Elections involving incumbents aren’t supposed to be this close. Yet, with less than a day left, all the polls are within the margin of error. A dead heat. Impossible to call. Are we on the way to another 2000, but with more lawyers?

Quite possibly, but more likely is one of these candidates will emerge with stronger support than the polls see. Indeed, there’s a growing belief that surveys, as we know them, are antiquated and not particularly accurate.

One problem is polls rely on most-likely voters – in other words, people who participated in the last election. This cycle there is a particularly huge number of first-time voters – an estimated 12 million. Those people are completely out of the pollsters reach, and no one knows if they are voting out of disgust with Bush, or fear of a weakening America. Or if they’re simply 18.

A second problem that is growing worse is the generational gap in the current surveys. You see, polls are conducted on landlines. Larger and larger segments of the population don’t even have landlines, getting their Internet via cable and their communication via cell phones. These folk are usually from the younger generation, one that in recent years hasn’t been politically involved, but this year finds Iraq, the possibility of a draft, and the degradation of the environment rallying issues (much more so than the public at large). Conversely, if you happen to be an up-and-coming business leader you are unlikely to let the neo-hippies on your dorm floor vote without counteracting them.

It’s hard to accurately gauge how many people are underrepresented in current polls. In addition, the different companies have greatly varied practices, i.e. Zogby makes an effort to get proportional party representation equal to recent voting patterns, while most of the others do not take declared party into consideration. Small changes in wording can also greatly affect the resulting replies, and each company is asking different questions.

In the early part of the 20th century telephone polling was invented, but it had a rough beginning. Back then only the wealthy had phones and, like today, they tended to vote Republican. Survey numbers were skewed and just wrong until the phone became cheap enough to be owned by most households. Now polls face a similar problem, only this time they are behind the curve. John Zogby, the CEO of Zogby polling, is talking about moving more towards Internet surveys and has already started experimenting in that direction. In an open letter he wrote in September, though, Zogby openly admits that polls in today’s age aren’t the accurate representations they once were. Times are changing.

Break it all down, and you realize something; this election isn’t likely to be decided by the undecided. This race is going to come down to people who haven’t been in a voting booth before, by the dot-commers who spend their time glued to a Nokia, the college kids who don’t bother with phone lines when they’ll be moving again in 9 months. Those who haven’t yet given an opinion will decide this election. The only reliable way to find out what they want is coming this Tuesday. This is more exciting than a dead heat – this is a complete unknown.