I agree that the stats, such as they are, say little about the relationship between educational level and votes cast for Obama/Romney among white Americans. The influential variable of race has certainly to be factored out before attempting to draw conclusions. How far it's legitimate to discount for wealth, I don't know, since wealth and education are so entangled; and I'm not a statistician.
But the limited point I made about the liberal politics of academics (or future academics) is valid; after all, it's not likely that post graduates are less white than graduates, so their greater liberalism needs a different explanation; and suggests that graduates who pursue post graduate qualifications, and follow a path more likely to lead to an academic career, are more likely to be liberals than their compatriots who leave academic life.
