One important point before we move on to the demographics - Obama rarely if ever underperformed his polling going into a primary. He did very much underperform early exit polling. Thus, if we are looking for a Bradley Effect, it will show up in raw early exit polling, not in underperfomance of pre-election polls. And the reluctant respondent phenomenon is very much about white Democrats, not white Republicans. Indeed, if anything, the Bradley Effect would have been something we would see in the primaries, not in the general election. Whites just vote less for Democrats period.
Remember how in the primaries we could pretty much peg the result based on educated guesses on the demographic turnout?
We were able to do this based on the assurance that Obama would take 92-95% of the African American vote, about 35-38% of the white vote (this number would vary based on region - i.e in the South Obama would rarely exceed 30% of the white vote, and age - younger white voters would vote for Obama heavily) and about 38-40% of the Latino vote. At that point, we just needed to feel confident on the projected turnout.
For the general election, you can feel quite confident that Obama's low among African American voters will be 95% of the vote and you can also expect a significant uptick in African American turnout. Thus, while John Kerry only won 88% of the A-A vote and A-As only comprised 11% of the electorate; by contrast Obama will win at least 95% of the A-A vote and A-As are likely to comprise at least 13% of the electorate. That alone gives Obama a 2.5% increase over John Kerry's total in 2004.
But of course there is more favoring Obama. He is winning Latinos by 2-1, while John Kerry only won Latinos by 53-44. Further, Latinos only comprised 8% of the electorate in 2004. It is fair to project that Latinos will be 10% of the electorate in 2008. These shifts provide Obama with another 2.2% over Kerry's total.
The final piece of the puzzle is the increase in young voters. John Kerry captured 41% of the white vote in 2004. If for no other reason than the larger share of younger voters, Obama seems assured of exceeding that total. In 2004, 17% of the electorate was voters 18-29 and Kerry won that group 54-45. This group will likely top 20% of the electorate in 2008 and Obama will capture close to 2/3 of this group. This should be worth at least a couple of points for Obama.
Thus even if Obama does not improve on Kerry's performance among voters older than 30 (which seems highly unlikely in my opinion), Obama should win the election by more than 4 points. Since he is likely to improve on Kerry's performance among all groups - the margin is likely to be more than that. Personally , I am picking a margin of 6 points.
By Big Tent Democrat, speaking for me only
