Associated Press
In this July 24, 2012, photo, President Barack Obama speaks at a fundraising event at the Oregon Convention Center in Portland, Ore.

This week's conventional wisdom is that the presidential campaign is over. "Mitt Romney is in a death spiral, caused by his own ineptness, and Barack Obama is beginning to coast. The proof is in the polls."

In response, the blogosphere is full of debate over how the various poll samples are being weighted, with Obama's supporters saying current readings are perfectly accurate and Romney's claiming that they are deliberately skewed towards the Democrats. At least something strange is going on, because, as I write this, the two most respected polls, Gallup and Rasmussen, are farther apart than they have ever been. Gallup shows Obama ahead by 6 and Rasmussen says they are tied.

There is really no way to determine who is right. Polling an election is as much an art as it is a science because you cannot know, in advance, whether or not the sample you are measuring is really representative of those who will actually vote. Historically, to achieve that result, pollsters have used the turnout at the last election as their guide to this year's voter behavior.

That may or may not work this time. 2008 was an unusual year. When we break down its electorate by racial make up and measure 2008 turnout rates relative to the eligible population, we get these figures: Among ethnic groups that are growing as a percentage of the population — those referred to as the "coalition of the ascendant" — black turnout increased by 4.9 percent, Hispanic by 2.7 percent and Asian by 2.4 percent. Exit polls showed that each of these groups gave the majority of its votes to Obama.

However, among white voters, there was a 1.1 percent decrease in 2008. If that had not happened — if they had turned out at the same rate they did for George W. Bush — there would have been 1.7 million additional voters last time, the majority of whom presumably would have voted for McCain because white males are Obama's worst demographic, and three times as many white men stayed home as women. A Republican majority of an additional 1.7 million white votes, distributed in the right states, could have given the presidency to McCain.

So, the question of whether or not the election is over depends on whether or not the current polls are telling us the truth about what will happen with respect to turnout. Will the "coalition of the ascendant" vote at 2008 levels, or even higher? Will the white vote stay away as much as it did last time? Or will the 2012 electorate differ enough from 2008 that the polls based on 2008 turnout will be proven wrong and the pundits will be surprised?

Team Romney takes heart from polls showing their voters to be more enthusiastic than Obama's. They also point to the fact that Romney's events, so far, are drawing more attendees. That information is anecdotal; the only statistically based indicators I have seen are requests for early absentee ballots. In Ohio, in 2008, these requests favored Democrats 39 percent to 19 percent. (The balance went to independents.) In 2012, that ratio has narrowed to 30 percent to 25 percent, a trend that holds even in Democratic strongholds. In Cuyahoga County, Democratic requests are up 2 percent while Republicans are up 9 percent.

It's dangerous to try to draw too many historical analogies, but polls have been wrong in the past. On Oct. 26, 1980, they showed Jimmy Carter ahead of Ronald Reagan by eight points. Ten days later, Reagan won by 10. It's still too early to be calling this one over.

Robert Bennett, former U.S. Senator from Utah, is a part-time teacher, researcher and lecturer at the University of Utah's Hinckley Institute of Politics.