Several others around the web have gone through the polling data so far and it seems that the easy answer was not completely correct. As Kevin Drum discovers over at Political Animal, terrorism and Bush's projected image of the economy may have made the major points.
A few other folks that come to a similar conclusion: Slate's Paul Freedman, The Time's David Brooks (kind of, he's also pushing the out-of-tough-liberal line), Matthew Yglesias over at Tapped. These are just a few going around.
The "values" points may not have been the silver bullet, but it may have been the final straw.
Update: The folks over at Donkey Rising (analyst extraordinaire Ruy Teixeira's blog) have come to similar conclusions, via Laura Rozen's War and Piece.
An analysis of the results of last week's election indicates that the presence of gay marriage referenda on the ballot had no effect on the outcome of the presidential election at the state level.
There was a very strong correlation between President Bush's share of the vote in 2000 and his share of the vote in 2004 across all 50 states and the District of Columbia. The president consistently ran a few percentage points ahead of his showing in 2000, but he did not improve on his 2000 performance any more in states with gay marriage referenda than in other states. In 11 states with gay marriage referenda on the ballot, the president increased his share of the vote from an average of 55.4 percent in 2000 to an average of 58.0 percent in 2004--an improvement of 2.6 percentage points. However, in the rest of the country the president increased his share of the vote from an average of 48.1 percent in 2000 to an average of 51.0 percent in 2004--an improvement of 2.9 percentage points.