One perpetual problem in the social sciences is figuring out what people really think. For example: How much of a bias do you have against black people? You could just ask people, but their answers aren't reliable — they often lie to researchers, and sometimes they don't even realize how biased their snap reactions really are.
So in 1998, a group of scientists developed what is called an implicit-association test. In short, it asks subjects to quickly categorize images as either "good" or "bad." The decisions need to be made at a speed faster than conscious thought, and that turns out to be very, very hard when they go against the subject's instincts. The result is that the test measures racial bias in ways that are very difficult to manipulate, and that have been shown, in study after study, to predict real-world racial bias, too.
Researchers Shanto Iyengar and Sean J. Westwood used this test to figure out how people feel about Democrats and Republicans — and the results were stunning. You should read the story we wrote about that here, but you can see how you stack up by taking the test below: