV networks are the big offenders here. Right after each of the big debates, CBS, NBC, ABC, and CNN usher us into the presence of nine, 12, or 15 members of a carefully chosen focus group to wait breathlessly for their responses. But wait! What possible vital information can this little group of people offer that you couldn't get just as well talking to folks on your own street corner?

The same fallacy is at work when USA Today devotes most of a page, as it did last week, to Pinellas County, Fla.-suggesting that Pinellas County could well be a microcosm of what we saw statewide in Florida in 2000. Such an editorial choice might make sense if Florida itself were genuinely considered a swing state right now. But The New York Times is not alone in suggesting (in its Oct. 10 issue) that Florida may not be close this year, reporting a recent 51 percent to 44 percent polling spread in Mr. Bush's favor and providing analysis on why that indication will probably hold.

Loaded questions. The tone of a political race is also shaped not just by the answers voters give to inquiring reporters but by the nature of the questions those reporters ask in the first place. The Media Research Center quotes ABC's morning host, Charles Gibson, as explaining that he would have the power to select which "uncommitted" voters would ask questions of Mr. Bush and Mr. Kerry during the second presidential debate. "We will eliminate some that seem, you know, superfluous or seem redundant, or whatever. And then I simply pick some at random, trying to cover the subjects that I think are important to voters." MRC's analysis shows that in the town hall "debates" of recent years, the selection process yielded 23 informational questions, 17 liberally oriented questions, and just six from the right.