Every single day every major news web site has a new "tracking poll" number. Reporters for these organizations make liberal use of these numbers in stories. You see phrases like "Bush surging" and "Gore rallying" peppered throughout various stories without any real reason to do so. The reporters don't know if the numbers are accurate, and they act like they do. In fact, reputable pollsters have no idea who will win. They'll make predictions, but they'll be drinking Maalox all night until the vote tallies come in.

I don't like to get didactic, but I know for a fact that lots of journalists don't understand statistics. Simple ratios are beyond the grasp of even some business reporters. When Apple Computer dropped from about $40 a share to less than $20 a share, a news story I saw talked about the stock losing more than 100% of its value.

That's just silly.

In any case, there are lots of reasons to not trust these polls. So I'm gonna try to explain how these polls do and don't work. Let me preface the first couple of questions by saying all of the major polls (all the TV networks--including CNN and MSNBC, Washington Post, New York Times, Los Angeles Times, etc.) all have Bush holding a 2-to-4 point lead. These polls have been consistent for a week.

Now, the questions:

Q: If I just take the average of all those polls, will that be more accurate than any one poll?

A: No. You can't average even two polls from different pollsters because everyone uses a different method of sampling. In fact, the more polls you throw into the mix, the less sure you can be of the results.

Q: If every poll on the planet has George W. Bush with a 2-to-4 point lead, doesn't that mean he has a 2-to-4 point lead?

A: No. Most of these polls have a margin of error of between 3 and 4 percent. That means that their results could be off by as much as that amount either way. You might recall that Bill Clinton held double digit leads in 1996 right before the election (someone had it at 15 or something), only to win by about 8 percent. The polls blew that election (some outside their margin of error), which might lead a sensible person to question exactly how it is that the pollsters determine their margins of error.

Q: How do pollsters determine their margins of error?

A: It's a complicated formula involving witch hazel and eye of newt. What, you think I'm kidding? In many cases, these margins of error are determined by statistical models that are based on statistical models. There is nothing hard at the foundation. Polling is an astonishingly inexact science. Just as the pollsters who had Jeb Bush well in front (at least 5 points in some polls) of Lawton Chiles the day before the 1994 Florida gubernatorial election. Yes, Lawton whipped out that wonderful "The he-coon walks just before the dawn" nugget at the last debate, but the simple fact is that the pollsters were so wrong that Jeb conceded soon after the election polls closed.

Q: If I vote for Ralph Nader, am I wasting my vote?

A: This one doesn't have anything to do with polls, but I'm tired of talking about them. Plus, I want to address this. In states where Nader is on the ballot, you are most certainly NOT wasting your vote by voting for him. If he gets 5 percent of the national vote, the Green Party will qualify for federal funds in 2004. That's hardly a waste.

In a state like North Carolina (where I live), where Nader is not only not on the ballot but ineligible to be a write-in candidate (we have very southern ballot access laws here), even the write-in votes will not be counted. Those will be wasted. I say vote your conscience. The only thing to worry about is the Supreme Court, and since most of the retiring justices are dead-to-right (you really think we can come up with someone worse than Rehnquist?), well, I think it might be safe to take your chances.

Q: Does George W. Bush's DUI conviction matter?

A: From a qualification angle? No. From a hypocrisy angle? Yeah. But then, the hypocrisy charge card on both sides maxed out long ago.

Q: Should I vote?

A: If you don't vote then you are not upholding your responsibility as an American citizen. Any attempt to justify not voting is self-serving and wrong.

Q: Shouldn't we have more than two viable choices for president?

A: Of course. There should be a real progressive party, not to mention a moderate-to-conservative workers party, a socialist party (not to be confused with the progressives) and a national front party (just because they're so much fun). But they don't really exist, at least in terms of putting forth candidates who actually might win. You're stalling.

Q: Damned right I'm stalling. I don't want to vote for Bush or Gore or Nader or Buchanan or Harry Browne. Who should I vote for?

A: Personally, I would like to vote for a ticket of Jesse Ventura and Jeffrey Feiger (best-known as Dr. Kevorkian's lawyer, but also the most recent Democratic nominee for governor of Michigan). Neither of them like the twin sacred cows of politics--the drug war and religion--and they both like to have fun with politics. The only fun in national politics these days seems to be in catching ugly old guys groping not-terribly-pretty sorta-young things. That's not fun at all.

Jon Worley is thinking of starting up a national political advice column. As if anyone would read such a thing, of course.