In other words, almost all voters in 2004 were firmly committed to one party or the other. Bush political supremo Karl Rove was right in saying there were few uncommitted voters and that his campaign's big task was to turn out the faithful. The Kerry campaign operated on the same assumption.
But in recent years, lots of American voters, at least by historical standards, have flipped from one party to the other, and in both directions.
The conventional wisdom is that we know with certainty the identity of the dozen or so battleground states. But the list has changed since 2008.
In 2008, Obama carried Indiana and lost Missouri by only 3,903 votes. Today, Indiana and Missouri aren't on anyone's target list.
In contrast, most analysts' battleground list this year includes Michigan and Wisconsin, which Obama carried in 2008 with 57 and 56 percent of the vote.
There's another difference between 2004 and 2012 that is salient. In 2004, George W. Bush's Republican base was pretty much united on issues. Foreign policy realists and neocons were all on board.
Cultural conservatives supported the Bush tax cuts. Few economic conservatives had much problem with Bush's stands on abortion or embryonic stem cell research.
Barack Obama's Democratic base is more heterogeneous. He probably increased turnout among young voters by endorsing same-sex marriage. But he risked turning off the many black voters who are solidly opposed.
Blocking the Keystone XL oil pipeline from Canada pleased gentry liberals who embrace every green cause. But private-sector labor unions don't like it a bit.
Bashing Romney's record at Bain and Co. may be helping him with some modest-income voters. But it risks antagonizing the affluent, which is a problem for a candidate who last time ran even, 49 to 49 percent, among those with incomes over $100,000.
Every campaign cycle is different, and 2012 is more different from 2004 than many Democrats think.