Consequences of unconscious racism
March 26, 2016 6:51 PM   Subscribe

 
It's been posted here somewhere before, but anybody can test their own implicit biases and help researchers out over here.
posted by saulgoodman at 7:16 PM on March 26, 2016 [1 favorite]


Seems to require flash.
posted by notreally at 8:17 PM on March 26, 2016


It's been posted here somewhere before, but anybody can test their own implicit biases and help researchers out over here.

Wikipedia lists many criticisms of that test — this one jumped out at me when I took it (and when I read Malcolm Gladwell's description of taking the test in Blink):
researchers have recently claimed that results of the IAT might be biased by the participant's lacking cognitive capability to adjust to switching categories, thus biasing results in favor of the first category pairing (e.g. pairing "Asian" with positive stimuli first, instead of pairing "Asian" with negative stimuli first).
More info on that here.

From a New York Times article in 2008:
The test is widely used in research, and some critics acknowledge that it’s a useful tool for detecting unconscious attitudes and studying cognitive processes. But they say it’s misleading for I.A.T. researchers to give individuals ratings like “slight,” “moderate” or “strong” — and advice on dealing with their bias — when there isn’t even that much consistency in the same person’s scores if the test is taken again.

“One can decrease racial bias scores on the I.A.T. by simply exposing people to pictures of African-Americans enjoying a picnic,” says Hart Blanton, a psychologist at Texas A&M. “Yet respondents who take this test on the Web are given feedback suggesting that some enduring quality is being assessed.” He says that even the scoring system itself has been changed arbitrarily in recent years. “People receiving feedback about their ‘strong’ racial biases,” Dr. Blanton says, “are encouraged in sensitivity workshops to confront these tendencies as some ugly reality that has meaning in their daily lives. But unbeknownst to respondents who take this test, the labels given to them were chosen by a small group of people who simply looked at a distribution of test scores and decided what terms seemed about right. This is not how science is done.”
posted by John Cohen at 11:21 PM on March 26, 2016 [7 favorites]


I was once a substitute teacher for the Cleveland Municipal School District. One day, I was subbing for a teacher that had a class of 7th grade students. If you're a teacher, you probably know how 7th grade is pretty much the worst. It's all these kids, antsy and bursting with energy, raging hormones, really eager to show how tough and cool they are.

So here I am as a 24 year old white guy with no training in education trying to get a class of 7th graders under control. They were all black. They knew they could mess with me because I'm a substitute. And they do it spectacularly. And yeah, I definitely DID think what the hell is wrong with these goddamn *N-word*? What the fuck is wrong with them?

Later that night, I sat on my couch feeling ugly and disgusted that I even thought those words. I always thought that I was one of the good ones. I was raised to hate that racist shit. And yet there I went, trading in racist thoughts at the first sign of trouble.

That experience really taught me how deeply ingrained American racism is. It's a good thing to be aware of. I'm not perfect, but with like articles like this, I hope those slip ups can help us white people get better at true empathy.
posted by mcmile at 12:10 AM on March 27, 2016 [19 favorites]


“One can decrease racial bias scores on the I.A.T. by simply exposing people to pictures of African-Americans enjoying a picnic,”

Perhaps mainstream visual media should try doing this sort of thing more often, then.
posted by iotic at 1:02 AM on March 27, 2016 [49 favorites]


“One can decrease racial bias scores on the I.A.T. by simply exposing people to pictures of African-Americans enjoying a picnic,” says Hart Blanton, a psychologist at Texas A&M. “Yet respondents who take this test on the Web are given feedback suggesting that some enduring quality is being assessed.”

I'm a little baffled by this critique. Yes, exposing people to images which counter implicit biases is one of the easiest ways to counter implicit bias - as well as being a solid argument for why our media should be increase diversity and have a plurality of characters so we can build empathy with all people. This isn't an argument against the empathy gap or against implicit bias, though - simply an indication that both things are learned via the vehicles of white supremacy and thus can be unlearned.

Objecting to anti-racism by focusing on how racism and racist are insults rather than things to be changed in ourselves is deployed to perpetuate racism, not to counter it. Luckily, this is also a learned skill and thus can be unlearned.
posted by Deoridhe at 2:07 AM on March 27, 2016 [9 favorites]


Later that night, I sat on my couch feeling ugly and disgusted that I even thought those words.

omg this is the whitest thing I have read today.

Humans will categorize humans into groups, and then in frustration will "other" groups. Most of the time, when I'm disgusted with and "othering" a group, it's men, and then it's whites. Guess what? I'm a white male! Who the fuck cares. That was directed at myself, btw.

Moral of the ramble: get over yourself, and instead of being sad about thoughts be sad about actions, or the lack thereof. Either one does something constructive, or one doesn't. But it's not the thought that counts here. That was also directed at myself, btw
posted by special agent conrad uno at 3:19 AM on March 27, 2016 [6 favorites]


wow, yes well I just did that Race IAT test - having never run across it before and I was rather shocked at just how difficult I found it to answer quickly when the pairing was Black + Good. It was really cognitively more difficult to perform quickly, and I found myself becoming confused in a way that totally didn't happen in the White + Good pairing.

It also really drove home one of the points Frantz Fanon makes in "Black Skin White Masks" (which I happen to have been reading just recently), about the almost universal association of white and "good", "peace", "purity" etc. Fanon was right.
posted by mary8nne at 3:28 AM on March 27, 2016 [2 favorites]


researchers have recently claimed that results of the IAT might be biased by the participant's lacking cognitive capability to adjust to switching categories, thus biasing results in favor of the first category pairing (e.g. pairing "Asian" with positive stimuli first, instead of pairing "Asian" with negative stimuli first).

Having just taken the test, I agree with this. I wonder if they switch it up at all; but I definitely felt that I was being trained to do something one way, then had to switch, and that was what was being tested more than anything else.
posted by maggiemaggie at 4:25 AM on March 27, 2016 [1 favorite]


researchers have recently claimed that results of the IAT might be biased by the participant's lacking cognitive capability to adjust to switching categories, thus biasing results in favor of the first category pairing (e.g. pairing "Asian" with positive stimuli first, instead of pairing "Asian" with negative stimuli first).

Every real world test has limitations and weaknesses. That doesn't necessarily mean the test is invalid or unscientific. The Project Implicit people are aware of the task-switching issue, as acknowledged in their FAQs:
Could the result be a function of the order in which I did the two parts?

This is a very common question. The answer is yes, the order in which you take the test does have some influence on your overall results. However, the difference is very small. So if you first pair gay people + bad and then pair gay people + good, your results might be a just a tiny bit more negative than they would be if you had done the reverse pairing first. One way that we try to minimize this order effect is by giving more practice trials before the second pairing than we did before the first pairing. It is also important to know that each participant is randomly assigned to an order, so half of test-takers complete gay people + bad and then gay people + good, and the other half of test-takers get the opposite order.
And, while a given individual's single trial might be skewed by the pairing order, over the course of a large number of multiple independent randomly ordered trials, the aggregate results (in the absence of other confounding factors) will yield valid findings.

(edit: So to answer Maggie's question, yes, the second pairing is longer on purpose, and the order is randomized between trials.)
posted by xigxag at 4:53 AM on March 27, 2016 [1 favorite]


And, while a given individual's single trial might be skewed by the pairing order, over the course of a large number of multiple independent randomly ordered trials, the aggregate results (in the absence of other confounding factors) will yield valid findings.

Without test-retest reliability a test CANNOT be a valid measure of a punitively stable trait. There really are not any "if, ands, or buts" about this. Either trait isn't stable or the test is a rubber ruler.
posted by srboisvert at 5:27 AM on March 27, 2016


A test like this is bound to be extremely sensitive to context, which is going to decrease test / retest reliability. Why would it be sensitive to context? Because context is what matters to humans in terms of which group identity is most important at any given time and we're really good at switching them (as the team studies show), even if the measured brain activity remains at least to some extent "racially biased." Since group identity is profoundly important to human survival, this is going to be hard to test without being confounded by context.

I find it hard to believe that people would disbelieve in these implicit biases when they are so obvious in the behavior of people who genuinely either don't want to act in racist ways or who believe they aren't doing so when they clearly are.

I also think it's going to be really hard to change racism without recognizing that it isn't a matter of evil racists v. everyone else. We all have this in us due to the culture— no one can they that they aren't influenced by it. The question is, are we going to try to fight this tendency and how much do we try to do so, not whether we have it or not.

I think that sort of understanding would help people address their own racism because if you either are a racist or you aren't, you can't change it so you are just going to try to avoid the label. And probably avoid dealing with the issue at all, lest you be found out.
posted by Maias at 6:17 AM on March 27, 2016 [7 favorites]


Measuring people's attitudes does tend to make them think about and possibly change those attitudes. People change. That's a good thing, often. The relevant "test-retest reliability" factor here has to be at a population level, not an individual level. Kind of like measurements in quantum mechanical experiments: it's okay to measure the velocity of a bunch of particles produced from similar circumstances, and measure the momentum of a different group of particles produced from the same set of circumstances, and draw conclusions about relationships between particle velocity and momentum. Measuring both for the same particle is well-known to be, shall we say, without particular scientific merit.
posted by eviemath at 6:18 AM on March 27, 2016 [1 favorite]


The article itself goes into almost overwhelming detail about several different studies that show racial bias, and I would never argue that it doesn't exist, despite my discomfort with the IAT test. It's especially frightening when we're talking about police and the use of force.

For myself, yeah, I long ago realized that the best way to deal with these inherent racial biases was to acknowledge them and take the time to think if a particular negative reaction was because of bias.
posted by maggiemaggie at 6:31 AM on March 27, 2016 [3 favorites]


You've got to be taught
To hate and fear,
You've got to be taught
From year to year,
It's got to be drummed
In your dear little ear
You've got to be carefully taught.

You've got to be taught to be afraid
Of people whose eyes are oddly made,
And people whose skin is a diff'rent shade,
You've got to be carefully taught.

You've got to be taught before it's too late,
Before you are six or seven or eight,
To hate all the people your relatives hate,
You've got to be carefully taught!

from SOUTH PACIFIC
by Rogers and Hammerstein
posted by pjsky at 6:59 AM on March 27, 2016 [3 favorites]


"...all of us see other races as less sensitive to pain than ourselves."

Unless they're complaining about ill-treatment, in which case they are more sensitive than ourselves, i.e. "too sensitive."
posted by Weftage at 7:23 AM on March 27, 2016 [3 favorites]


Measuring people's attitudes does tend to make them think about and possibly change those attitudes.

Seems a lot of people just go "the result doesn't match my self-image, so there's clearly something wrong with the test, let's see if I can find someone on the internet that agrees with me" instead of taking it as a hint to start thinking about and observing their own behaviour in non-test conditions.
posted by effbot at 7:38 AM on March 27, 2016 [3 favorites]


Without test-retest reliability a test CANNOT be a valid measure of a punitively stable trait

I don't think they claim implicit bias is a fixed number. In fact they seem to claim the opposite. What they do claim is that the effect of order is small.

Plus, randomized blocking is a standard and well-accepted component of experimental design, in psychology and in other fields. It seems a weird thing to hang objections on.
posted by Kutsuwamushi at 8:28 AM on March 27, 2016 [1 favorite]


It's been posted here somewhere before, but anybody can test their own implicit biases and help researchers out over here.
posted by saulgoodman

Seems to require flash.
posted by notreally


It's already working!
posted by ActingTheGoat at 10:10 AM on March 27, 2016


Plus, randomized blocking is a standard and well-accepted component of experimental design, in psychology and in other fields. It seems a weird thing to hang objections on.

Test-Retest reliability is a central component of measurement validity. If your replicators can't get it up over 50% then your measure is rubbish. This is independent of the issue of randomized blocking and really takes precedence over it because it is an invalidating failure rather than a question of degrees of accuracy. Who even cares about order effects when your measure is already unreliable?
posted by srboisvert at 12:40 PM on March 27, 2016


Either the implicit bias test is wrong, or all of MeFi is wrong, because MeFi's pretty consistent in saying "Everyone's a little bit racist" and yet the implicit bias test for race said "Nope, Bugbread, you're not biased at all," and I don't even believe that.
posted by Bugbread at 9:27 PM on March 27, 2016


MoonOrb: "Is this thread really going to be the place where we sort out whether the Harvard Implicit Bias test is sufficiently reliable, or is it possible that maybe this criticism has been addressed before?"

I'm sure it's been addressed before. I'm also fairly confident that without mod intervention this thread is going to be the place where we sort out whether the Harvard Implicit Bias test is sufficiently reliable.
posted by Bugbread at 9:28 PM on March 27, 2016


"So here I am as a 24 year old white guy with no training in education trying to get a class of 7th graders under control. They were all black. They knew they could mess with me because I'm a substitute. And they do it spectacularly. And yeah, I definitely DID think what the hell is wrong with these goddamn *N-word*? What the fuck is wrong with them?
Later that night, I sat on my couch feeling ugly and disgusted that I even thought those words. I always thought that I was one of the good ones. I was raised to hate that racist shit. And yet there I went, trading in racist thoughts at the first sign of trouble."


I think it's easier to have thoughts like this about a particular group if a large number of people corresponding to a particular racial/ethnic/gender/orientation/whatever group is giving you a hard time, especially if their behaviors are corresponding to negative stereotypes. There is a certain population I have to serve at work that I have zero problems with anywhere else and wouldn't really even think about as a problem outside of work, but dealing with their particular special snowflake complications at work unfortunately makes me think not the most charitable of thoughts at times, because people of that population and their particular recurring problems that we can't fix for them due to things out of our control generally make my work life just plain harder. I'm not proud of this whatsoever, but as Avenue Q put it....we all have our moments. And then try to get over them as best we can manage.
posted by jenfullmoon at 9:38 PM on March 27, 2016


Perhaps mainstream visual media should try doing this sort of thing more often, then.

A lot of white people growing up in the 1980s, in minimally mixed suburbs and towns, nevertheless had a great model for a deeply normal, functioning, two-parent, black, upper-middle-class family: The Cosby Show.

It makes the recent revelations about the star even sadder, because that show really did made a difference in presenting positive images of a black family, week after week.
posted by theorique at 8:59 AM on March 28, 2016 [2 favorites]




Either the implicit bias test is wrong, or all of MeFi is wrong, because MeFi's pretty consistent in saying "Everyone's a little bit racist" and yet the implicit bias test for race said "Nope, Bugbread, you're not biased at all," and I don't even believe that.

Not quite everyone is a little bit racist.
posted by a snickering nuthatch at 9:09 PM on March 28, 2016


It's sort of depressing to think that the only way a person can't be racist is if they have a literal neurological disorder.
posted by theorique at 4:53 AM on March 29, 2016 [1 favorite]


I totally agree. Yet another reason to add to my list of "Why We Can't Handle Aliens When We Still Can't Handle People With Slightly Different Skin Tones."
posted by jenfullmoon at 6:07 AM on March 29, 2016


I totally agree. Yet another reason to add to my list of "Why We Can't Handle Aliens When We Still Can't Handle People With Slightly Different Skin Tones.”

Or we could be more sympathetic to aliens than we are to people with slightly different skin tones.
posted by savvysearch at 8:02 PM on March 29, 2016


Only if they're Superpeople and fictional, I'm afraid. I fear in real life we'd just have all the Republicans grabbing their guns and Trump going for the nukes if aliens show up for real.
posted by jenfullmoon at 10:22 PM on March 29, 2016


« Older “They have something to say.”   |   Maybe I should have marinated the chicken a little... Newer »


This thread has been archived and is closed to new comments