Friday, January 27, 2006

How (whatever)-ist are you, really?

The problem with surveys to determine how biased we are is that we are smarter than that. We know the right answers to all those questions about how we feel about races, genders, religions, ethnic groups, disabled or overweight people, etc. In short, we answer questionnaires as we know we should feel, not necessarily as we do feel.

The Implicit Association Test gets around that by operating on the basic principle that your mind works faster when you are doing less mentally straining tasks. On the surface, these tests are all one task: sorting words or phrases into two groups, and to make it even easier, they tell you the groups up front. For example, the test for skin tone preference has you sorting drawings of light-skinned faces from dark-skinned faces, and positive words like "happy" and "laughter" from negative words like "hate" and "pain." If you end up sorting faster when asked to put light-skinned faces in the same column as positive words than when asked to put dark-skinned faces with positive words, you have an implicit bias toward lighter skin because your brain has to think a little more to pair up dark and good. It's very hard to fool the test.

On the ones I have done so far, I show a strong preference to women being in the home and men in the workplace, am pro-Semitic, and show a slight preference for lighter skin tones. No real surprises there, except that I didn't think my housewife bias was as strong as it apparently is. More telling than anything are the tests I am reluctant to try because I'm not sure I want to confirm what I suspect are the results.

No comments: