Implicit Bias

I think you're seeing his test score as conclusive proof that he is biased.Apparently I have a bias and I favour older people to younger people. My experiences are far more complex than that! I admitt that I find older people more peaceful, but I also have bias towards younger people, in other areas. Similarly, Riverrat could have a bias against whites in some manner. Or even just the colour white, which he associated with nazis.Perhaps the test is faulty in that it is offering choices between one option and another option. It is 'looking' for a conclusion of bias where there might not actually be one. A bit of a false dillemma.If Riverrat disagrees with the result, and he knows he doesn't have biases, because his life experiencesare contrary to the result, then it's reasonable to conclude that his life experiences show a clearer picture, as evidentially, it's a far more thorough picture.. The overwhelming evidence of his experiences of none-bias in his life, are far more important than one mere test.I don't think Riverrat has a race-bias just because of the outcome of this tests. That would be jumping to a conclusion. No scientist would advise this IMHO. I think he's an honest person from what I have experienced of him.In my experience, tests such as these are far from reliable. Sure, they might work generally, but not specifically. Even if generally qualifies as 99%. It's like IQ tests. I've taken the same test and gotten a very different score.( General instance doesn't negate specific instance,). It is both possible that this test works, mostly but that Riverrat and others are exceptions.I agree with Riverrat about unforeseen variables.Maybe RR should re-take the test. If he hasn't got a race bias this time, then what does that mean?Maybe he should take it ten times. If he scores 8-2 in favour of being biased, then I think it would be slightly more scientific to then conclude he has bias, IMHO (heavy induction). But then, what would explain any miss?Apparently, according to people here in EvC town, if I have one prayer miss then God doesn't exist, and I haven't counted my misses. It seems reasonable to apply that logic in this instance.Perhaps one miss means God doesn't exist. So then, it is then fair to say that one miss will mean this test doesn't work all the time? Maybe God doesn't work all the time. Because of the overwhelming evidence of his life experiences, it is required that the test prove itself in an extraordinary capacity, to counter his claims. With great claims comes a great requirement for evidence don't forget! Therefore I suggest that Riverrat scores 10 out of 10 in order to qualify as racially biased, IMHO.This message has been edited by mike the wiz, 02-13-2006 11:03 AM

it is important to note that these biases could come also from simply being raised around one group. that's how sexual selection works and i can't imagine it not coming into play in the test.

Is it real bias then?Or is is it pressure to conform? If it is pressure to conform via cultural pressures, then I would say that equivocation with the term bias, is very easy to do.If we're talking about generally tuned behaviour in our wiring, via environmental pressure, then I suppose many biases are entirely innocent and even inconsequential.I prefer cats to dogs. Pressure; in my life dogs have attacked me, and made unecessary large noises at me, but cats haven't.-----BIAShttp://www.dictionary.com1. A preference or an inclination, especially one that inhibits impartial judgment.2. An unfair act or policy stemming from prejudice.I suggest that the second definition of bias is something different from a mere inclination to favour this rather than that. A self-motivation of bias, seems a stronger form of preference wouldn't you say?This message has been edited by mike the wiz, 02-13-2006 10:59 AM

i don't think that there is a real difference between the behaviour denoted by "i hates me sum niggas" and "i really only date white guys because my family is white". the problem with such thought processes isn't with the thought processes, but rather with the behaviour it elicits. if your attitudes toward someone else cause you to impede on their rights, you are in the wrong. if your behaviour simply affects your personal choices and not the rights of others (say, whom you choose to date) then your thoughts are your own. they may be preventing you from enjoying something (jungle fever?) but no one has the right to make you think otherwise.and that defines my opposition to hate crimes legislation. it's legislating thought which is wrong. we can only legislate action. a person who murders a gay person randomly is no less culpable that a person who murders a gay person by choice. they are both murderers.

They are both murderers, but one is perhaps a serial killer. I would fear the person who murdered because the person was gay, more than I feared the random fruitcake who got drunk and lost his marbles and killed someone. Both have a problem ofcourse. Yes. The latter seems to be just as prejeudice but not as obvious, perhaps.Yet the first saying incorporates a supreme self- motivated prejudice. I disagree that there is no difference.There is a difference if the person is genuinely only dating whites because of environmental pressure (family).There is an active racist, and a passive conformist. I still think the active person is more dangerous and his bias is ethically wrong. I don't think environmental pressure makes me a person have such a bias. I think many biases are innocent through sloppy thought processes and unconscious conformity. Back in the day, there might have been many white racists, but I would bet a lot of money that it was the willful ones that were dangerous, rather than the people who merely went along with the group.While your example makes the actions of both type of person, equally murderers, that doesn't mean that people with a more passive bias are just as likely to kill.I would fear the person who willfully hates ni**ers rather than the person who is a victim of society's opinion.

what if the person who randomly killed a gay person killed him because he was left handed and the killer is a serial killer of left handed people?the thing is that through education, we can try to remove prejudice from society, but we cannot legislate opinion and it is ethically wrong to try to do so. that is why it is important to protect hate speech. the most important protections the constitution gives us are the ones no one approves of.i think the passive racists are more dangerous because, chances are, they don't know they are biased. and they have the possibility to hurt someone unknowingly; for example, the HR people cited above who claim to be actively recruiting minorities but don't demonstrate this in practice. active racists are easy to spot. i believe it was syndey poitier, but i saw an interview wherein a black actor decided to move to mississippi because the upfront racism there was better than the hidden racism he'd been confronted with in the north (which i quite agree exists).

Hi all,Just a general reply. I use the IAT in my research and it is a very robust measure of implicit bias. It has predictive and convergent validty, and is used in many areas of psychology, from consumer to clinical/abnormal. Here's a couple of studies that show its predictive validity... Both of these studies used the IAT and showed relationships between the IAT and the experimental measures of race behaviour.As mentioned before, it measures differences between the strengths of category associations using RT/latency measures (which are used extensively in cognitive psychology). The exact mechanism underlying the task is unclear but it is very robust. There is a clear distinction betwen explicit prejudice and implicit prejudice, and it's suggested that in situations of ambiguity and/or complexity, these unconscious biases may be important. Many of the criticisms mentioned in this thread are important (e.g. familiarity, task-order) but they are known and accounted for.Implicit bias helps us understand why, although explicit measures of prejudice show little prejudice, we still have prejudice in society. For example, in the UK, asians make up a large minority of doctors in low positions, but are rare as consultants.I like this quote to highlight implicit prejudice... and don't forget it works both ways, the OJ trail may be a good example of prejudice. Thus, most of the jury had a positive bias towards OJ and negative towards LAPD.edited for typos.This message has been edited by melatonin, 02-13-2006 06:16 PM

10 times?
Ugh.Thanks mike.

I was rasied around gay people, and blacks.So where does that leave me?But for real, my parents always beat it into to my head to not be racist, or bigoted. I was rasied mostly liberal.If you are taught that, and then cross that line, I actually found that black people treated me better on a whole than whites. Forget about Spanish, I love them, married one.Most of the time it takes someone like me to make that first move, to show them that the racial barrier just doesn't exist in my mind. It exists in society, and that just sucks.

I would like to actually see those formulas, and how they work. Then I would be able to talk more about it. But we are to just assume that everything has been accounted for?Like I said earlier, what if I actually had no bias, but was just able to recognize guns and blacks faster than whites and non weapons. It would stand to reason that I would associate them faster, leading to a false result. It's like red light green light 1,2,3. Or Simon says.You think because someone says Simon says put your hand on your head, while actually displaying a hand on his stomach, and you go for the stomach, does that mean you are bias towards stomachs?The test is like some kind of trick truth serum Welcome to the forum

Thanks for the welcome Well there are no formula's to account for these factors, just methodological approaches.Familiarity is controlled for by using names/stimuli that are rated to be relatively similar for familairity across the categories (but it as been shown that as long as they are not very unfamilar it has no effect), the task-order problem has been controlled for by increasing the number of reverse block practice trials. The task-order improvement is relatively recent (reported in 2005). Previously, it was known that the IAT effect was slightly greater in one task-order than the other, but this has never affected its ability to measure individual differences in implicit attitudes, just the ability to locate a definitive zero-point. We would counterbalance over the sample to control for such effects.If you were able to react faster (via recognition and response) to black/weapon compared to white/no-weapon within a condition it wouldn't really matter. It measures the RT latency between the conditions (experimental blocks)...1: black/weapon-white/no-weapon (one block), and2: black/no-weapon-white/weapon (another block).If the RT means are smaller in one one block (both pairs combined), it shows you have a stronger association between the two pairs of categories, i.e. they are easier to categorise. On an individual level the results are not that important, I wouldn't worry too much. As a researcher, I'm more interested in what happens with a large group and the same would apply for Greenwald et al.This message has been edited by melatonin, 02-13-2006 08:02 PM

For anyone interested in any issue about the IAT try Greenwald's site...Dr. Anthony Greenwald/Home Pagehe also has his publications available there and a section about validity issues and some alternative interpretations of the IAT effect.

1) A strong preference for progress over tradition -- I expected a moderate.
2) A slight preference for books over tv -- I expected a strong
(I cut off cable two years ago -- haven't watch most of the shows listed and have a compulsion for more books than I can find time to read -- I think the problem is that I haven't read most of the books listed or seen most of the tv shows listed. They are more titles to me than anything but maybe I'm denying something. )

Your posts have been informative and well written, though I am still not satisfied with what can be obtained (or should I say "extrapolated") from association results on these tests. And more specifically I have some amount of doubt regarding the level of confidence about these tests as expressed by the thread's author.I want to state at the outset that I am not concerned with order of task issues. While I suppose there could be some minor effect, I thought the strategy used to minimize such effects were valid.My concerns regard assumptions used to design the test, which ignore differences between associative button pressing issues (cognitive skill tests), and real world feelings towards others, as well as actions.1) Familiarity. While you state that IATs are robust and familiarity issues have been dealt with, I am not seeing that within the literature. From the Greenwald site... While this effect can be addressed as much as is humanly possible, it seems to me there is no way to create a generalized test in such a way as to guarantee such an effect will not be crop up when testing any particular individual. Some items in a list (pictorial or word) might be wholly unfamiliar, and more in one list than another. I know for certain that I had some problems with names that meant nothing to me.2) Diversity/Rarity: This does not seem to be addressed anywhere. Especially with regard to tests that pit a singular group against a collective of many other groups, it seems to me there might be an effect (annoyance or interest) based on the difference in frequency between the singular group and any of the rest. Clearly we are creating relationships as we conduct such tests, and one can be "I am sick of seeing these face/names already" and so likely to associate negative words, or "hurrah its something different" and associate positive with items from the other group.3) Cross-category association: I find this to be one of the most important, given that we are dealing with finite lists. This is from the greenwald site... For any particular test taker, I do not see how this can be avoided by a generalized testmaker. This is a bit ironic given that this is supposed to be a test regarding biases, and yet makes an assumption that biases cannot exist on th individual level. It can happen that any particular list may contain more "liked" or "disliked" items, or in any case associated based on wholly other criteria than general class.As an example, a person may have a definite bias against some specific names and facial types (we're talking specific arrangement of features, not just racial type). This is from associations based on direct physical abuse or other negative/positive events. There is no way that researchers can know this about anyone and create a list that avoids those names/faces from within a generalized list of a category. And indeed they can inadvertantly deselect neutral names for a person, based on their own biases of what is culturally popular/unpopular. Enough of these anomalies (even just one) could skew results.In a way there are even biased assumptions of what is good/bad or "biased" in an association. One might not only be familiar with but actually like blacks and guns, and so such an association does not indicate a person would negatively react to blacks or associate them with violence. A culture where guns are not viewed as negative would have to produce a different result. Same for links between gender and occupation typed.I think there was one more issue I had, but I'll stop there for now. You did have two citations linking IATs with other tests, but those seemed to be flawed in a similar manner... especially familiarty.It does make sense that a strong association may very well indicate a bias within an individual, but lower order associations do not seem to be as clear cut given the finite nature of the tests and potential for association or "bias" beyond the simple connection being assumed.Look forward to your reply.

---

holmes
"What you need is sustained outrage...there's far too much unthinking respect given to authority." (M.Ivins)

I am a drag racer. A perfect light in bracket racing is .500. That means you left at the same time the green light turned green. Very rarely do you a) cut a perfect light b) repeat the same reaction time.Just based on that evidence alone, I can't see how a timed reaction test can determine you bias. I am sure they are counting hundredths of a second here.You could do something as simple as lift your finger before pushing i or e, and that would change the result dramatically, as opposed to just straight pushing down.