homeaboutarchives + tagsshopmembership!

Implicit Association Tests

posted by Jason Kottke   Aug 16, 2004

An IAT, or Implicit Association Test, attempts to measure people’s “conscious-unconscious divergences”. Basically, an IAT puts you in a situation with no time to think and compares your reaction with your behavior when you’ve had time to think things over. On the Project Implicit Web site (click on “Demonstration”), you can take a number of IATs to get an idea of your own divergences, including factors such as age, gender, race, sexuality, and religion.

I took the Race IAT:

This IAT requires the ability to distinguish faces of European and African origin. It indicates that most Americans have an automatic preference for white over black.

and the Arab-Muslim IAT:

This IAT requires the ability to distinguish names that are likely to belong to Arab-Muslims versus people of other nationalities or religions. It frequently reveals an automatic preference for other people compared to Arab-Muslims.

On the Race test, my data suggested “moderate automatic preference for White American relative to African American” which isn’t surprising because I spent the first 17 years of my life in northern Wisconsin where there weren’t any black people and all my associations were from what I saw on the news, TV, in the movies, or from what my friends and relatives told me. Of course, I’ve since come to believe that people of all races, while culturally and physically different in some ways, are deserving of the same level of respect. The IAT reveals that though that’s what I consciously believe, my subconscious mind still shows a preference for white people. From the Race FAQ:

Automatic White preference may be common among Americans because of the deep learning of negative associations to the group Black in this society. High levels of negative references to Black Americans in American culture and mass media may contribute to this learning. Such negative references may themselves be more the residue of the long history of racial discrimination in the United States than the result of deliberate efforts to discriminate in media treatments.

On the Arab-Muslim test, a more surprising result: “your data suggest a slight automatic preference for Arab Muslims relative to Other People”. Maybe I’m one of those self-hating white American liberals the neo-cons are always talking about. Growing up, I can’t really remember any positive or negative associations about Arabs or Muslims…I wasn’t really aware of them as a culture. And even with 9/11 and the American military action in the Middle East…those events aren’t something that I associate with any particular religious or cultural group.

Even though these tests are just demonstrations, the differences in what you believe and your mind’s true feelings are important to be aware of for many reasons, even though it may be uncomfortable to know — as I am — that you prefer whites to blacks or thin people to obese people. For those who wish to balance your conscious and subconscious minds and lessen your divergences, the site’s FAQ offers hope:

One solution is to seek experiences that could undo or reverse the patterns of experience that could have created the unwanted preference. But this is not always easy to do. A more practical alternative may be to remain alert to the existence of the undesired preference, recognizing that it may intrude in unwanted fashion into your judgments and actions. Additionally, you may decide to embark on consciously planned actions that can compensate for known unconscious preferences and beliefs.

If you’re interested, you can sign up on the site to participate in the research effort.

Reader comments

eloisaAug 17, 2004 at 6:26AM

“Your data suggest little or no automatic preference for White American relative to Black American”. It may not account for much, since i’m European and you know how we all feel towards Americans, n’est-ce pas!

Just kidding. Anyways, i find this interesting since just like you i grew up in a small town with just white people- and now that i’m in Belgium, most of the black people i see are africans, either imigrants or first generation.
Even though on a rational level i believe 100% that we are all equal and should be treated as such, i would have thought that i’d be biased for whites. Interesting.

karlAug 17, 2004 at 8:08AM

You must unlearn what you have learned.

lovesFranceAug 17, 2004 at 10:45AM

Americans learnt to be politically correct, so they know how to answer questions on tests, not to sound biased. Except against French of course.

souloniceAug 17, 2004 at 12:50PM

….a slight preference for Black American …..

I’m not surprised considering my personal ethnic makeup and the fact that I had never seen more than 5 white people at one time until I went to college.

I’d be interested to see how my parents would do on such a test. They both come from small towns in the south (Georgia & Louisiana) and grew up during the 40’s and 50’s.

ZeibinAug 17, 2004 at 2:03PM

With all due respect to the older generations (on the Age test for Canada), it really threw me for a loop seeing a picture of an older gentleman right after acknowledging “Pleasure” on the same side…

KatieAug 17, 2004 at 2:42PM

What interesting tests - thanks for finding this. I am ashamed to admit that my result also showed a “slight preference for white americans” while I’m supposedly not ageist at all.

My election 2004 results show a “moderate automatic preference for Kerry,” but I was surprised to see that there were no percentages listed for other users’ results as was included with the other tests. This could just be due to the third variable of Nader, but it still seems possible to list percentages for Bush vs. Kerry.

essAug 17, 2004 at 3:04PM

I always show a slight pref for black on these things - and I suspect that perhaps I have a larger or different vocab or faster fast-twitch muscles or a much lower IQ than the baseline group. Some other quirk must be driving these results.

Not that I want to have a bias against African-Americans or anyone else, just that as a typical nomially white person in the US, you’d think my results would skew toward the ofay.

I’ve always wanted the people who create these things to create one that would predict food preferences, and then test their results against the subjects known food preferences.

Anyone remember when Hank and Peggy Hill did this?

MattAug 17, 2004 at 3:36PM

Don’t bank too much on these things folks. They’re internet sideshows, not automatic guilt vendors.

KatieAug 17, 2004 at 3:41PM

Thanks, Matt, that helped. 8)

RaphyAug 17, 2004 at 3:49PM

It’s not my intent to be controversial, but must one feel “ashamed” with a result that shows a “preference” to one’s own racial group? I don’t think the answer is black and white, so to speak.

The test is interesting nonetheless.

Paul GriffinAug 17, 2004 at 4:06PM

My problem with these sorts of things (much like personality tests) is that I quickly get bored with just taking the test, and start forming hypotheses about the inner workings and trying to deliberately skew the results toward a preconcieved outcome to test the validity of my assumptions. I guess I’m just predisposed to be far more interested in how the test works than I am in the actual results of the test.

nathanAug 17, 2004 at 4:24PM

“you show a high preference for not bothering to enter fake data into an awful web registration form, especially when bugmenot.com is down.”

Charlie ParkAug 18, 2004 at 8:55AM

Somebody must have gotten his preview copy of Blink. Hope you’ve enjoyed it.

Jared WhiteAug 20, 2004 at 6:45PM

I’m White, a Christian, and a conservative. You’d think that I’d have some kind of underlying mistrust of Arab Muslims, right? I was hoping I’d at least be neutral, but I was pleasantly surprised to see the test results come back as me having a moderately strong preference towards Arab Muslims. Hmm…maybe it’s because I recognized their names faster. :) I did put in the form that my beliefs require me to be initially warm to all types of people, so I’m happy that perhaps my subconscious really is following that path.

I don’t know if these tests are all that useful, but they’re interesting nonetheless. Thanks for the link.

This thread is closed to new comments. Thanks to everyone who responded.