BHO: QED: why Obama is scientifically certain to win in November.

• Author: Balz, John
• Position: TEN MILES SQUARE - Barack Obama

If you spend time among the chattering classes, it s easy to get caught up in Campaign '08's most tantalizing questions. Has the length and tenor of the primary campaign irreparably split the Democratic Party in half? Could the Jeremiah Wright videos hurt Barack Obama more in the general election than they did in the primary? Might improvements on the ground in Iraq make John McCain's support for the war less of a problem with voters and burnish his image as a leader on national security? Well, sure, all these things could happen. But the disinterested, square students of politics, otherwise known as political scientists, find this kind of distress silly. That's because Obama will win this fall. Period. As a political scientist, I'll stake the reputation of my profession on it. Maybe some money, too.

Over the past few decades, political scientists have come up with increasingly elaborate statistical models for predicting presidential election outcomes. The basic idea is that voters evaluate presidents largely on the basis of how the economy has fared under them. That extends to potential successors in the same party. As such, McCain basically inherits Bush's record and Obama inherits Clinton's record. These statistical models have also gotten quite accurate in their presidential forecasting. In 2004, during a very close race, eleven out of twelve studies correctly predicted a Bush win.

So why is Obama certain to prevail? Much of it boils down to what you've already heard in the press: what political scientists call the fundamentals--primarily presidential approval, the state of the economy, and incumbency--are indisputably in the Democrats' favor this year. And don't be deceived by fluctuating poll numbers, either. At this time in 1988, George H.W. Bush was solidly trailing Michael Dukakis in Gallup polls. Likewise, at this time in 1992, George H.W. Bush was solidly leading Bill Clinton. Enough said.

While political science models differ from one another in the variables they take into account and in how they weigh them, nearly all of the models are quantitative. That means they not only pick a winner--they also forecast what his or her margin of victory will be. So what are the models saying for 2008? Douglas Hibbs, who has been studying the relationship between voting and the economy since the 1970s, has predicted that the Republican candidate will win only 46.7 percent of the vote. (That means the Democrat would win 53-3 percent of the...