Race & Intelligence (UK TV, Channel 4)

[Paul W]

I’ve off the board for a short while, so I haven’t had a chance to reply to the various posts. However, here goes.

I clearly underestimated the extent to which the prejudices of readers of this thread would contribute to a misunderstanding of what I wrote. Thanks, though, dlorde, I think you got what I meant, but for those who didn’t …

The problem often seems to me to be that psychologists (and other specialists) use words in specific technical ways which don’t always correspond to common usage.

First, those who have trotted out the old racist arguments about the supposed relative ability levels of different “racial” groups appall me. I am merely trying to point out that “measuring” “intelligence” is not a simple matter. The early Binet tests were heavily biased towards French middle class values – as, indeed various versions of the Wechsler tests were (and probably still are – I haven’t used the latest versions) biased towards white, Anglo-Saxon Protestant (WASP) values. For example, in one version of a Wechsler test the question appears “What should you do if you find an envelope in the street which is sealed, addressed and has an unused stamp on it?” The “official” answer is “post it”: in many subcultures the “correct” answer would be “open it to find out if there is anything of value (to me) inside”. Neither is “right” or “wrong”, but each reflects the values of a particular subculture.

Secondly, let me explain, casebro, what I mean by a test “breaking down”. Tests have typical error patterns, and with experience it is easy to spot deviations from the pattern for a given test. Indeed, the Manual for Raven’s Standard Progressive Matrices (a test I mentioned in my previous post) gives a table for “Normal Score Composition”. The test is divided into five Sets, and for any given total score, there is an expected score for each Set. These scores were, incidentally, derived from actual performance data and not from anybody’s ideas about what they should be. Deviations from these patterns are a cause for concern, and may mean any one of a number of problems. One may be a lack of understanding of the instructions, or may be something else: the problem is to find out what it was. My point about having used Raven’s Matrices with deaf and dyslexic clients, as well as with those whose first language was not English is that this does not seem to cause a problem. What does quite frequently cause a problem is, incidentally, brain damage – usually from illness or trauma (head injury), although it might be genetic (eg Downs syndrome).

With the Wechsler tests, which consist of a number of different sub-tests, the normally expected pattern is a roughly similar (standardised) level of performance across all of the sub-tests. This may be high, middling or low depending on the overall level of ability of the individual, but will be consistent across the sub-tests. If it isn’t, you have a problem.

One of the skills of using tests is identifying when something has gone wrong – which is what I meant by a test “breaking down”. This is not a comment on the individual, but on the test. It would, no doubt, be possible to construct a test based on ability to survive in a tropical jungle: your average European or US WASP would probably “fail” it miserably. Mind you, a survival test which your typical backwoods red-neck would romp through would likely baffle a London (or New York) urbanite! My point was that the individuals I was referring to gave unusual error patterns, which suggests to me that the test had failed the individuals, not that the individuals had “failed” the test.

By the way, I find your comment about “ … trying to measure IQ as related to living in the First World?” incredibly ethnocentric and patronizing. I, as a psychologist, am interested in “intelligence” as a generalized concept – not what it is in relation to our society, here and now. Incidentally, which bit of “ … non-European populations … ” don’t you understand? My experience of psychometric testing has basically been with European populations, although I suspect that it would not be much different with other populations of predominantly European descent – eg USA, Australia, and so on. I am not prepared to make generalisations to populations of which I have no experience.

Thirdly, several posters have made comments about reducing the measurement of ability to a single number. Unfortunately, Gould and Gardner notwithstanding, the empirical evidence does lean towards the idea that there is a single underlying ability factor. In the UK, NFER-Nelson – one of the major suppliers of ability (and other) tests – have produced the “Cognitive Ability Test”, widely used in schools. OFSTED also use it to evaluate the performance of schools. It consists of eleven (I think) subtests covering Verbal, Numerical and Spatial abilities. Unfortunately NFER-Nelson’s own factor analysis of the data shows that around 60% of the variance in the scores – across a massive proportion of the general teenage school population – is accounted for by a single factor. Most statisticians would stop at that point and say that they had effectively produced a single factor test. Indeed, their attempts to separate out verbal, numerical and special factors are a bit dubious, at best.

Similarly, although the Wechsler tests have a range of verbal (including numerical) and performance scales, the assumption borne out in practice is that performance across the scales will be similar for an individual: as I noted above, if it isn’t there is cause for investigation.

Fourthly, I am at a loss over GreyICE’s “contribution” to the debate. If we take but one part of it, that “… IQ doesn't mean you're creative … ”: unfortunately research with the 16PF – one of the most respected and widely used personality inventories (please – not “tests”) –does, indeed, suggest that Factor B (roughly “intelligence”) is the single most important contributor, although other factors are also important.

Put simply, ability testing is a minefield, and contributions from those who are not familiar with the detail are entertaining – but irrelevant. Most are driven by political agenda, not reality. Gould, et al, take note.

[GreyICE]

I'm at a loss to identify yours, except to be ridiculous condescending and assert things like '60% correlation means we're basically done' which has to be one of the more ridiculous things I've ever read. Are you just assuming we don't understand basic statistics? A 60% factor is useful for broad generalizations, and nothing more.

[Fnord]

[opinion]
Intelligence knows no racial pre-condition. However, culture and nurture put tremendous limitations on what little intelligence nature may have provided. This could explain the "racial" and "sexual" correlations that racists, sexists and other bigots are so desperately searching for.
[/opinion]

[bpesta22]

Welcome to the board Paul W.

You will not find another topic in the forums here where more skeptics simply ignore scientific data when forming opinions. Many of the flatly wrong statements made above could be easily looked at in the literature.

It is most bizarre.

[Paul W]

Oh dear, I do seem to have stirred up some prejudices.

bpesta22: thanks! You are absolutely right, except it’s not sceptics – it’s the ignorami, of whom GreyICE is a good example. BTW, it’s “back to the board”, not “welcome”.

GreyICE: I didn’t say “ … a 60% correlation … ” because such a statement doesn't make sense: I said " ... 60% of the variance ... ", and to account for this the correlation would be around 0.775. Please, GreyICE, read what I wrote, not what you think you read. And yes, that kind of statement does mean that you don’t understand basic statistics – or advanced statistics for that matter. Yes, I do. My degree was in psychology and statistics.

Fnord: not sure I agree – but it sounds interesting. Let’s (argue) discuss. Sorry, I use “argue” in the rhetoric sense of “let's have an entertaining dispute”, not in the sense of “let's fight”.

[GreyICE]

Oh, okay, so you were vague and rude. And that means I should read you better? Even .775 is pretty shoddy in many respects - and that's correlation across one battery of tests, in 3 factors (Verbal, Numerical, spacial). In fact, .775 is too low to to establish correlation into a single factor - as you probably should know.

Your inability to communicate is not my problem. Your credentials are not that impressive.

[Skeptic]

Originally Posted by Jeff Corey

From the link;"In 2007, Nobel Prize winning US scientist James Watson was quoted referring to research suggesting that black people were less intelligent than other races. His comments caused a storm of controversy, Watson was condemned."
Watson's DNA test showed he had 16% African heritage. Now that's irony.

Well, it explains why he made such a stupid comment...

(run away)

Okay, jokes aside, even if true, so what? Why would or should a small difference in the average IQ of two groups matter to me in treating individuals of those groups?

[GreNME]

I'm still getting hung up on the measure of the term "intelligence" as a process. While, yes, intelligence plays a role in creativity or social ability, it's certainly not a requirement for hand-eye and physical coordination and aptitude, nor does a high capability for general "intelligence" mean that one is, indeed, intelligent.

Also, while I understand that the comparison is between group averages, but how would this translate to the individual in any meaningful fashion? Are the alleged differences enough to be predictive in any measurable amount or significant proportion? It seems that this would all be important considerations to cover if we're looking to derive some kind of reasonable (meaningful) conclusion based on the data instead of expecting the data to have some sort of self-evident conclusion jump out at us.

I have to admit to asking from a purely personal motivation: I tend to scoff at groups like Mensa, despite having tested well above the necessary scores before (despite certain handicaps), and I'd be devastated to learn that I'd have to not be as scoff-prone.

[leon_heller]

I saw it, it was quite good. There was another programme the other evening in the same series about mixed-race children. It concluded that having parents from different racial groups was advantageous from a genetic point of view, just as it is with plants and non-human animals.

Leon

[Paul W]

leon heller:

Let me tackle the “simple” problem first. In the late 1960s I remember a paper written by an eminent biologist which argued that a proportion of any herd based animal population (eg humans) were somehow “programmed” to go outside of their immediate social group (aka “herd”) to find a mate and reproduce. The obvious advantage of this is genetic mixing and ultimately the survivability of the species. Certainly where I grew up there were several populations where mating outside the local group was infrequent, and this seemed to have led to questionable genetic (d)effects, often expressed as:

“X” born,
And “X” bred:
Strong of arm,
And thick of head.

(sorry should be “**** of ’ed” - and the "strong of arm" is a bit dubious anyway.)

Fill in your choice of X.

Sorry about the **** above, but the auto censor thought I was trying to evade it. I wasn't, but it rather destroys the meaning of this. Replace the "th" with " 'f "

Interestingly, I have recently seen suggestions that the marriages between first and second cousins prevalent in some ethnic groups has led to an unusually high level of physical birth defects.

The paper was, I think, published in Nature, but I may be wrong and I have not been able to find it again. It was of particular interest to me because at the time I was in the process of going well outside my own herd! I would be most grateful if someone can find the reference.

The serious point about this is that genetic intermixing between widely diverse populations is essential for the genetic health of a species. As the 19something pop song went “ … coffee coloured children by the score … ”

I’m sorry I missed the programme: could anyone please give me a catch-up reference.

I'll come back to the intelligence issue later, GreyICE.

[Undesired Walrus]

Christ bpesta, is there an alarm that sounds when the terms 'race' and 'intelligence' are brought up in this forum? I don't think I've ever seen you talk about anything else.

[bpesta22]

Originally Posted by Undesired Walrus

Christ bpesta, is there an alarm that sounds when the terms 'race' and 'intelligence' are brought up in this forum? I don't think I've ever seen you talk about anything else.

I do occasionally search out the topic. Anything IQ I like to discuss, and boobs. We don't get too many of the latter, so here I am.

I'm sure there's people who publish in areas and then post lots whenever those topics come up (francesca in econ; FLS for health).

I have been trying not to be repetitive in newer threads on the issue.

I dunno what field you're in, but what if vast amounts of skeptics were discounting it / trashing it based on simplistic comments illustrating they have not read the literature?