Test your Hidden Bias

January 3, 2003 10:29 PM

Test your Hidden Bias. Tolerance.org has a set of Java-based tests designed to shed light on personal hidden biases w.r.t. race, gender, sexuality, and body image. Your results may surprise you. See also this New York Times article mentioning these tests and more rigorous studies.
posted by tss (37 comments total)

I don't know how to put this, but this isn't the first time this has appeared here in the last year.
posted by y2karl at 10:54 PM on January 3, 2003

yep.
good link none the less.
posted by Espoo2 at 10:56 PM on January 3, 2003

at least this seem to be a better method of assessment than the silly Quizilla tests.
posted by supershauna at 11:01 PM on January 3, 2003

I took the Native American test: This test probes for an automatic stereotype that Native Americans are not as "American" as White Americans.

My results were along the lines of, you show a strong association between white americans and America.

I don't deny this. I do not see native americans as being less american I just see less of them. Of course I am going to associate white americans with America, they are the overwhelming majority.

I assume they only mean North American because Peru was considered foreign in the test.
posted by Recockulous at 11:26 PM on January 3, 2003

I am not sure that the version of these tests given on the web tests anything but muscle memory...

In both that I tried the concept toward which one is supposed to be biased (Native Americans are not Americans and Gay People are bad, in my cases) is set up in several quizzes, and the alternative, unbiased concept in fewer, given at the end.

So the version given online is more a test of how many video games you've played. All this really tests is how quickly can you associate new ideas with learned key-punching motions.

Sounds like the tests for the actual study were more elaborate, however.
posted by 23lemurs at 11:27 PM on January 3, 2003

I agree with 23lemurs in this instance...I found myself using typical motor learning mechanisms, such as speaking the associative term I wanted my hand to go to.

Sort of humorously, though, at the beginning I thought it would be more "trick-question" - y: I deliberately chose "straight" for the first appearance of the two brideish women, figuring that it really could go either way. I was assuming that they would test my tolerance by showing me images that many would associate with homosexuality, such as two men holding hands.

That said, I apparently have a "slight bias" towards gay people.
posted by babylon at 11:41 PM on January 3, 2003

Your data suggest a slight automatic preference for Gay people

Uh oh.
posted by bobo123 at 11:43 PM on January 3, 2003

I didn't get that test at all. When they showed me Indians, I pressed Native American every time. What the hell else was I supposed to do? I should have scored 100%. {glances side to side}
posted by Witty at 11:56 PM on January 3, 2003

y2karl: your response shows a hidden bias against repeat postings.
posted by alms at 12:23 AM on January 4, 2003

I have an extreme bias, one might even go so far as to say a blood-curdling irrational abhorrence, an execration, a repugnance if you will, to the term "hate crime".
posted by hama7 at 1:16 AM on January 4, 2003

Ah, but no one could accuse you of having hidden biases, hama7. You air them so nicely for us!

Kidding friendly-like. The best I can say for my results is that I'm not quite as homophobic as I am racist.
posted by hippugeek at 1:29 AM on January 4, 2003

Uhmn, I took the Arab prejudice test and it told me that I have a "strong automatic preference" for other people -- which is way off. I mean I like "middle-eastern" food, but I really have a strong automatic bias against all people. This clearly means that the software they are using needs an upgrade.
posted by stvc15 at 1:42 AM on January 4, 2003

I did the gay test, and I kept getting confused about the wedding pictures. Instead of thinking (the way I guess I was supposed to) that the picture with one person in a white dress and the other person in a tuxedo was heterosexual, I assumed it was either a butch-femme lesbian wedding or else it was two men, one of whom was a crossdresser.
posted by faustessa at 3:00 AM on January 4, 2003

Your data suggest a strong automatic preference for darker skin

Your data suggest a strong automatic preference for Gay people

Surprising on both points. But does a 'strong preference' mean it informs behaviour...? what point are they making?... must admit I tried to read thru the notes but my head got foggy (my strong automatic reaction to somewhat opaque text).

...black gay men with GSOH apply here.
posted by keno at 3:29 AM on January 4, 2003

It's already been said, but my big problem with these tests is, if you cannot choose associations (or if you do all you get is an agressive red cross reminding you that the way you see things is just wrong), isn't the test already playing wiht prejudices?

And to push the controversy a little further, isn't the tests themselves are biased ? Why even wanting to test bias, finally? It just enlightens the fact that people do discriminates before evn starting.
posted by Sijeka at 6:14 AM on January 4, 2003

There is bias in these bias tests! I agree with Sijeka, these tests are structurally flawed at best. empirically I may out the word "HOMOSEXUAL" on the "BAD" side, but is that because I think "HOMOSEXUAL" is bad or is it because "BAD" is on the same side as "GAY" was in the first case of the test?
posted by Quartermass at 7:28 AM on January 4, 2003

I'm a bit curious about how these tests work on a "scientific" level. I put each item into the correct column (aside from one when I accidentally pressed the key twice), and it told me I had a "strong preference for heterosexuals".

In truth, I don't really give a flying flip what people do in their bedrooms. I generally don't obsess about people's sex lives, homosexual or otherwise.

I'm very dubious about these tests. This reminds me a bit of the Scientologist "personality test", where they tell you "Yep, you're crazy, come on down and will show you how to fix yourself."
posted by Erasmus at 7:34 AM on January 4, 2003

I have to add two things:

1/ It is really stressful to click rapidly on the "e" and "i", which leads your brain to a certain confusion and a quite not 'normal' state. Is it a really scientific behavior?

2/ A gay friend just did the test and he had 'slight preference for gay people'. So...that leaves me confused.

And i truly apologize for the typos in my last post.
posted by Sijeka at 10:12 AM on January 4, 2003

I agree with 23lemurs. It reminded me of the Stroop Test more than anything else.

It would be interesting to see this test done the same way, only substituting, say, images/words related to "Smurfs" for Native Americans, Arab Muslims, etc. I have a sneaking suspicion that I would come out as showing a slight bias against the Smurfs.

[And quite rightly, too, I hate those little freaks!]
posted by filmgoerjuan at 10:15 AM on January 4, 2003

The idea behind the tests is that you can more quickly sort things into combined categories if you naturally associate the categories being combined. So, if you're sorting into "cars + trucks" on the left and "puppies + kittens" on the right, you'll be a lot faster and make fewer errors than if it were "cars + puppies" and "trucks + kittens".

The assumption is that the converse is true, and that if you're faster sorting into "blacks + weapons" than into "whites + weapons", then you associate blacks and weapons at some liminal or subliminal level.

I got "Your data suggest a slight automatic association between White and weapons" which I didn't expect, although I know a lot of white cops so that may explain it. (Also, if you don't identify a 16" Mag-Lite flashlight as a weapon, you've never seen someone hit with one...)
posted by nicwolff at 10:30 AM on January 4, 2003

From the intro page to the tests: Yet reports of unequal treatment because of skin color, religion, gender, age, sexual orientation, country of origin, ability and other differences persist.

Notice the word "ability" in that list? Damn straight people treat others of varying "abilities" different. We pay Michael Jordan $$$ for his "abilities." We pay Bill Gates $$$$$$ for his "abilities." We pay Joe Six-Pack $ for his "abilities. Like the old Sesame Street song, from the list of supposed "biases", one of those things just doesn't belong. One is born with skin color, gender, age, country, and arguably born with sexual orientation (not activity, necessarily) and religion. But "ability" to me is subject to change, based on one's aspirations, education, motivation, resources, etc. and able to be changed and improved upon, and is, in fact, one of the foundations of society: we pay and perceive people differently based on their abilities, which the vast majority of people are able to improve, enhance, change, etc.

Having said all that...I find myself lately to be somewhat puzzled over the term "Native American." I am a Native of this land - I was born here. I am an American. However, my ancestors came from Europe. Nevertheless, I am a "Native American". Would it be wrong/controversial/stupid of me to suggest that those who currently identify themselves as "Native Americans" henceforth use the term "Tribal Americans," seeing as how their primary identification as individuals and as a culture is the Tribe? Would this help reduce the confustion or linguistic tangle over the terms "Indian," "Native American," etc? No hard feelings, I hope -- just trying to figure out a better way, language-wise, of thinking about this issue.
posted by davidmsc at 11:36 AM on January 4, 2003

very interesting! took the native american test. although the test showed me to be neutral in my bias, i think it's flawed. i found it much more difficult at first to associate white with foreign in the final stretch of the test. the reason i came out neutral in the end is that i was able to very quickly "re-learn" the association while i was taking the test (basically, i just stopped for a few milliseconds after getting a bunch wrong and began to think of the pictures in terms of natives and foreign invaders instead of in terms of citizenship). so i failed to match most of the first set of images while i aced the later sets.

a number of interesting questions come to mind:

- is the concept of "subconcious" bias bogus?
- is the test?
- are some people more able to change their associations than others?
- how long will my changed association last?
posted by muppetboy at 11:39 AM on January 4, 2003

nicwolff - Thanks, now I understand what the test is trying to tell me. I am open to the notion that I have a "strong automatic association between White American faces and American," but I'm concerned the first combined trial had Native American and American together, and I stumbled badly trying to understand what they were getting at. Unless that trial was discarded I'm sure it skewed my results badly.

I 'spose I could take the test over and see, but, alas, I have the attention span of a web surfer and must now go find other links to click.
posted by chipr at 11:39 AM on January 4, 2003

davidmsc, the "ability" in that kind of writing generally refers to whether or not a person is physically disabled. It's a shorthand way of refering to it, but my guess is that's what they meant, rather than making judgments between people with different amounts of skill or aptitude.
posted by Chanther at 12:04 PM on January 4, 2003

Would it be wrong/controversial/stupid of me to suggest that those who currently identify themselves as "Native Americans" henceforth use the term "Tribal Americans," — davidmsc

Yes, it would. They were native when the Europeans got here, and will always be the most-native Americans (even if they are not truly autochthonous); and that's what distinguishes them as a group, so it's the simplest and most applicable title for that group. You're a native American, but not a Native American.

On preview: 0 for 2!
posted by nicwolff at 12:40 PM on January 4, 2003

This test makes no sense. I understand there is racism everywhere - I was even the one in college who was convinced that Brave New World had racist undertones, even though no one else did. However, what does this test tell you? Absolutely nothing.

Biases mean nothing. Actions are what mean something. On the news, you see more black people with guns, so I assume one might, at a subconscious level, make the association. However, I think very few here would say that they're afraid of black people. Likewise, American is typically associated with apple pie and Ms. Cleaver. Does this mean I don't think of Asian-Americans as American? Of course not.

But what do you expect from an organization that, as noted, considers "discrimination by ability" to be a bad thing. It reminds me of a Vonnegut story I once read...
posted by Kevs at 1:00 PM on January 4, 2003

It reminds me of a Vonnegut story I once read...

Oh God, Harrison Bergeron gave me nightmares for weeks after I read it in Junior High....
posted by bradth27 at 1:18 PM on January 4, 2003

Crashes in mozilla.
posted by elwoodwiles at 1:40 PM on January 4, 2003

Biases mean nothing. Actions are what mean something.

Biases almost always inform actions, especially if you have a broad definition of "action". First impressions, for example, can be self-reinforcing -- if you're an asshole to someone because your brain automatically files them under "dangerous criminal" after seeing the color of their skin, they'll likely be an asshole back. If you're a teacher, you have to make decisions about how to treat students based on their intelligence and maturity; if bias colors those observations, minority students can be systematically shortchanged throughout the school system, internalizing the expectations of their teachers. Feedback loop, etc.

If you have biases, try to eliminate them. Almost everyone has biases; you're not being singled out, and there's no need to get defensive.
posted by Tlogmer at 3:57 PM on January 4, 2003

I also agree with 23lemurs. All of my "mistakes" were due more to fatigue and the fact that I have terrible eye-hand coordination more than anything else. I suck at video games.

If you have biases, try to eliminate them. Almost everyone has biases; you're not being singled out, and there's no need to get defensive.

Actually, I take issue with that. Of course everyone has biases, but then, if that's so, how are we expected to get rid of them? The reccoemndation of the folks giving the test is not that we get rid of our biases, but that we trade them for s new set. Favoring tolerance is a bias, after all. Morals are a bias.

And sexual preference is a bias.

it told me I had a "strong preference for heterosexuals".

Huh. Me too. Fact is, I prefer to have sex with other heterosexuals 100% of the time. I also don't find androgeny or corss-dressing sexually attractive. I also do not care what people do in thier bedrooms, so long as all parties consent. But I don't find a lot of that stuff inticing or even interesting. Does that make me a bigot or just straight?

The assumption is that the converse is true, and that if you're faster sorting into "blacks + weapons" than into "whites + weapons", then you associate blacks and weapons at some liminal or subliminal level.

The problem I have with that is that they "trained" people to respond one way and then switched things up - beginning with the "bad" way first. Why not begin with the "good" way and then switch?

Then, too, I question the statistics. I was paying attnetion - I'm pretty certain my mistakes were about equal in all directions - and my most common mistake was classifying Niagra Falls as Foreign. I only got that one right once (I associate it with Canada.) So does that mean I'm biased for or against Niagra Falls? Canada?

Crashes in mozilla.... means you're biased towards open-source. Keep up the good work.
posted by eustacescrubb at 5:19 PM on January 4, 2003

Eustace, if you were just fatigued, then how come you were still biased in favor of heterosexuals? If your mistakes really were even, they should cancel out. That's why the test is so damn long--it takes the subject from "I wonder if I'm biased" to "Let's just get this damn thing done," which is when one starts making mistakes that expose one's bias.

And considering that the majority of cross-dressing is done by heterosexual males (90%), it does look like you have a bias.

Bias is bad--it's taking a cognitive shortcut, rather than fully assessing a situation. Morals are not a bias, but a methodology for reaching a conclusion. Moral decisions can be tainted by bias, though--e.g. "Because The Bible/Chomsky/Bin Laden said to do (task), (said task) must be good."

Hell, let's try it now. Is the following argument logical?
1. People who believe in democracy believe in free speech.
2. Dictators do not believe in democracy.
Conclusion: Dictators do not believe in free speech.

Now try this one:
1. Grass is green.
2. Frogs are not grass.
3. Frogs are not green.

Bias is a crappy substitute for real logic, even when the conclusions are agreeable.
posted by LimePi at 7:54 PM on January 4, 2003

//Eustace, if you were just fatigued, then how come you were still biased in favor of heterosexuals?//

I'm guessing you didn't take the tests then? There were more than one, and the one having to do with sexual preference is a different one than the one in which i made all my typing mistakes. Guess which one I did last.

//And considering that the majority of cross-dressing is done by heterosexual males (90%),//

Um what I actually said was : "I also don't find androgeny or corss-dressing sexually attractive"

You're confusing gender-related activities with sexual ones - I know a good many straight men who've cross-dressed for a lark - they were having fun fucking with a cultural norm, but they weren't trying to get laid. I've thought it was fun when male friends (straight or gay or bi) have gone drag, but I didn't want to sleep with them.

// it does look like you have a bias. //

Uh, no shit Sherlock? Didn't I just say that? What I was taking issue with was not the assumption of bias (we all have them, all the time) but the faulty conclusion that all bias is bad. In the case of seual preference, my "bias" for straight stuff has to do more with my biology than any choice I make - but you, like the test, assume that preferring straight sex and images of the same is a bias and bad, which is to say that you're telling me being straight is bad.

//Bias is bad--it's taking a cognitive shortcut, rather than fully assessing a situation.//

Ah. It's dictionary time! My Webster's Unabridged Dictionary defines bias first as a kind of line drawn over fabric, and second as: "an inclination of temperment or outlook."
So, let's, for laughs, look at your above statement: "bias is bad". You're assigning moral value to bias, which is cool, but let's plop Ye Olde Definition in there in place of the word itself: "an inclination of temperment or outlook is bad." Hmmm. Now that sounds fishy. Sounds fishier when one notes that you don't know what a bias is:

it's taking a cognitive shortcut, rather than fully assessing a situation.

So when you say bias what you mean is sloppy homework?

Morals are not a bias, but a methodology for reaching a conclusion.

No, ethics are based on methodology. Morals are a conception of right and wrong, of good and bad. That is, to think one thing is good, and another bad is to have an inclination of outlook - one is inclined to believe X is good and Y bad or vice-versa.

Moral decisions can be tainted by bias, though--e.g. "Because The Bible/Chomsky/Bin Laden said to do (task), (said task) must be good."

Hell, let's try it now.

Okay:

Your argument above is implied, but goes like so:
Statements like "Because The Bible/Chomsky/Bin Laden said to do (task), (said task) must be good" exemplify flawed reasoing.

I'll buy that. But let's represent it a little more abstractly:

Statements like "Because Fact A is true, therefore Fact A must be Moral Value A" are logically flawed.

Gee, this is fun. Let's plug some other words into that formula. How about, Fact A = "this statement is biased", and Moral Value A = "bad".

Statemetnts like "This statement is biased, therefore it must be bad" exemplify flawed reasoning.

Preferring nonbiased statements to biased ones is an inclination of temperment or outlook. Never mind the fact that objectivity is inversely proportional to how much one cares about a topic. Are you suggesting that we all become emotionless robot people? Like Vulcans? Even Spock was biased (in favor of being logical over being emotional).

You assumed, I think, that because I disagree with the test's epistemic basis that I disagree with the society it envisions. I don't: I think people should not be hateful, and should respect each other's differences. But I do not think that eliminating bias will solve the problem, because one cannot eliminate bias. Every group has outsiders - our group - the dominant culture in the USA, values tolerance, which means we are biased towards it and against other stuff, like nontolerant stuff. I think this is a good thing. But from the POV of, say, a Grand Dragon Wizard in the KKK, our "tolerant" society is oppressive and nontolerant - because we don't conisder bigots to be tolerant, and thus consider them enemies (rightly so). They want to destroy that which we want to build.

In short- we all have biases. Even you. You're biased against biases, it sounds like.
posted by eustacescrubb at 5:11 AM on January 5, 2003

Egads, this is all semantics. Obviously the dictionary definition of the word is not what is being referred to here -- the "bias" that this test purports to measure is something more akin to mindless, unconscious prejudice than a well-thought-out outlook. (That said, I suppose one could make an argument that one should be prejudiced against certain groups, but most people don't make that argument -- most people think they treat everyone equally, regardless of skin color, national origin, etc. That's why the test is important: to make people aware of bias they didn't know they held, so they can get rid of it, and -- if you insist on rushing to the dictionary -- to replace it with another view, a conscious one, that corresponds with ideals that they hold.)
posted by Tlogmer at 7:52 PM on January 5, 2003

What a meaningless exercise. It's sorting kittens and axe murderers on the left against puppies and Jehovah's Witnesses on the right, then sorting beanie babies and mouthwash on the right and oriental vegetables and muscle cars on the left. Wait, or was that the other way around? The results had nothing to do with any damned bias, but with one's ability to adapt to changing associations.
posted by GhostintheMachine at 4:49 AM on January 6, 2003

I thought it was a fascinating test even once I got to the result - I'm strongly biased toward straight people, uh, yeah, right - and started looking for the methodology to see how this had been arrived at. When I found there wasn't a link, I felt like Charlie Brown missing the football again. Why can't these online tests explain themselves? Is it because they're bogus? Or what?

What I really wanted to know was, was my whole score skewed because the first time the two men together came up, joined at the side of the hip, I assumed they were Siamese Twins?

Oops. There's my bias. I meant, conjoined...
posted by soyjoy at 1:41 PM on January 6, 2003

I discovered a very strong bias against java and abysmal usablity.
posted by dagnyscott at 3:14 PM on January 6, 2003

I HAVE NO BIAS - YOU ARE ALL GREEN AND TILTED SLIGHTLY TO THE RIGHT and you smell like asparagus...
posted by troutfishing at 11:16 PM on January 7, 2003