The guessing game: How the US has gone poll crazy

If children could vote, John Kerry would be the next US president, according to the Nickelodeon network's Kids' Vote, in which the Massachussetts senator received 57 per cent to George W. Bush's 43 per cent. Nickelodeon has held the Kids' Vote in every election year since 1988, and claims that "its voters have correctly predicted the winner in the last four presidential campaigns".

But wait: Scholastic, the global children's publishing and media company, has just announced the results of its 2004 Scholastic Election Poll and found that "students in first through eighth grades prefer Bush, 52 per cent to 47 per cent". Scholastic claims its poll "has accurately predicted the winner in every Presidential election since 1960".

Welcome to the number-crunching nerd heaven that is the 2004 US presidential contest. No election in history has been so relentlessly polled as this one, and never have those polls been so readily available, via a bewildering range of websites, graphs, maps, multiple interpretations, disputations, and allegations of bias.

You don't have to pay any money, you just take your pick. I'm partial to the Electoral Vote Predictor at electoral-vote.com, largely because of its dinky, interactive map. Like the candidates themselves, the EVP doesn't pay much attention to the overall popular vote. It's all about the electoral college, and the magic number of 269. As I write this, it's predicting 271 votes for Kerry to 257 for Bush, with Minnesota a dead heat. The site is run by an avowed Democrat, but seems scrupulously fair. The more overtly Republican-leaning electionprojection.com has Bush at 274 to Kerry's 264.

Meanwhile, Slate.com has Bush with 262 votes to Kerry's 276 . . . hold on, though, it's just changed. Good news coming in for Bush: "He's suddenly safer in West Virginia, he's in a better position to take Ohio, he's in a good position to take New Hampshire, and he has taken Wisconsin outright. He loses one electoral vote in Maine, but the 10 from Wisconsin bump him up from 261 to 271, giving him the election." Whew. From zero to hero in the space of an hour.

All these sites happily allocate states to one candidate or the other on the basis of the most marginal leads in the most recent polls. As such, it could be argued that they're completely meaningless, or even, as some have suggested, pernicious in their effect on the US electorate. But with everyone pretty much agreed that this one is going to be a cliffhanger, they offer wonderful entertainment for amateur psephologists. Buzzing around their fringes are thousands of bloggers, offering their views on everything from The Great Cellphone Debate (are mobile users under-represented?) to Security Moms vs Soccer Moms.

Can the polls be trusted? Bush won the electoral college in 2000, but Al Gore won the popular vote, 48.4 per cent to 47.9 per cent. No pre-election poll reported those numbers exactly, but some came a lot closer than others to predicting Gore's half-point margin of victory. The best of the bunch were Zogby, CBS, Democracy Corps and Fox/Opinion Dynamics.

If you tally the results from the four pollsters who did best in 2000, they're currently showing Bush with a lead of 1 per cent. But a lot of pundits prefer to look at the President's job-approval rating, which currently averages at around 48.6 per cent. Is that enough for him to win the popular vote? Possibly. It was enough for Al Gore to win the popular vote in 2000 - and lose the electoral college.

Hold on, that map's changing colour again . . .