According to the IAT, I have a slight preference for black people over white people. However, the IAT found I have a moderate preference for Rick Santorum over Herman Cain. This proves to be very interesting to me because Rick Santorum is white and Herman Cain is black. I don't completely understand how they can come to this conclusion. The way the system works is to press "K" whenever a picture of the specific subject or a "good word" (love, friend, success, etc.) appears, and to press "D" whenever the other subject or a "bad word" (hate, terrible, enemy, etc.) appear. I believe they judge your beliefs off of how long it takes you to answer and how many errors you make in each set. Their results would conclude I answered the questions faster when "black people" or "Rick Santorum" was mixed in with "good words". These results seem absolutely ridiculous. I don't believe it is accurate to assume I have a preference for either of these subjects based off of me making more or less mistakes when seeing quick images appear on screen. I believe many people have preferences/prejudices over liking white or black people more, but I don't believe this test is the one to prove it. I don't think the IAT is very credible but it is funny to see what this data reveals.
I got "little or no preference" as a result, which I think mirrors my (conscious) opinion fairly well (I am neither a Caucasian nor an African American.). But I'm not sure if the test represents people's real unconscious. In other words, I doubt the test's validity. Since I can recognize both black people's and white people's faces quite well and fast, all I needed to do was just sorted them into the right category as fast as I could. Or there might be some confound: intelligence, ability to react fast to sensation, etc.