Implicit Bias

In the debate over internal biases, there has been a bit of a problem: How to measure actual bias when people are highly motivated to claim they don't have any? Three scientists at Harvard, the University of Virginia, and the University of Washington, Mahzarin Banaji, Anthony G. Greenwald, and Brian Nosek, have developed a methodology to try and work around it by measuring people's unconscious reactions. They have a program that pits one variable, such as gender, against another variable, such as science/arts aptitude.First, you are presented with references to just one variable. You are asked to sort them as quickly as you can. For gender, you might be asked to sort terms such as "woman," "uncle," "male, "grandmother," etc.Then you are presented with references to just the other variable and must sort them as quickly as you can. For example, you might be asked to sort "biology," "philosophy," "painting," "neurology," etc.Then, the two variables are paired off one way and you must sort the references. All things that are "female" or "science" go in one bin while all things that are "male" or "arts" go in the other bin. The pairings are then inverted ("female" and "arts" against "male" and "science"). Again, sort them as quickly as you can.It seems that when you are asked to sort items that you are biased against, it takes you longer to do it. That is, if you are biased against males and the arts, it takes you longer to react when you see one of the terms because your mind is associating it with the other term, too, and is having a hard time reconciling it.Note, this doesn't mean that you are a raving bigot, though further research has shown that the tests are very good predictors of behaviour. If one is aware of bias, one can work against it, but many people don't realize their own biases.You can take the tests at Implicit BiasNote, you will need JavaScript, cookies, popups, and Flash to take the test.

Neat.Let me be the first to openly admit that I have to think long and hard about taking this test; I find the fiction that I'm relatively bias-free to be comforting.

RrhainI took the test on gay straight relationships and the result was this. . This is pretty much in line with my impression of myself however, several questions arise in my mind.
First I cannot find any reference to research and experiments that are available to see the science being conducted. Also I think I would like to try such a test in the future using a voice response rather than a typed one to see if there is a change in the results as a matter of reaction times.
On the matter of reaction times I fail to see how these reactions can be correlated to the psychology of the person in question. I am assuming here that the ability to respond a good or bad correlation to the main categories of straight or gay is supposed to be a measure of a person true feelings toward a particular sexual orientation. Why does the teast need to be done at high speed to achieve a real result in their view.
Perhaps the test should be conducted in voice response time to avoid fatigue for people such as myself since I found that the athritic joints of my hands became harder to move as the test went on which could affect my score.I am also going to take the test again, this time without filling in the questionaire at the end to see if I recieve the same score.Thanks for the link Rrhain.This message has been edited by sidelined, Thu, 2006-02-09 01:28 PM

Here's how my test came out sidelined. It was an interesting test.

---

There are two ways of spreading light: to be the candle or the mirror that reflects it. -Edith Wharton

A biased study to say the least.No offense, but what a bunch of BS.So I took the first test, the weapons test. They are trying to find out who you associate weapons with more, whites or blacks. But the problem with the test is that it is based solely on timed reactions, which by itself might not to bad of an idea, but again science proves that it just doesn't know all the variables. That is how I see it as biased. It is not knowingly biased, but biased in regards that all the facts are not accounted for.Then make a faulty test to prove something?So first they have you see how fast you can recognize whites from blacks, then how fast you can recognize weapons from non-weapons.Then they put the 2 categories together, and on one side they have weapons, and blacks, and the other side weapons and non-weapons and whites. In my mind I am having enough trouble trying to separate all the categories and which side they go to, but I still have it clear in my mind. As I go, I get better at it, and the program is actually teaching me to associate blacks with weapons, and whites with non-weapons.Then just when you get all used to it, they switch it up, and put blacks with nonweapons, and whites with weapons. Well you have already trained your mind to categorize things in a certain way, and now you have to change it. Of course it will go slower, and it will appear that you do not associate whites with weapons that well.You see in my head, I don't associate either with or without weapons at all. If anything the whites I saw in the pictures looked like white supremacist anyway, and the blacks reminded me of people starving in Africa. Because they labeled them as African Americans, not blacks, so I had an image of people starving in Africa, not that I always associate blacks with starving people.It would have been no different is it would have used car parts and ice cream. The result would have been the same.I see it as next to impossible to actually measure if someone is biased or not.Plus like you said, if one is aware of biased you can work around it. I could take the next test and do just that. As soon as the first cup of coffee settles in.I associated balcks with weapons in my first test, But I do not agree with those results. In my life I have been attacked by more whites with weapons than blacks. I also grew up in a neighborhood where both races would carry weapons. I also have just as many, if not more black freinds than white freinds, so I see no reason to be associating things that way. I will take one more test.This message has been edited by riVeRraT, 02-10-2006 06:30 AM

Of course they make you associate gays with bad stuff first, just like the weapons. It would seem that most of the tests results would indicate that the first thing they associate something with would be the favorable thing.As I did the test, I tried to not pay attention to the first association much, then concentrated on the second one more, and this was the results I got. Now we know it is a bogus test.
Science fails again I'm Gay!TRuthfully, I must say the only thing in my life that has been able to show me my short comings has been the Holy Spirit. Lord knows I am full of them

I would like to retract that statement and say that the study did in fact work. It showed that the scientists who made it are in fact actually biased towards every stereotypical thing they are trying to test.The subconsciously put all those things first, thus showing their true biases.

quote:

---

Of course they make you associate gays with bad stuff first, just like the weapons. It would seem that most of the tests results would indicate that the first thing they associate something with would be the favorable thing.

---

I think they should have conditioned us by associating the term good with the term gay first and then do the reverse.It would be interesting to see how those results would turn out.Would they be the same?

---

"Peshat is what I say and derash is what you say." --Nehama Leibowitz

That weas my reaction to the test.I presumed that was why they made us do each test twice: once to train; once to test.

it's a classic brain association sorting test. we need to get zhimbo in on this topic. anyways. my brain doesn't sort very well and since i figured out the test in the first instance, my brain has tripped itself in the proceeding ones and skewed the results since i was trying so hard. ah well. at least i've got no bias for or against fat people.

They should reverse it, and let people who never did it before do it for the next few weeks and then re-average the results.

Your data suggest little to no automatic preference between Thin People and Fat People.
Your data suggest a moderate automatic preference for Young compared to Old.
Your data suggest a slight automatic preference for European American compared to African American.
Your data suggest a slight automatic preference for Straight People compared to Gay People.
Your data suggest a slight association of European American with Harmless Objects and African American with Weapons compared to African American with Harmless Objects and European American with Weapons.
Your data suggest a moderate association of White Am. with Foreign and Native Am. with American compared to Native Am. with Foreign and White Am. with American.
Your data suggest a moderate automatic preference for Richard Nixon compared to George W. Bush.ah well. no one's perfect. more later. tired now.

sideline responds to me:

quote:

---

On the matter of reaction times I fail to see how these reactions can be correlated to the psychology of the person in question.

---

Because the reaction time is correlated to your personal attitude toward the concept in question. If you associate X with A and not with B, it takes you more time to come to recognize X and B together. If you have a strong association of X and A, seeing an image of X will have you expecting to see an image of A. If, instead, you get an image of B, you have to work harder to recognize it. It's that (*blink!*) reaction: Something unexpected just happened and you have to get past it.And from the other side, having a strong association of X with A has you thinking about A when presented with X, even if B never comes around. It's simply easier to respond to the implicit association.

quote:

---

Why does the teast need to be done at high speed to achieve a real result in their view.

---

Because they are going for automatic responses. If you sit around and cogitate on the matter, you blow past your implicit responses. Again, the test isn't claiming that everybody who has an implicit bias is raving bigot. When asked about the test with regard to courtroom use, Banaji has said she would testify against such a use. Biases can change and there is a difference between implicit bias and overt action.But again, just because people say something doesn't mean they do it. One experiment asked HR managers about white and black applicants. The overwhelming response was that they were actively recruiting black people. The offices were then sent two sets of identical resumes: One set was for an average worker but one had a white-sounding name while the other had a black-sounding name. The other set was for a highly-qualified candidate but again, one with a white-sounding name and the other with a black-sounding name. The only thing they measured is whether or not the resume triggered a callback for an interview.The resumes for white-sounding names received more callbacks (half again as many) and on top of that, there was no difference in the number of callbacks for black-names with regard to the average or highly-qualified resume. Even though these HR people specifically said they were looking for highly-qualified black people, their actions indicated otherwise.

quote:

---

Perhaps the test should be conducted in voice response time to avoid fatigue for people such as myself since I found that the athritic joints of my hands became harder to move as the test went on which could affect my score.

---

Perhaps, but such a problem would affect both sides equally. That is, your physical impediment in sorting would affect all items being sorted. By trying to go as quickly as you can, even though your fingers are a bit slower, your brain makes up its mind quickly.

---

Rrhain

---

Thank you for your submission to Science. Your paper was reviewed by a jury of seventh graders so that they could look for balance and to allow them to make up their own minds. We are sorry to say that they found your paper "bogus," specifically describing the section on the laboratory work "boring." We regret that we will be unable to publish your work at this time.

riVeRraT responds to me:

quote:

---

But the problem with the test is that it is based solely on timed reactions

---

But that's why the test works: You're not given time to stop and pause and consider. Instead, the test bypasses your conscious controls on behaviour and goes toward your automatic response.

quote:

---

Well you have already trained your mind to categorize things in a certain way, and now you have to change it. Of course it will go slower

---

Right, but that would affect both sides. If the test has acclimated you to associate X with A and Y with B, then making the switch to X with B and Y with A will affect you on both sides.Take some time and think about it: Suppose someone has an implicit bias of X with A and Y with B. When asked to sort references with X and A paired together in one bin and Y with B paired together in the other, such a task will be fairly simple: You're being asked to do what you automatically do without thinking.But now, reverse the associations so that it's X with B and Y with A. This will cause both aspects to slow down. X is not normally associated with B so it takes time to figure it out. But just the same, Y is not associated with A so it also takes time to figure out. Since you have slowed down on both sides, you have just demonstrated the implicit bias.

quote:

---

You see in my head, I don't associate either with or without weapons at all.

---

Obviously not or you wouldn't be complaining.

quote:

---

Plus like you said, if one is aware of biased you can work around it. I could take the next test and do just that.

---

No, not really. Not without working on changing your implicit bias or cheating by slowing down.The point behind working around it is that when you have the time to stop and think and contemplate your actions, you can behave in ways that you want to even if they go against your implicit biases. But that doesn't mean you don't have an implicit bias or that it has changed. It simply means that you are going against them.What the test is measuring is what your brain associates when unencumbered by your conscious thoughts telling you that you're not supposed to think that way.I'm sorry the test showed you something about yourself that you didn't like. But you can either gripe about it and get offended or you can learn from the experience.

---

Rrhain

---

Thank you for your submission to Science. Your paper was reviewed by a jury of seventh graders so that they could look for balance and to allow them to make up their own minds. We are sorry to say that they found your paper "bogus," specifically describing the section on the laboratory work "boring." We regret that we will be unable to publish your work at this time.

purpledawn writes:

quote:

---

I think they should have conditioned us by associating the term good with the term gay first and then do the reverse.

​

It would be interesting to see how those results would turn out.

---

Turns out it makes no difference. Which one is done first, which direction is done first, even reversing the left-right direction, none of that matters.Now, a highly motivated person can influence the test by specifically and deliberately holding counter-bias images in mind. But, that's cheating. That's part of the reason why you are told to sort the items as quickly as you can: You want to be thinking about the time rather than on the outcome.

---

Rrhain

---

Thank you for your submission to Science. Your paper was reviewed by a jury of seventh graders so that they could look for balance and to allow them to make up their own minds. We are sorry to say that they found your paper "bogus," specifically describing the section on the laboratory work "boring." We regret that we will be unable to publish your work at this time.