I believe that Obama will win fairly easily. I never really thought Kerry would win. The odd thing is that though my gut-feeling is 180 degrees different there is not a gigantic numerical difference between the two so far.
In August 2004 the Gallup tracking poll had Bush up by a few points for most of the month while most other polls showed Kerry leading. (Gallup was flamed every day, though they tuned out to be more accurate than some favored polls, like Zogby.)
There were a lot of polls in August 2004, and they were all over the place:
"This month, a Zogby Poll has Kerry up 7 percent; Gallup has Bush up 3; Time has Kerry up 7; and CNN/USA Today has Bush up 4." ("Electoral College holds key to election, again"
By John Baer. Knight Ridder Newspapers 8/23/04 http://www.arbiteronline.com/home/index.cfm?event=displayArticlePrinterFriendly&uStory_id=e56d7142-7819-44cb-8900-9d13959fd1da)
Despite the uneven national polling, Kerry was romping in electoral vote projections:
"Two Web sites, one pro-Kerry, one pro-Bush, are tracking state polls, then assigning corresponding electoral votes and showing results on color-coded maps. Both show Kerry way ahead. On the pro-Kerry site, www.electoral-vote.com, for example, Kerry has 317 electoral votes (270 are needed to win) and Bush 202. On the pro-Bush site, www.electionprojection.com, Kerry has 327 and Bush 211." (ibid.)
The point here is not that Obama will not win. He will. He may even win in a landslide. The point is that John Kerry was very strong in electoral college projections four years ago, but not so strong in national polling.
People like to say that national polling is meaningless and you need to look at the states. That argument is beguiling because it is true in principle... the president is indeed picked by states, not popular vote. But state polling is, for whatever reason, more variable than national polling.
And on election day the popular vote and electoral college count tend to match. I think three presidents have been selected while losing the popular vote. (And since more Florida voters intended to vote for Gore than Bush in 2000, an *accurate* electoral college count would have matched the popular vote in 2000.)
CONCLUSION: Obama is a prohibitive favorite (2:1), yet the race is incredibly close. It's a paradox. Both are true. I think it is because everyone assumes Obama is in the challenger role, and thus the almost unprecedented numbers of currently undecided voters will break for Obama disproportionately, as is typical of elections with an incumbent. But McCain isn't an incumbent, so it's hard to say. Everything I know about politics points to an Obama win, but most numerical measures people point to *today* to show that are historically unreliable.
This is an unusual election.
_________________
DISCLAIMER: There is a sense here that cheer-leading on the internet wins elections. That is the "Save Tinkerbell" theory the Republicans used to make the Iraq war such a success. (If things aren't going well it's because people are not clapping hard enough.)
Democratic Underground has existed for only one presidential election so far. Everyone was encouraged to say Kerry would win, and was flamed for suggesting he might not. So the historical track record for the effect of cheer-leading on DU is 0-1.
Did Kerry lose because DU encouraged complacency by stating he would win? Or did Kerry lose because DU wasn't enthusiastic enough... didn't do enough to visualize victory?
I would suggest a third option; that posts on DU did not decide the 2004 election one way or the other. Since DU cheer-leading probably does not win national elections, or lose them, it is best for people to prognosticate based on their most sober analysis, not based on whether their observations will be applauded or flamed.