John Hood exposes some of the myths of the Presidential election:

[T]he notion that Bush won primarily because religious voters turned out for him does not seem to be backed up by any real evidence. Few reporters or commentators appear to have gone back to examine the 2000 exit polls, which would seem to be necessary if one wishes to assert a trend.

I did. I found that the percentage of voters sampled who said they attended church at least weekly was the same—42 percent—in both 2000 and 2004. The percentage never attending church was also the same, at 15 percent. The middle group, those attending occasionally, was, you guessed it, 42 percent each time. Interestingly, while Bush slightly improved his standing among frequent churchgoers, by about a point in 2004, his support grew by 3 to 4 points among those attending seldom or never.

Yep, it was the atheist vote that really put Bush over the top in 2004.

...OK, what about issue positions? In 2000, about 40 percent of voters in the exit poll said that abortion should be mostly or always illegal. In 2004, it was 42 percent. Not exactly a huge jump. And we don't know how many of those are single-issue voters on abortion. Both parties have significant minorities who disagree with the official party position: about a quarter of pro-lifers voted for Kerry, while around one-third of pro-choicers picked Bush. On same-sex marriage, the issue was not polled in 2000 so it is impossible to say with certainty how the two electorates compare, but it is unlikely that this year's voters were significantly more conservative on it. In fact, the public's position is more nuanced here than the insta-spin would have you believe. About as many favored civil unions but not official marriage (35 percent) as favored neither (37 percent), and Bush was preferred by both groups over Kerry.

Well, perhaps there was no national trend but it happened in selected states such as Ohio. Nope. In the 2000 exit poll for Ohio, the percentage of frequent churchgoers was higher (45 percent) than in 2004 (40 percent). Bush did win a larger majority of religious Ohio voters in 2004 than he did four years ago, but there were fewer of them proportionally. Besides, saying that the religious-vote affect mattered in a few key states changes the nature of the media spin, which has been trying to assert it as a sweeping national "explanation" for Bush's popular vote.

...[Y]ou can track the impact of foreign policy over time. In 2000, only 12 percent said that "foreign affairs" was the most important issue in the presidential race, and they broke 54 percent to 40 percent for Bush over Gore. In 2004, a combined 34 percent identified foreign policy (either Iraq or the war on terrorism) as the most important, and they appear to have broken for Bush by 59 percent to 40 percent. Put it all together, and the increase in salience and small increase in Bush preference for foreign policy constitutes a gain of 13.5 percentage points in the Bush vote in 2004.

Obviously, he didn't win by that much. He lost ground on economic issues, because of the recession. But without his edge on war on terrorism, Bush would have lost. And that proposition—unlike the "it's all about gay marriage meme"—is testable and fits the available data. Voters worried about partial-birth abortion, same-sex marriage, and other cultural issues are obviously an important constituency within the current GOP majority, but they are no more responsible for Bush's national victory on Tuesday than voters motivated by other issues to re-elect the president.