Most of us are familiar with the Stroop Test. The subject is shown a series of words, each of which is written in a different color. The only task is to say what color the word is. What’s so hard about that? Well, nothing; it would be easy, except that each word is the name of a color, and usually different from the color in which it’s written. You can try it online here, and see how surprisingly difficult it is. Believe it or not, a similar test can reveal your hidden racial biases.
The Stroop effect exists because it pits your conscious analytical skills against your native perceptions. By native perception, you see this image immediately as RED. But you are forced to use your conscious intellect to figure out the correct answer, BLUE. It’s very hard not to let the automatic impulse take over.
A related effect is laid embarrassingly bare by Harvard’s Implicit Association Test (try it for free here). IAT tests reveal many demographic biases, such as attitudes toward age, disability, race, sexuality, weight, etc. The one that I took involved racial biases. This particular IAT presents you with 24 items:
- 4 white people (each with a white-sounding name)
- 4 black people (each with a black-sounding name)
- 4 Asian people (each with an Asian-sounding name)
- 4 Hispanic people (each with a Hispanic-sounding name)
- 4 positive words (love, pleasant, great, wonderful)
- 4 negative words (hate, unpleasant, awful, terrible)
Each round of the test uses one of the four races as the target. There are two keys on your keyboard: one to use for positive words or the target race, and the other to use for negative words or any other race. The test randomly flashes either a word or a face and name (cropped so as not to show clothes, hairstyle, or other cues) on screen, and you press either the “positive” or “negative” button. If you push the wrong button, it tells you, and you can then push the right one, merely adding on to your reaction time.
It’s mainly about conscious intellect: identify the race or the type of word, and push the correct key. But your native impulse intercedes to enough of a degree that you find it easier or harder to associate certain races with your finger poised over the “positive” button. Once enough trials have been run — in the version of the test I took, each race was the target race 3 times for a total of 12 rounds of about 20-30 images each — enough data from your reaction times have been collected that a bias is shown. In my particular case, I found it easiest to associate Asians with positive words, and significantly hardest to associate Hispanics with positive words. Why? I don’t know; that’s a question for the psychologists. I’m unaware of either such correlation in my real life. Any comment I might make would be tainted by the famous Onion headline “I’ll have you know I have several black Friendsters.”
It should go without saying that little valid science is produced by a single run of this test. The idea is that demographic information is collected from each subject, and statistical trends are derived from very large data sets.
Nevertheless, taking yourself on a quick run through an Implicit Association Test can be an eye-opening experience, particularly for those who consider themselves above racial, gender, or body stereotype biases.
Update: I am alerted that a book has been recently released, Blindspot: Hidden Biases of Good People by Mahzarin R. Banaji and Anthony G. Greenwald, focusing on the results of their research using the Implicit Association Test. I haven’t read it so cannot comment.