Remember that in our first five-day tracking sample, involving interviewing conducted April 11-15, Romney had 47% of the overall vote of registered voters, Obama 45%. Using Obama as the base, let's call that a -2 Obama advantage. (Note: We are going to provide in-depth analysis of the general election across subgroups using the very large sample sizes that accrue when we aggregate several weeks of our tracking together on a rolling basis. The examples I'm outlining here are based only on the first five days of interviewing with slightly over 2,200 interviews).
Obama's advantage over Romney among blacks is an extraordinary +89 (based on 93% of blacks who say they would vote for Obama, and the 4% vote who would vote for Romney). Obama's margin is +46 among Hispanics. When we put all "non-whites" together, we have Obama with a +60 advantage. (Again, these gaps are based on the smaller sample sizes in our first five days of interviewing, but it's obvious that Obama is doing well among non-whites, regardless of the precise numbers involved).
Of course, that leaves whites, where Obama's advantage is -20 (based on 35% of non-Hispanic whites who say they would vote for Obama, and 55% who would vote for Romney). Whites are the majority of the U.S. population, of course. Their lack of support for Obama is counterbalanced by the very high support Obama gets among non-whites.
Then there is age. In these first five days of interviewing, 18- to 29-year-olds are +17 for Obama, a clear area of strength. Those aged 30-49 years are +1 for Romney, a margin that climbs to +6 among those aged 50-64 years, and +11 among those 65 years or older. So, as was the case for race, Obama's great strength in a smaller segment -- in this case young people -- counterbalances his lack of strength among the majority.
Now let's move on to marriage -- a very powerful predictor of vote intention. Among married Americans, Obama's margin is -18. Among non-married Americans -- a group that includes single/never married, divorced, separated, living with domestic partner, and widows -- Obama's margin is +19. That's a huge pivot point -- a swing on the gap of 37 points (from -18 among married Americans to +19 among non-married Americans).
Marriage is, in fact, so important that it overwhelms the famous gender gap wherein women skew toward Obama and men toward Romney. Non-married women have a +26 Obama vote tilt. Non-married men have a +9 Obama vote skew. This reinforces the basic finding that non-married Americans are pro-Obama, and also that women are more pro-Obama than men.
Married women, on the other hand, are -15 Obama. Marital status trumps gender, in other words. Even though women, in general, are disproportionately Obama supporters, if they are married, they end up being significantly more likely to vote for Romney than for Obama. And, of course, putting both marital status and gender together produces the least Obama-supportive group of all, with married men coming in with a -22.
Finally, there is religion. There is no doubt that religiousness is a major correlate of political status in America today (and has been since at least the 1980s). Basically, there is an R & R rule in American politics: religious equals Republican. That's certainly the case in voter intentions. Among Americans who attend church weekly, Obama's position is -21. Among those who attend nearly every week or monthly, it's -5. Among those who seldom or never attend religious services, it's a +11.
Obama is, by all accounts, a personally religious man, but his political appeal skews strongly toward Americans who are not religious.
These data on religion and Republicanism include black Americans, who are anomalous in that they are both very religious and very Democratic. Looking only at non-Hispanic whites, we find that Obama's position goes from an extraordinary -46 among weekly white church attenders to -25 among almost-weekly/monthly white church attenders to -2 among whites who seldom or never attend church. So the gap swing among all Americans from weekly church attenders to seldom/never church attenders is 32 points. The gap swing among white Americans from weekly church attenders to seldom/never church attenders is a significantly larger 44 points.
From a demographic standpoint, therefore, we see that Obama's political strength is among Americans who are not white, who are not married, who are not religious, and who are young.
Just to underscore the power of these demographic variables, I put together two groups of voters (based on our April 11-15 five-day rolling average):
- Married, weekly-church-attending, whites: 17% for Obama, 73% for Romney
- Non-married, seldom/never-church-attending, non-whites: 86% for Obama, 8% for Romney