Clay Hambrick ([info]dclayh) wrote,
2006-02-07 21:29:00
  • Previous
  • Memorize
  • Share
  • Next

Find out how biased you are

Take the Implicit Association Test. Here's a Slate story on it.

The most disturbing part of the test is that you can see what's happening while you're taking it, at least for moderate or strong biases. I scored as strongly preferring white people to black, moderately preferring straight people to gay, moderately associating men with science and women with humanities (I also kept trying to put "biology" under "humanities"—thanks, Mudd!), and, most bizarrely, slightly preferring George W. Bush to Lincoln.

I have a couple of quibbles with the methodology, but I think you all should try the tests first and see what you think.

  • Post a new comment

    Error

    switch

    Your IP address will be recorded 

     
    Help
  • 8 comments

[info]sophonax

February 8 2006, 19:08:55 UTC 6 years ago

I would recommend not taking the tests one after another; the results seem to get more skewed the more you take. You know, not that I'm questioning your undying devotion to Bush or anything.

An interesting phenomenon that occurred in the gender/science test was that the "male"- and "female"-associated words spanned a range of ages. If a science-related word popped up immediately after an older female-associated word like "grandma," I'd have a harder time telling my brain which category to assign it to. This effect wasn't nearly as pronounced with the younger-sounding gender words like "boy," "girl," "son," and "daughter."

The most disturbing part of the test is that you can see what's happening while you're taking it, at least for moderate or strong biases.

Absolutely true; that was freaky.

[info]dclayh

February 9 2006, 05:37:12 UTC 6 years ago

Actually I took the tests in exactly the opposite order from how I list them above—but see my comment below about the ordering.

[info]sithjawa

February 8 2006, 23:58:21 UTC 6 years ago

I don't believe that they correct for order enough, though I expect it's hard to come up with a standardized correction for order because different people would need a different degree of correction.

I only took two, but the order was the same on both (the socially expected bias was tested first). Did you find that the order varied?

[info]dclayh

February 9 2006, 05:36:04 UTC 6 years ago

Yes, the order disparity is my main problem with the test as well. I could feel myself getting better as I became more familiar with the words. I remember 3 out the 4 I took having the expected bias first as you describe; I don't remember the last one.

[info]sithjawa

February 9 2006, 00:07:06 UTC 6 years ago

Also, I apparently think that Music is a science, Astronomy is liberal arts, and I have difficulty categorizing Latin.

Which actually makes sense.

Music is a science because it's like math. Astronomy is liberal arts because you have to take it down at Pomona. Latin is a science because it's used for naming things, but a liberal art because it's a language.

[info]squirrelloid

February 9 2006, 22:49:14 UTC 6 years ago

I had a problem not with time, but remembering which sides went with which things. I don't think the test actually tested anything, cause i just made a lot of stupid mistakes, and seemingly on both sides, the longer the test went on. The last section (with everything backwards from the first time) i made a lot of mistakes solely because i couldnt remember which side things were on anymore. (I also made some mistakes due to keystrokes not registering, all also late).

Yeah 'moderate preference for light skintones' meaning 'i can't remember which finger good is on anymore'

[info]squirrelloid

February 9 2006, 23:00:02 UTC 6 years ago

Ok, 2 more tries got me slight and slight (in the light direction), i'm not sure why its so consistent since i've figured out why i have so much trouble - i can't compute whether a word is 'good' or 'bad' especially fast, and have a tendency to hit the same direction as i hit the previous time if i'm being indecisive. Like, i match faces much faster and with less error because i can identify color faster than meaning. Most of the time i've hit a direction on words before i've actually understood what the word was. (Or the words take longer to assign, which is going to end up penalizing me). So yeah, timed tests that require you to comprehend words? not so hot for revealing bias.

[info]squirrelloid

February 9 2006, 23:01:00 UTC 6 years ago

Oh, note that it uses the just faces sections as its 'baseline' i think, which means its penalizing you for comprehension time.
Create an Account
Forgot your login or password?
Facebook Twitter More login options
English • Español • Deutsch • Русский…