In 2005, an American implicit association test revealed my views about black American males. These computer tests infer your ‘implicit’ attitudes by how quickly you link positive or negative concepts with photographs of persons of particular ethnic groups, or ethnic names where these are easily linked to particular groups.
Back in 2005, the conclusion was that:
Your data suggests a moderate automatic preference for white people relative to black people.
One of the things I really did not like about my trip to the US in June was the way I absorbed the racial culture. I quickly fell into the habit of doing quick risk assessments on young black men. I could not recall the precise statistics, but I was well aware that they are massively over-represented in the criminal justice system. Most times I concluded that they were no threat and I never actually found myself in a worrying situation (unlike my first trip to the US in 1992, when I am pretty sure I at least would have been robbed, had not the police arrived and arrested the guy who was harassing me and my friend – they had been looking for him anyway).
I wanted to take a moral shower every time I thought this way, but my self-defence instincts were too strong to stop the thoughts entering my mind.
In a new Australian implicit association test, organised by Andrew Leigh and Alison Booth, I was spared any need for a moral shower. My responses were neutral as between Anglo and Chinese names, though I expect that if terms like ‘studies hard’ or ‘drives well’ had been used some stereotyping may have been revealed.
While I encourage readers to take the Australian test to help my friend Andrew L to get a reasonably large sample to analyse, I am lukewarm on these tests – or at least think that only very limited conclusions can be drawn from them.
As I noted in 2005, even for stereotyping purposes race is rarely the only information we use. It is only one aspect of general appearance.
My reading of the research suggests that few people are doctrinal racists, believing persons from other groups to be inherently inferior or always to be avoided. However many have prejudices that can be contradicted by information about specific individuals. This is how we can get the phenomenon of people who say that they don’t much like blacks/Asians/Jews/whatever, but that the particular black/Asian/Jewish/whatever persons they know are all ok.
Also, and importantly, prejudices are rarely the only thing guiding behaviour. Social norms and sanctions guide our interactions with others. General norms of civility can maintain good relations between people even when prejudices are widespread.
The internet notoriously weakens those norms and sanctions, as the bad behaviour in comments threads shows. It makes web the ideal place to test underlying prejudices, even if a poor place for working out how people might behave in the real world.
Andrew Norton 
Speaking of prejudices, it’s good to see economists start doing social psychology :).
My suggestion (if they want to do it properly — everyone else should ignore this from here on in, as it’s just something to do with experimental design) is that they should consider masked priming, parafoveal preview, or better paradigms than the IAT, since otherwise they’ll end up with a database of effects that are rather hard to interpret and rather open to bias.
Social psychologists (and economists also) need to start reading the cognitive literature a bit where people basically gave up on these sorts of paradigms more than a decade ago.
That’s my prejudice anyway, which is especially unfair since I think one of the people can’t even reply even if he wanted too :). Damn economists :).
LikeLike
I don’t feel any need to prove any lack of prejudice on my part. I’m like most of the population, but the majoity will rationalise their prejudices. When I’m driving I notice certain groups of drivers are worse than average, and if an Asian, a westie, a red P plater, a hat-man, etc etc. does something stupid, I mutter (if I’m with my wife), or yell (if I’m on my own), “effing xxxxxxx”. Makes me feel better. Sorry guys, I’m obviously morally bereft because I don’t feel any need to prove my moral superiority. Something missing from my genes, I suppose.
LikeLike
One interesting aspect of the test was the question listing options about “how close to someone of X racial group you’d be willing to be.” Apparantly, accepting someone as a family member is seen as a closer tie than accepting someone as a good friend. I disagree with this view – I accept my in-laws as members of “the family”, and would do so even if they were baby-killers, pedophiles or church goers; I doubt that I would choose to have such people as my friends, however.
LikeLike
Tom – This is a common order in social distance surveys, but I take your point. As they say, often you can’t choose your family, and people will accept in-laws to keep the peace.
The question about the Lebanese was perhaps not specific enough – many people would probably distinguish between Lebanese Christians and Lebanese Muslims. (Highlighting that in these issues race itself often isn’t that important; it is race as a proxy for culture.)
LikeLike
Within about five seconds of this test, two words sprang to mind: “confirmation bias”.
Even so, I got ‘liitle or no bias’ which is rubbish. I can tell you here and now if I am walking down a street and see a gang of aboriginal young men on one side and a gang of anglos on the other, I will walk on the Anglo side. Yet this test classes me as totally blind to such racist, evil, and possibly UN-Convention breaking thought crimes!
If I were told those names were Scottish people or Swedish, I doubt the results would have been any different. And one wonders how many people from full-blood aboriginal communities or those where they actually have names like those on the IAT – rather than Smith, O’Donohue, and Jones – were tested for Thought Crimes Against Whitey!?
Set up a similar test for Leftists but instead of aboriginal names use Cohen, Schwartz, Dershowitz, etc. Heh.
LikeLike
Andrew
If, as you say, young black males are ‘massively over-represented’ in the prison system, then why is computing a ‘risk assessment’ of young black males a matter of prejudice? are you not simply acting logically?
LikeLike
While there is undoubtably bigotry against blacks/asians/jews/whatever… that does not mean that every generalisation about those groups is caused by bigotry or that those generalisations aren’t useful and appropriate.
Who here seriously thinks they can’t predict the winner of the next Olympics 100m sprint?
Generally speaking, people from asia do study hard. And generally speaking, the average person from asia looks more asian than the average person from australia. I’m not saying this because I hate/love asians or because I think they are better/worse than me.
LikeLike
The “Aboriginal” names in this test are bunk.
Most of the Aboriginals in Australia cities do not have traditional indigenous names – they have regular English or (in particular) Irish names.
The aboriginals who had little contact with white settlers and therefore retained their own names are not the aboriginals who are currently in your local train station causing trouble. They are still in places like Arnhem land and the Kimberly where they have always lived.
The average australian would have very little contact with anyone named Yunipingu or anything similar unless they live in the Northern Territory or far-north Queensland or WA.
LikeLike
Your missing the point Yobbo. There may well be more Aboriginals called O’Reilly than Yunipingu, but if the name O’Reilly flashes up on your screen you don’t think about Aboriginals.
LikeLike
I did the American version of the IAT a while back and I still think that language familiarity has a lot to do with the test more than racism. For example, even though I am Anglo and Celtic by ancestry and Australian by nationality, I am Muslim and speak Arabic so I am familiar with Arabic names. The American IAT used a lot of Arabic-derived names for their ‘black’ examples. Consequently, I scored a slight preference for African-Americans on that test. On this Australian test, I was offered Italian and Anglo names, and showed a slight preference for Anglo. I am pretty sure this has more to do with the fact that I do not speak Italian and thus have less familiarity with Italian names than anything else.
LikeLike
I think you are missing the point Spiros: When I hear the name Yunipungu, I think of an entirely different group of Aboriginals than I do when I hear say Rielly.
The 2 groups may be of the same race but their culture is completely different and to associate the 2 groups makes about as much sense as lumping Chinese in with Eskimos.
LikeLike
Yes Yobbo but the test is about your reactions to Aboriginal names, Italian names etc. If you see the name Reilly you are very unlikely to think about Aboriginals called Rielly and react accordingly. Likewise if you see the name Schnitzengruber you aren’t going think that this could be a northern Italian, even though there are actually a lot of northern Italians with German names. For you to think Italian, you’re going to have to see a name like Rosso.
LikeLike
“I quickly fell into the habit of doing quick risk assessments on young black men. I could not recall the precise statistics, but I was well aware that they are massively over-represented in the criminal justice system. I wanted to take a moral shower every time I thought this way, but my self-defence instincts were too strong to stop the thoughts entering my mind.”
Those are perfectly natural human instincts which are shared by Asians and even other Blacks (Jesse Jackson noted being relieved when he noticed someone following him was white).
Every group has in-group/out group biases, but whites tend to be more ashamed of this & see it as morally wrong. Most other groups are more ethnocentric, you see this with various race based groups or societies. Also in terms of immigration policies, for instance Japan & Singapore have quite clear policies to preserve a homogenous population.
LikeLike