This was a disturbing, yet I suspect accurate, test. I came out with a slight automatic preference for white people over black people. That's not something I would be totally happy admitting, had I been asked, I think. Yet, chances are, it is true.

More happily, I had no preference on the gender test that links science and liberal arts to male and female.

Posted by: Patrick Samphire at May 21, 2004 05:55 AM

In addition to the demonstration tests, you can sign up to take a random test -- which, for me, was Apple vs. Microsoft. Interestingly, I have a slight unconscious preference for Apple, despite (or perhaps because of) the fact that I work as a Windows support technician.

I think I might respond to this kind of test in an odd way, though. In the two more socially-weighted tests (Age and Gender/Science/Humanities), my results came out somewhat opposite to the usual prejudices, which I wasn't at all expecting.

Then again, I kept classifying Latin and History as sciences and Astronomy as a humanity.

Posted by: Dan Percival at May 21, 2004 01:03 PM

Er.

"The two more socially-weighted tests that I took..."

Should have previewed before posting, sorry.

Posted by: Dan Percival at May 21, 2004 01:05 PM

I took the Gender/Science vs Liberal Arts test and came out with a moderate association of science and women. Yay!

I also moderately prefer fat people to thin people ('Cepting you, Ben, 'cuz you don't have the thin people traits I don't like). And a strong preference for young people to old people, which is the one that's making me wince.

I'll have to check a few more out.

law

Posted by: law at May 21, 2004 08:57 PM

You guys have much cooler unconscious preferences than I do

Posted by: Benjamin Rosenbaum at May 24, 2004 09:50 AM

I don't think there's much room for pride on my part...

I mentioned the web-IAT to a friend this weekend who's a recent PhD in social psychology. He remarked that the IAT is is everybody's sweetheart in the field for its easily quantifiable data, but some recent experiments have undermined its utility as a measurement -- and regardless, there's no concrete theory of what, ultimately, it measures (said my friend). He also noted that it's relatively easy to cheat (there's enough wiggle room in the timing for a cheater to consciously slow down on "wrong" answers), that the rapid-fire stimulus can cause people to unplug from the experience entirely, and that it may measure familiarity with commonly-held prejudices more than the prejudices themselves.

It's still interesting, but don't let an internet test convict you. After all, I can look around my circle of friends and see, clearly, that I have a strong prejudice towards youth, but it wasn't at all reflected in the IAT.

Posted by: Dan Percival at May 24, 2004 06:59 PM

I noticed the rapid fire stimulus coming into play when I did the second test (the gender one). It was one of my reservations about that result.

Posted by: Patrick Samphire at May 25, 2004 06:00 AM

I can well see how those factors would distort the test. Indeed, I'm not sure if "preference" is really the right word for what they're testing. What I found ominous was not just the result, but what I observed phenomenologically while taking the test.

When trying to associate "white and good" or "Jewish and good", I could feel myself coalesce them as a single concept, giving a single set of instructions to my fingers. When trying to associate the opposite pairs, I could FEEL myself fighting an uphill battle.

It wasn't that I was silently objecting to myself "but black people are bad" or "but other religions are bad", obviously. But it felt like an unnatural association, a strange and unfamiliar way to parse the world. This doesn't say much about "preference" as an individual, personal act of the will. Rather, it's a way of viscerally feeling the effect of culture. But that's almost scarier. In every instant that I look at a given human being's face, the insidious, silent, unremarked voice of culture is whispering in my ear, associating that person "naturally", almost "linguistically", with good or bad.

Not that this is news, but it's a frightening demonstration of it.

Shudder

Posted by: Benjamin Rosenbaum at May 27, 2004 09:02 AM

I did the young versus old and it said I had a strong preference for young; however, I think the methodology is off. It initially lumped old and bad together, which pissed me off. Anyway, I adjusted to this ordering but then once accustomed to one set of methodology, it was a question of mental dexterity to remember that the ordering switched. What if they initially paired young and bad? I'd have had a similar mental hurdle to leap with the switch.

I think it's a crock, but interesting to ponder. Thanks for the site.

Trent

Posted by: Trent Walters at June 2, 2004 10:16 PM

Trent, they claim they adjust the results to take account of the set order, which is random. I expect they studied precisely that question -- "What if they initially paired young and bad?" -- by comparing results with different set orders and random samples of people.

Which doesn't mean the thing is valid -- it may be based on flawed assumptions, and certainly I think the implied definition of "preference" is a complex question -- but they did at least try to address that particular methodological issue.

Posted by: Benjamin Rosenbaum at June 3, 2004 09:40 AM

Crock is a harsh term. I should have used another. Still interesting to speculate about.

Take care.

Posted by: Trent at June 3, 2004 01:08 PM