The alleged dean of the Washington press corps writes this:
The exaggerated reaction to the election among many liberals was set off by the belief that Bush owes his victory to a bunch of religious zealots bent on imposing their views on the whole society. That impression was based on exit polls showing that Bush won overwhelmingly among the 22 percent of voters who said moral values were the most important issue to them.
But as columnist E.J. Dionne Jr. has pointed out, even if he had won every vote in that bloc, Bush wouldn't have gotten close to a majority. The real Bush success was in fighting John Kerry to a near-standoff among self-described moderates.
A near-standoff? Bush lost among moderates by nine points, comparable to his eight-point loss with moderate voters in 2000. But moderates went down from 50% of the electorate in 2000 to 45% in 2004, while conservatives went up from 29% to 34%. This change in the composition of the electorate was apparently not driven by "moral values" as most people think, but it was where Bush made most of his gains in going from losing the popular vote in 2000 to winning it by more than three million votes in 2004. The self-identified moderate vote is the wrong place for Broder to look for Bush's "real success."
Update: The Wall Street Journal editorial board pushes the same argument Monday (subscription required):
True, weekly churchgoers voted overwhelmingly for Mr. Bush. But they comprised the same 42% of the electorate as four years ago. It's self-described moderates (some 45% of the electorate), who made the difference this year. Nearly half went for the President, and if opposition to gay marriage was one of the issues that stirred these swing voters, don't blame bigotry or ignorance. It's more accurate to say that proponents have overreached.
This is an incoherent comparison of answers to two different questions from the 2004 exit poll, which the Journal assumes (incorrectly) to be mutually exclusive. One question asks if you are liberal, moderate or conservative. The other asks how frequently you attend church. It's true that Bush did well with frequent churchgoers, but the converse group -- people who aren't frequent churchgoers -- are not necessarily moderates! Moreover, to say moderates "made the difference" and Bush got "nearly half" of them is absurd -- they represented a smaller proportion of the electorate than in 2000, and Bush lost to Kerry among moderates by nine points. The place to look for more useful answers is in the expansion of the conservative electorate, as I argue above -- Bush probably did well with conservatives who are not frequent churchgoers.