The Implicit Association Test (IAT) is a psychologicaly test supposed to show unconscious biases and beliefs. It works by having sit at a computer and sort words or pictures into two groups, each of which is a combination of two groups. For instance, a common test is the race specific IAT. One group would be “White or Bad” and the other group would be “Black or Good”. The words that the subject is asked to sort in this test would go something like this: Martin Luther King; Hurt; Pain; Bill Clinton; Candy; Bill Cosby; Sunshine. . .Etc. Each word clearly belongs to one group or the other. The words are flashed on a computer screen one at a time, and the subject must hit one of two keys to sort the words.
Now, supposedly, the amount of time invloved in sorting the words is supposed to demonstrate underlying biases, so for instance, people are supposed to sort faster when “Black” is associated with “Bad” than when it is associated with “Good”. And that’s what the test reveals.
My beef is with the interpretation of the test. The test is generally interpreted to mean that the test takers are latent racists, that they have negative associations that make it more difficult (i.e. more time consuming) to sort against racial stereotypes.
But wouldn’t it also be expected that someone who doesn want to be seen as a racist would perform worse on a test that might reveal her to be so? Wouldn’t the extra stress that causes also slow down her reaction time, because she must not only sort the words, but then check to make sure she’s avoiding doing anything offensive?
One might suggest that this possibility can be dismissed based on the fact that people perform faster on test that are equally loaded, but where the groups are “White or Good” and “Black or Bad”, that is, the groups are in line with the supposed stereotypes. But in this situation, the potential for offense is decreased greatly. A mistake would actually make the subject appear less racist, so the preasure is off. Hence, the performance improves.
I’m not saying that this is all that is going on, but I doubt highly that it is not at least a part of what is going on. I suppose the best way to find out how much the fear of negative social perception affects the outcome of the test would be to compare it with more socially benign groups like “Apples or Yellow” vs. “Bananas or Red”.
Thoughts?