Over the last couple of months, I've spent a good deal of time looking at exit polling data pertaining to the Presidential nomination, especially on the Democratic side. Since the media story has largely been one of "personality" (to which I'd ascribe most demographically-related voter considerations), why have exit pollsters not queried voters on gender, racial, and religious motivations for the way they voted? The demographic data on the voters themselves are definitely useful, but it'd be interesting to know what percentage of voters report they voted for Hillary because she's a woman, Obama because he's black, or Edwards because he's white.
Polls consistently show much of the country is uncomfortable voting for a Mormon, with less opposition to a female, less still to a Hispanic, and even less opposition yet to a black. A USAToday/Gallup poll from a year ago phrased the question in a satisfyingly direct way: If the candidate is nominated by your party and is otherwise qualified, would you vote for the person if he's black, a woman, etc? The percentage of people who wouldn't vote for the following:
Black - 5%
Woman - 11%
Mormon - 24%
Other polls have found the level of opposition to candidates in their seventies and older falls between that of a woman and a Mormon, and the level of opposition to a Hispanic falls between that of a black and a woman.
I suspect the reason the exit pollsters didn't ask the question initially (and continue not to) is because what it would likely reveal wouldn't fit the script. It would require not just asking about the oppressed candidates, but also the oppressive ones. That's something polls like the one excerpted from above have been able to avoid (although it would certainly be interesting if they included an item gauging what percentage of people wouldn't vote for a white; it would likely be higher than the opposition to a black is).
Would (that is, have) more people vote(d) for Hillary because she's white or Obama because he's black? The Good people don't want to know the answer.