From Charlie Cook's column in the National Journal:
. . . .
The metrics of this election argue strongly that this campaign is over, it's only the memory of many an election that seemed over but wasn't that is keeping us from closing the book mentally on this one. First, no candidate behind this far in the national polls, this late in the campaign has come back to win. Sure, we have seen come-from-behind victories, but they didn't come back this far this late.
Second, early voting has made comebacks harder and would tend to diminish the impact of the kind of late-breaking development that might save McCain's candidacy. With as many as one-third of voters likely to cast their ballot before Election Day, every day more are cast and the campaign is effectively over for them. The longer Obama has this kind of lead and the more votes are cast early, the more voters are out of the pool for McCain.
Third, considering that 89 percent of all voters who identified themselves as Democrats voted for John Kerry four years ago and 93 percent of Republicans cast their ballots for George W. Bush, the switch from parity between the parties to a 10-point Democratic advantage would seem to almost seal this outcome irrespective of the candidates fielded on each side. The unprecedented surges seen in Democratic party registrations in those states that require party affiliations confirm that.
Fourth, just look at the money and spending. With Obama now outspending McCain routinely by margins of 3- and 4-to-1 in advertising in so many states, it's hard to see how the Arizonan's campaign can drive a message. For a time, Obama was matching McCain one for one in negative advertising, then spending double or triple on top of that in positive advertising. Now Obama seems primarily doing positive ads, probably the right move given his lead going into this final stretch. Organizationally, it's hard to find any state where McCain is organized as well as President Bush was four years ago or Obama is today, a product of both money and enthusiasm.
Fifth, while many are talking about the so-called "Bradley effect," voters telling pollsters that they will vote for an African-American candidate when they won't, putting aside the question of whether it ever existed, it hasn't been seen in at least 15 years and the likely surge in turnout among African-American and young people seems sufficient to offset it anyway.
Finally there are the states. Obama is now leading in every state that Al Gore and John Kerry both won, including Michigan, Pennsylvania, Minnesota and Wisconsin, and he is ahead in Iowa, New Hampshire and New Mexico, the three states that went once but not twice for Democrats in 2000 and 2004. He is also ahead in Florida, Colorado and Virginia. If that weren't enough (and it is), he's running basically even in Indiana, Missouri, Nevada, North Carolina and Ohio, and even threatening in Montana, North Dakota and West Virginia.
As things are going now, this election would appear to be on a track to match Bill Clinton's 1992 5.6 percent margin over President George H.W. Bush, the question is whether it gets to Bush's 1988 7.7 percent win over Michael Dukakis or Clinton's 8.5 percent win over Robert Dole in 1996.
Maybe some cataclysmic event occurs in the next two weeks that changes the trajectory of this election, but to override these factors, it would have to be very, very big.