HOW BUSH WON THE ELECTION….So what really made the difference in this election? Most of the attention has been on “moral values” and terrorism, and I was curious to see what the exit poll results showed. So I decided to take a look at the 2004 exit polls and compare them to the 2000 exit polls.

Details are below ? and there’s a bonus issue at the end that might actually be the most important of all in Bush’s victory. But first a caveat: this is just a suggestive, rough cut summary. I’m trying to get some idea of what happened, but I’m certainly not trying to pretend that anything I say here is conclusive. Serious analysts will be going through this stuff with a fine tooth comb later, but in the meantime take this for what it’s worth.

A Few Highlights

First off, George Bush won 48% of the vote in 2000 and 51% this year, meaning his overall support increased by 3%. So what I’m looking for are areas where his support was significantly higher or lower than +3%. Here are the things that stood out:

  • Bush’s support among women increased by 5%. This isn’t an awful lot higher than his overall 3% gain, but I’m including it because women make up 54% of the voting population, which makes this important even if the gain is small.

  • Bush’s support went up by 9 points among Latinos.

  • His support was up 7 points among those 60 and older. Apparently the Medicare boondoggle didn’t hurt him much.

  • His support was up an astonishing 10 points among those with no high school education, a traditional Democratic stronghold.

  • His support was down 2 points among gays and lesbians. No surprise there.

  • Among those who think government should do more to solve problems, Bush’s support was up 10 points. I’m not really sure what to make of this, but I guess it means that Bush really is perceived as a big government conservative.

  • Finally, his support was up by 10 points in urban areas and down by 2 points in rural communities, including a surprising 9 point decrease from residents of small towns. This goes against a whole bunch of conventional wisdom (including mine) about the growing urban/rural divide in America. If anything, it seems to have narrowed in this election.

So: support was up among Latinos, the elderly, and high school dropouts. Support was down in the gay community and in small towns. Trying to weave some kind of coherent story around this data is more than I can manage, I’m afraid.

Moral Values

22% of voters said “moral values” was their most important issue. Among these voters, 80% voted for Bush, while in 2000 voters who said “moral leadership” was a higher priority than managing government gave him 70% of their votes. Although this suggests that Bush made some inroads with this group, the 2000/2004 questions aren’t really comparable enough to draw a conclusion. So, since this is the infamous “God, guns, and gays” vote, let’s take a look at each of these individually:

  • God: Bush’s Protestant base showed up to the polls in slightly lower proportion than in 2000, and their support increased by 3 points, the same as his overall increase in support.

    Regular churchgoers voted in about the same proportion as in 2000, and their support increased by 1 point. In other words, their relative support for Bush actually decreased a bit compared to other groups.

    Conclusion: religious voters supported Bush heavily, but no more so than in 2000. What’s more, they didn’t turn out any more strongly than any other group. Religious belief doesn’t seem to have made much difference in the election.

    (The Washington Post has more on this today. As they say, Bush won 79% of the white evangelical vote, and evangelicals were instrumental in grass roots organizing for his campaign. However, it’s not clear that they were any more instrumental than in 2000. In fact, what data there is suggests that evangelicals probably showed up in slightly lower proportion than in 2000.)

  • Guns: Gun owners came to the polls in substantially smaller proportion than in 2000, and their support for Bush increased only 2 points.

    Conclusion: the gun vote was a net negative for Bush this year compared to 2000.

  • Gays: There were no questions about gay rights in the 2000 exit poll, but Paul Freedman convincingly argues that gay marriage didn’t make much difference this year, and Andrew Sullivan has some statistics that indicate the same thing.

    Conclusion: gay marriage might have made a difference, but the evidence seems to indicate that it was pretty slight at best.

Overall, despite the hype, the data seems to indicate that conservative moral value voters didn’t have any more impact this year than any other year. In fact, if anything, maybe slightly less.


This is a much harder area to judge because terrorism wasn’t an issue that was explored in the 2000 exit poll. But here’s a suggestive comparison:

In 2000, 12% of voters said “world affairs” was their most important issue, and of those 54% voted for Bush.

This year, 19% of voters said terrorism was their most important issue, and of those 86% voted for Bush. 15% said Iraq was their most important issue, and of those 26% voted for Bush.

Add those together, and 34% of voters chose world affairs of some kind as their most important issue. Of those, 59% voted for Bush.

Conclusion: obviously, these two aren’t directly comparable, but they suggest that terrorism was a pretty important issue. Compared to 2000, three times as many people thought world affairs was the most important issue in the election, and among those people, Bush gained 5 points of support. Multiply those two things together, and that’s a lot of extra votes for Bush.

The Economy

18% of voters cited the economy as their most important issue, and 80% of them voted for Kerry. So that means the economy was a net negative for Bush, right?

Not so fast. Check this out: more people think the economy is doing well today than thought so in 2000. And among people who think the economy is in good shape, a stunning 87% voted for Bush. Among that same group in 2000, only 48% voted for the “incumbent,” Al Gore. Bush apparently has done a great job of persuading people who think the economy is doing well that his policies were responsible.

Need more evidence? Among voters who say their family financial situation is better than before, 80% voted for Bush. In 2000, Gore won only 61% of their vote.

What makes this all the more remarkable is that only 32% of voters said their financial situation was better than before. In 2000, 50% of voters said that. So even though fewer people personally think they’re doing better this year than thought so in 2000, more of them think the economy is in better shape. Go figure. And virtually all of those people voted for Bush.

Conclusion: despite the conventional wisdom that the economy was a good issue for Kerry, I think in the end it was probably a net positive for Bush. He managed to convince a lot of people that the economy was in good shape even if they personally weren’t doing very well, and he convinced them that he was responsible. And those people voted for him in droves.


Based on this, my tentative conclusion is that the “moral values” vote is a red herring. It played no bigger a role this year than in 2000.

Terrorism played a bigger role, mostly by being a more important issue to a lot more people. Bush’s actual level of support among people who based their vote primarily on world affairs increased only modestly.

And that good old mainstay the economy was the most important of all. Compared to 2000, fewer people personally think they’re doing better but more people believe the economy is in good shape anyway. And Bush was overwhelmingly successful in convincing those people that his policies deserved the credit.

Interesting, no?

UPDATE: Just to make this absolutely clear: I’m not suggesting that “moral values” voters played no role in the election. I’m just skeptical that they played a bigger role than in 2000. In fact, I’ve been poking around a few other statistics since I wrote this, and I’m even more skeptical than I was last tonight.

There’s no question that evangelical Christians organized heavily and turned out heavily for Bush, but they did that in 2000 too. Basically, Bush increased his total vote by about 20% from 2000 to 2004 (from 50 million to 59 million), and if Karl Rove managed to turn out those extra 4 million white evangelicals he kept talking about, it means the white evangelical vote increased by about….20%. Bottom line: it doesn’t look like there’s anything special about their turnout.

But I could be wrong, especially since the data is thin. If I come across any interesting contradictory figures, I’ll let you know.