Researchers at Harvard, the University of Virginia, and the University of Washington have created a system they claim can uncover hidden biases that we all carry. They have developed the Implicit Association Test, which is actually a series of tests that assess your conscious and unconscious preferences for over 60 different topics ranging from pets to political issues, ethnic groups to sports teams, and entertainers to styles of music.
Project Implicit examines thoughts and feelings that exist either outside of conscious awareness or outside of conscious control. The tests are fast and visual, and have been crafted to avoid the kind of conscious self correction we tend to do when we think we know how a test expects us to answer a question. Here is how the site describes it:
Psychologists understand that people may not say what’s on their minds either because they are unwilling or because they are unable to do so. For example, if asked “How much do you smoke?” a smoker who smokes 4 packs a day may purposely report smoking only 2 packs a day because they are embarrassed to admit the correct number. Or, the smoker may simply not answer the question, regarding it as a private matter. (These are examples of being unwilling to report a known answer.) But it is also possible that a smoker who smokes 4 packs a day may report smoking only 2 packs because they honestly believe they only smoke about 2 packs a day. (Unknowingly giving an incorrect answer is sometimes called self-deception; this illustrates being unable to give the desired answer).
The unwilling-unable distinction is like the difference between purposely hiding something from others and unconsciously hiding something from yourself. The Implicit Association Test makes it possible to penetrate both of these types of hiding. The IAT measures implicit attitudes and beliefs that people are either unwilling or unable to reportYou can participate in the program or just take the demonstration tests.
It’s really interesting stuff.
I took the demonstration versions of the two race related tests. Here are my results:
Darker Skin-Lighter Skin IAT
Your data suggest a moderate automatic preference for Dark Skin relative to Light Skin
Black-White IAT
Your data suggest little or no automatic preference for Black relative to White
What’s it mean? Apparently I’m unconsciously color-blind. More information on the race related tests is available in the FAQ.
Strong automatic preference for Judaism relative to other religions.
Good news: I’m not anti-Semitic.
Bad news: I’m Catholic!
I tested with a strong preference for Abled persons over Disabled.
I have a son with autism.
I’m torn between “duh” and “hm”.
OK maybe the methodology leaves a little to be desired:
Your data suggest a slight automatic preference for John F. Kennedy relative to George W. Bush
Yeah, right.
The problem is that the Kennedy pix were easier to pick because most of them were portraits rather than photos AND the pix of Bush made him look goofy. (well about 50% of them)
I like the concept and the old young seemed better but the Bush/Kennedy one was horrible mostly do to poor picture choices… Besides, let’s use Bush/Clinton THEN we have a real battle. 30 years after he was killed, most everyone likes Kennedy.
Feh! I strongly prefer beautiful people to rich people? Like I freaking care one way or another.
Couldn’t be more wrong in that I am utterly ambivalent one way or another. I cannot make a judgement based on beauty or bankrolls.
I agree with atomic-amish.
I don’t think the test is totally accurate. No matter how you screw up it’s going to favor a negative view towards whatever you’re being tested on. It’s designed to favor the theories behind the test, not to be completely neutral.
I took the Old/Young test and it was a crock. I focused strictly on identifying the face or word. It was strictly an identify and type response, not some magical BS. I was slower a couple of times because I started to lose focus. I guess that was because of some psycho-mumbo-jumbo that says I got tired of looking at old people.
I have done several of them. I think they are kind of interesting, but I can’t help but wonder how much being right or left handed and be able to manage the shift in columns has an effect on it.
I showed a preference for African Americans (I am white, although in college most of my college friends were black, so maybe that affected it-don’t know)
I think they are kind of interesting, although not sure how accurate they are.
I took the race test and I have a slight preference for black people. So quit oppressing my peeps you honkey casper rednecks!!!!! Fight the power!!!
Word.
It’s a black thing, y’all just can’t understand.
(btw, I know myself well enough to know that I have a slight to insignificant preference for white people, so I am dubious as to whether this is an accurate test. A slight preference for one’s own ‘identity’ would be something I would expect in any case.)
You strongly prefer not to waste your time on silly web polls, tests and quizzes.
Who knew? 😉
I tried it, and sat there waiting while the damn thing loaded… I got impatient and quit the window
You strongly prefer not to waste your time on silly web polls, tests and quizzes.
I got the same thing McGehee!!
I am strongly tempted to “participate in the program” and work against myself, as in, answer everything in the opposite, just because of the mood I am in tonight.
Not like anyone could ever discern any discrepancies, not being able to correctly evaluate any standard to begin with.
Here’s what my bad mood tells me: people “test” and evaluate when they have something to prove. No kidding, all seriousness here, it’s like, hey, answer this…t.h.i.s…w.a.y…and, presto, I told you so: I’m right.
Adding an Ivy League definition to the process does not any difference make. You still drive your own results by the very testing process designed and by means of implementation.
It was a bad day.
Parlor trick. Remember how we tried to rub our stomachs and pat our heads simultaneously? The assumptions of the IAT may be valid, but I doubt it. But even I, a layman and not a psychometrist, sociologist, psychologist, etc., see that any “preferences or biases” as demonstarted by the IATs could very well be little more than the result of “cross-wiring” a given task immediately after having gotten the subject to reach a level of expectaion and comfort in that task. Seems to me this is far more what we layman would try to describe as a left-brain/synapse and hand-eye-coordination test than one to test prejudices.
Strikes me far more readily that our IAT designers themselves assumed certain biases inherent, and developed a test which instead of revealing them, merely confirmed their own biases and assumptions. I didn’t read the FAQs and maybe this was addressed, but a method to test the reliabilty of the IAT seems more like guesswork than a reliably scientific method. Exit polling, anyone?
Also, their conclusion that, as to the race IAT, that the good/bad concepts as applied to race can demonstrate bias, must have been a given with which they started, withoput questioning that assumption’s own validity.
By the way. I answered honestly on the questionaire and performed the word and/or photo key-association fairly and without hesitation: it failed absolutely in describing me. But then I suppose they would say I was one of those self-deluded types who refused to see the truth of the test heh
I did the “weapons” one, which turned out to be association of weapons with black people. It decided that I have a “moderate tendency to associate weapons with black people.” Fair enough; I even admitted as much in the questionnaire.
It then took the opportunity to present me with a more than usually turgid regurgitation of the Party Line relative to “profiling.”
It’s an interesting notion, and probably the underlying science is valid — reflex responses are more likely to reflect real preferences than studied ones are (well, duh!). But they probably ought to do some of it on themselves before they release their system into the wild.
Regards,
Ric Locke
This test’s article is on the front page of the Wash. Post magazine on 1/23.
yes, and the Post made sure to point out that conservatives were much more biased than liberals. Proving their own bias?