Intelligence test batteries are essentially tests of pattern recognition. The tasks contained vary from the relatively simple component tests of memory (remembering a series of numbers in the order presented, for example) and processing speed (rapidly copying symbols with matched digits, for example), up the ladder of complexity (reordering digits in a recall exercise, identifying embedded patterns in figural drawings, identifying conceptual similarities in verbal and nonverbal categorical reasoning exercises). The most comprehensive batteries also contain scales assessing real-world knowledge (of the type you might acquire in school), as well as scales in which you demonstrate understanding of how such knowledge is applied in real-world situations (which is also a form of pattern recognition, albeit one that relies on baseline knowledge). Even the relatively simple, ostensibly culture free tests (reaction time) rely on accurate/rapid identification of a stimulus association and accurate/rapid response.

But pattern recognition is itself not a unitary process. In the most simplistic sense, it relies on component processes: conceptual priming (activation of a set of related concepts in memory) and modulation (selective de-activation of competing sets). The process is (not by chance) analogous to the scientific method – associations between variables are identified, then subjected to tests of significance to determine whether or not those associations are actually robust, and finally subjected to a conceptual/methodological analysis.

One of the difficulties in comparing this general capability across groups or contexts is that, arguably, different operating fields entail different payouts for various biases in the process. For example, the fuzzy logic underlying social reasoning is qualitatively different than the hard logic underlying mathematics reasoning, and facility in one area does not necessarily generalize to facility in the other (in fact, there’s some evidence of the contrary, at least for many people, such as assessment of the social reasoning among physics professors, as a group). There is no reason the think that the broad range of environmental contingencies associated with different geographical/climate zones/cultures would be any different, though, identification of relevant differences are difficult to identify and quantify.

In our educational system, what we have essentially created is a container for the optimal application of a particular set of reasoning strategies – it is the conceptual container to fit our culture. With intelligence testing, we have created a measure by which we can predict the degree to which people can operate comfortably within that container (primarily limited to education, though also broader, in a sense). Is it any wonder that the people who created this container are naturally those best suited to operate within it? The unanswered question is whether or not the container itself is an optimal fit in the broader, concrete universe (there is an assumption it is, and that assumption is itself unwarranted, given the lack of comparisons).

Do ethnic groups vary in intelligence? In a qualitative sense, to the extent that geographical isolation covaried with substantive differences in operating fields, that is undoubtedly true, because there are different payout biases in each field. Do those differences impact relative odds of success in Western educational settings (and many career settings)? Undoubtedly, because those operating fields have been evolving under the direction of a certain set of biases. Is our educational system optimized for functioning within the constraints of the physical universe? Well, no – that would be saying there’s no room for improvement. But we can certainly see that people for whom the operating field is already optimized benefit from the pre-existing match between native predilections and the environment created by others who share those predilections.

But none of it will mean anything until we can have a much more thorough discussion about what intelligence actually is – a discussion we are not really ready to have, and one that Rushton, Gottfredson, Sailer and others are uninterested in anyway. There’s your real framing problem, Will.

An aside: the outcome of the No Child Left Behind act is to shift the calculus employed by school administrators as they determine how to best deal with “problem” students. There is an incentive now to shuffle underperforming students out of their school and into (relatively unenriched, and usually ineffective) alternative education programs, rather than to attempt remediation in the least restrictive setting (home school). It’s no wonder minority students aren’t showing differential improvement – the whole enterprise is a monumental joke. Any analysis of such programs that doesn’t take this into account is fundamentally flawed (also a joke), which I’m sure doesn’t dissuade people like Rushton from making as much hay from it as possible (ignoring such factors is attractive, as it complements his racist biases).

Oi, General! I haven't had to chomp my way through that many multi-syllabic words in a long time! Woof... I think you took some extra fiber this morning to get all that out. :-P

I get the point you're making, General: ethnic groups vary in selected traits due to both environmental and hereditary influences. Therefore it's valid to conclude that yes, different ethnic groups would have varying degrees of intelligence (overall, assuming a broad average). Individuals within said ethnic group will have varying degrees of intelligence (because nature deplores uniformity) but as a whole there will be noticable degrees of difference.

I would like to add one caveat: the elephant in the room that no one is talking about (quite possibly because most aren't conscious of it), is the social contract binding this society together and how the "intelligence as ethnic variable" theory collides with it. One of the reasons American society remains a cohesive whole is the belief of egalitarianism: we all believe that we are fundamentally equal to each other. Any theory that says we are not fundamentally equal, and this especially includes what Will is talking about, actively undermines the egalitarian bond that holds our different interest groups together in a society. If you get a majority of people to accept the conclusion that certain ethnic groups are dumber as a whole than others, then you've struck a major social blow against treating said groups equally.

We are still monkeys in our tribes, General. We just play at being human. If you take a carefully struck balance in which everyone believes in the ideal that we're all equal, and introduce a given inequality, you cannot continue to maintain equality. This is why people are up in arms about this whole idea, declaring it 'racist', and 'evil, etc. Their cherished notions of egalitarianism go right out the window with this information.

The differences that are most rational to posit are qualitative ones. They are quantitative only in the sense that the quantitative measurements we've developed are, themselves, biased in favor of the context in which they're developed.

To posit qualitative differences doesn't contradict anyone's sense of egalitarianism. To automatically assume superiority of qualitative differences, however, or to homogenize such qualitative differences into quantitative differences by employing a measure that biases in a certain direction, does - and should.

For people who assert that what is measured is a reflection of cognitive efficiency that spans contextual operating fields, I ask why is it, then, that chicken sexers (look it up) aren't all members of Mensa, and why I will never read an advice column, or book on parenting, written by a PhD in Engineering. Various qualitative differences might include optimal speed of association identification, optimal speed for obtaining cognitive closure, optimal speed for identification variability (and at what frequency), etc.

The most vocal proponents of such reasearch are the least sophisticated in regards to what, precisely, it is they're discussing. That's not accidental either.

Try this idea. It works better. And it's much simpler. Ockam's law, and all, you know. Families in which the parents have good jobs with good pay came from a solid educational background, they understand that to make it in life their kid has to get good scores on tests. These parents place high value on tests, they pay for practice, they tell the kids that tests are important. The kids understand value is attached to these tests. They study a bit and do OK, on average.

Families in which the parents may have had a lousy education sometimes don't know to attach value to these tests. The kids show up in class, the test is just some irksome unknown to do before recess, they never studied, and they don't do OK, on average.

Sum all these cases across school districts and states and nations, and what are the correlates? Poverty. Education. Parental income. School district. Socially defined race. I'm sure it would also correlate to percentage of kids with dangerous asthma (which characterizes poor areas for interesting, complex reasons). Does asthma cause intelligence? Does intelligence cause asthma? Among the thousands of correlates of test scores, we only obsess over a few. Or maybe, just one.

Why does the correlation of test scores with "race" continue to fascinate and compel us? Our history and our culture. Certainly not our biology, or any direct causation between genes from this-or-that-latitude, on average, and a score on a test made up by people who are trying to test this untestable thing.

True story: a GRE (an "IQ" test for entrance to grad school) representative bragged several years ago that the GRE score itself finally predicted future professional success better than did parental income.

There is a school of academic racists, however, that have generated a considerable amount of (flawed, including many correlated dependent variables) data that (in theory) controls for the factors you just mentioned. Myself, I think the entire study of ethnicity and intelligence is an unfortunate distraction, notable primarily for the manner in which it's prone to being misused in the service of dubious social agendas.

But watch out for these guys - for the uninitiated (which does not necessarily mean unsophisticated - see Watson), their argument is apparently persuasive. Saletan is arguably many things, but stupid is (probably) not one of them. [Insufficiently selective in regards to his friends/associates being the most easily substantiated allegation, I think.]

GRE and parental income: there are idiosyncrasies of range that make the reported association pretty much inevitable. Think about it.

General, there's no faulting the logic of your argument, but it's missing any consideration of the real world (the "context" as you describe it). Sure, there is no inherent superiority between qualitative differences in "intelligence" but I'm afraid it might be expecting too much to hope most people will see it that way. Not only that, but how do we react to these qualitative differences (if they exist)? It's much easier to assume that people all have the same capabilities and try to get everyone to fit in than change our society to give all different styles of intelligence (or whatever you call it) equal opportunity at success.

In other words... you say that intelligence measures are biased by context, but the fact is that we live in a certain context. If we could change the context to fit everyone that would be ideal, but that's not easy (although I'm sure we can do better). I think these types of issues are the "freight train" that Saletan is worried about and I think he's right to worry about it.

There's always a difficult trade-off between homogeneity and heterogeneity in various adaptive circumstances. Currently, this culture appears to be attempting a social correction by means of affirmative action, etc. - our meager attempt to change the nature of the container.

The linchpin upon which all of the racist studies of IQ and achievement rest is the idea of a "g" factor that explains everything important. My point - we are quite a ways away from having a very thorough discussion about cognition, let alone any meaningful discussion about why it is some people make better engineers, and others better artists. Whether "g" is a valid concept or a statistical artifact, there are other obvious factors these folks are too ignorant or thoughtless (or biased) to consider.

We don't need to accommodate everybody. But we definitely need a more thorough understanding of the problem space before we come to any potentially unflattering conclusions about the general cognitive abilities of specific groups. Too many confounds, as it is, and plenty to mislead.

Certainly not our biology, or any direct causation between genes from this-or-that-latitude, on average, and a score on a test made up by people who are trying to test this untestable thing."

1. If it was untestable then then it would not have external validity in predicting academic and other socioeconomic outcomes. <link>

2. Psychometric g is "extracted" from IQ tests and accounts for virtually all of the external validity of those tests. Elementary Cognitive Tasks (chronometric measures) can measure g as well as good IQ tests.

3. There are a number of neurobiological correlates, including myelination quality (affects processing speed & is largely hereditary - see New Scientist article March 2009 <link>) cortical thickness:

"In healthy adults, greater intelligence is associated with larger intracranial gray matter and to a lesser extent with white matter. Variations in prefrontal and posterior temporal cortical thickness are particularly linked with intellectual ability."

"Relationships between IQ and Regional Cortical Gray Matter Thickness in Healthy Adults" Cerebral Cortex 2007 17(9):2163-2171;

4. About 7% of the genome has undergone recent selection (past 10,000 years). The sweeping alleles we see are mostly regional - you see them in one group and not the other two. A fair fraction are neurological and likely to affect behavior in some way. For example, you see new versions of SLC6A4, a serotonin transporter, in Europeans and Asians. There’s a new version of a gene (DBA1) that shapes the development of the layers of the cerebral cortex in east Asia.

East Asians created our educational system?

There appear to be some similarities, however.

[Note: I claim no expertise in the area of education in ancient China.]

going through and refuting your points one-by-one. Two quick points: the reaction time assertion is patently ridiculous (really). I saw a paper awhile back indicating people with Tourette's do better on such tasks than other people - are they all smarter than me? [Note: could make the argument they make better jazz drummers - at least one case study to support.]

Other stuff: there's an active, growing body of research on neural plasticity that should, if you're paying attention, give you great pause about the conclusions you're implying from the findings you cite. Think about it. [Sigh...]

Trivia - didja' know that Verbal GRE scores are a better predictor of graduate school performance among engineering students than Quantitative GRE scores? Whaddya' think that's about?

[If you're the same person who used to post as BenK, you've slipped quite a ways since I last encountered you.]

"refuting your points one-by-one."

1. You haven't said anything to refute the fact that 'g' predicts educational outcomes (see the link above to recent studies). This applies internationally:

"International cognitive ability and achievement comparisons stem from different research traditions. But analyses at the interindividual data level show that they share a common positive manifold. Correlations of national ability means are even higher to very high (within student assessment studies, r = .60-.98; between different student assessment studies [PISA-sum with TIMSS-sum] r = .82-.83; student assessment sum with intelligence tests, r = .85-.86). Results of factor analyses indicate a strong g-factor of differences between nations (variance explained by the first unrotated factor: 94-95%)."

The g-factor of international cognitive ability comparisons: the homogeneity of results in PISA, TIMSS, PIRLS and IQ-tests across nations Heiner Rindermann *European Journal of Personality. Eur. J. Pers. 21: 667–706 (2007)

"growing body of research on neural plasticity"

2. Yes, is there any evidence that this is leading to Europeans catching up to East Asians or Ashkenazi Jews in terms of group average?

3. In terms of cortical thickness, here is a more recent study:

"A principal finding of this study is that it supports a distributed model of intelligence where multiple areas of the brain are involved with cognitive ability difference instead of the view that there is just one centre or structure important for intelligence differences in the brain," says Dr. Sherif Karama, psychiatrist at the MNI and co-investigator in the study. "Previous studies have shown a link between intelligence differences and individual brain structure or function. This is the first time that a correlation between a general cognitive ability factor and essentially most, if not all, cortical association areas is demonstrated in the same study."

<link>

4. There is also a later development in brighter children:

"One analysis found the cortex in kids with the highest IQs - 121 to 149 - didn't reach maximum thickness until age 11. Children who were just slightly less bright reached that point at age 9, and those with average intelligence at around 6. In all cases, the cortex later thinned as the children matured."

<link>

"going through and refuting your points one-by-one."

Also, you skipped the point about recent genetic changes. The SNPs so far associated with cognitive ability (so far only a few have been and they are almost certainly going to be of very small effect), cluster to one degree or another along traditional racial lines. Doesn't this suggest that the forces of genetic drift and selection did not stop at the neck?

Clearly, either you didn't read it, or you don't understand what I'm saying.

I have read it and you are ducking the points regarding g & educational outcomes. Also, the recent genetic changes over the past 10,000 years associated with agriculture & population growth. Some of the recent changes relate to neurological development and axon growth.

In terms of plasticity, what exactly do you think this means? That people can become Einstein's with a bit more education? You can improve muscles as well through working out. Does that mean that people don't have different levels of ability?

I think you need to tease out your argument a little more.