Selected exit poll comparisons, 2000-2004-2008

Politics, Presidential Elections, US Elections, US Politics


Share of voters: 48% in 2000; 46% in 2004; 47% in 2008.



Share of voters: 52% in 2000; 54% in 2004; 53% in 2008.

Compared to John Kerry’s vote, Barack Obama gained about equal ground among both men and women. But compared to Al Gore’s performance, Obama gained much extra ground among men, but little among women.



Share of voters: 39% in 2000; 36% in 2004; 36% in 2008.



Share of voters: 42% in 2000; 41% in 2004; 39% in 2008.

The same distinction noted above is even more apparent among white men and women. Obama won 4-5 points among white men compared to both Gore and Kerry, but won only 2 among white women compared to Kerry, and actually did less well than Gore did. Turnout among white women was also weaker in proportion to turnout among white men than it was in 2004 (i.e, it was still higher, but less so.)



Share of voters: 10% in 2000; 11% in 2004; 13% in 2008.

Speaks for itself. Note also the effect of the high turnout on the share of black voters in the electorate.



Share of voters: 7% in 2000; 8% in 2004; 9% in 2008.

Obama’s surge among Latinos this year (who said Hispanics would never vote for a black man?) has pushed the Republicans back to pre-2000 levels of support. On a side note, Latinos were among the very rare groups where the Nader candidacy still registered in 2004, possibly thanks to his VP candidate Peter Camejo.


AGE 18-29

Share of voters: 17% in 2000; 17% in 2004; 18% in 2008.

No sign of the expected surge in turnout among young voters this year; but those that did vote, voted for Obama in much greater numbers than their peers had done for the Democratic candidate in previous years.



Share of voters: 14% in 2000; 16% in 2004; 16% in 2008.

One of the rare demographic groups where McCain actually grabbed a larger share of the vote than Bush ever did.



(This one took some recalculation for 2000, as the exit polls then only broke the vote down into six more specific income groups.)

Share of voters: 47% in 2000; 45% in 2004; 38% in 2008.

The decreasing share of the vote this group makes out does not need to signal lower turnout; $50,000 just isn’t worth as much anymore as it was in 2000, so fewer people make less than that.

Obama did much better than either Gore or Kerry did among these working class voters, and partially that’s of course related to the larger proportion of African-Americans and Latinos in this group. But Obama did not do badly among poor and working class white voters either. He received 47% of the vote of whites with less than a $50,000 income, versus 43% of those making more than that. (Unfortunately there’s no comparable data for 2004 and 2000).


INCOME $50,000-$100,000

(This one took some recalculation for 2000 and 2004, as the exit polls then did not include this income group as one single category.)

Share of voters: 38% in 2000; 37% in 2004; 36% in 2008.



Share of voters: 15% in 2000; 18% in 2004; 26% in 2008.

A perfect split for Obama and McCain in both the middle- and upper-income ranges – an even performance if there ever was one. It does mean that he’s gained more ground among the upper-income voters, who were the most staunchly Republican – especially in 2004.

Note, by the way, that in spite of Kerry’s elitist image, the one income group he did not lose any ground in compared with Gore was the lower income group.



Share of voters: 42%* in 2000; 58% in 2004; 56% in 2008.

* I suspect a typo here; I’m guessing the CNN 2000 Exit polls page has the numbers for non-graduates and graduates mixed up.



Share of voters: 58%* in 2000; 42% in 2004; 44% in 2008.

One of the enduring puzzles is how the partisan preferences vary greatly by income group, but are pretty even across education groups. The Democratic vote does peak at the very top (postgraduates) and bottom (no high school), but it is even across the vast swathe in between, in which 75-80% of voters fit. And so we see both Gore and Obama getting practically identical numbers among college graduates and those without degree.

Kerry’s balance was a bit off, as he did a little better among graduates. That means that Obama won more ground among lower-education voters than among higher-income voters compared to 2004. If you break the numbers down by smaller subgroups, Obama won the most among the very small group (4% of voters) who did not complete high school (+13%), and won ground pretty evenly among other groups, his gains dropping off slightly among the higher-education ones: +5% among high school graduates and those with some college or associate degree; +4% among college graduates; and +3% among postgrads.



Share of voters: 39% in 2000; 37% in 2004; 39% in 2008.

That’s right: after all the hype about PUMAs and other embittered Hillary supporters who would refuse to vote for Obama, and after all the concerns about racially resentful white Democrats, more Democrats voted for Obama than had voted for either Gore or Kerry.



Share of voters: 35% in 2000; 37% in 2004; 32% in 2008.

The same goes here: after all the hype about Obamacans who would cross over, the share of Republicans who did was the same as in 2000, and just 3 points higher than in 2004.

Note the sharp drop-off in people defining themselves as Republicans. (This might hide a larger number of “Obamacans” than the above charts suggest, as some Republicans who crossed over to Obama probably changed their self-description to Independent while they were at it.)



Share of voters: 27% in 2000; 26% in 2004; 29% in 2008.

Independents, then, were the crucial electoral group – both because the lack of significant cross-voting within the respective party bases, and because there were more of them. And the Republicans got 3-4 points less among them than in the previous cycles.


Links to data: – Election 2000 – exit polls Election 2004 exit polls

Exit polls 2008 – President – Exit polls-



  1. Boris Doktorov  •  Jan 11, 2009 @7:36 pm

    Dear colleagues, thanks a lot for your wonderful web I am a Russian sociologist and analyze the Elections-2008. Do you have information about white voters in 2000?
    Best, Boris Doktorov

  2. nimh  •  Jan 11, 2009 @8:08 pm

    Dear Boris,

    An excellent further data resource, one I only found after writing these blog posts, is this National Journal page.

    You can use the drop-down menu to find back exit poll data by demographic group for every presidential election since 1988. The data include a breakdown of white voters since 1988. In 2000, 54% voted for Bush and 42% for Gore.

    For a further analysis of the behaviour of white voters in the 2008 elections, meanwhile, don’t forget to check out these blog posts of ours here as well:

    The red and blue states of white America in 2008: Southern whites constitute the real McCain Belt
    – a follow-up: Of two minds about the South.

  3. Richard Charnin  •  Nov 3, 2011 @9:15 pm

    Final Exit Polls are always adjusted and forced to match the recorded vote. The recorded vote is NEVER the same as the True Vote due to a combination of uncounted ballots and electronically switched votes. Gore won by 3 million True Votes; Kerry by 10 million; Obama by over 20 million.

12 Trackbacks