« How Polarizing Are They? | Main | Your Buck Will Not Pass Go. It Will Not Collect $200. »

December 04, 2007

"The Dilettante's First Law of Empirical Narcissism"

Over at Slate, Stephen Metcalf has written a frankly devastating rebuttal to William Saletan's series of race, genes, and IQ. The whole thing is fascinating, but this bit is particularly useful for folks interested in thinking clearly about the subject:

To understand what is really being fought over when we fight over the IQ gap between blacks and whites, its authors explain, you must think through an analogy. Imagine two wheat fields. Now imagine two genetically identical sets of seeds. (The analogy was first made famous by the Harvard evolutionary biologist and geneticist Richard Lewontin.) Now imagine each field is planted with these two identical seed stocks. Field No. 1 is given the best possible inputs: sunshine intensity, rain, soil nitrates, etc. Field No. 2 is given much less of all of the above. Within each field, inputs are kept uniform. Inevitably, the first field grows a healthier supply of grain than the second. But here is the rub: Within each field, the variation in outcomes is entirely hereditary. Between the two fields, the variation in outcomes in entirely environmental.

Metcalf goes on to calmly detail the compromised the research Saletan relied on, the problems in the Minnesota Twin study that Saletan never mentioned, and the findings of the APA panel he referenced. Well worth a read.

December 4, 2007 | Permalink

Comments

Yeah, good job by Metcalf. When I saw the information about Jensen and Rushton, I was surprised. Even setting aside Saletan's automatic contrarianism, citing dudes like that without realizing who they are is pretty sloppy.

Posted by: Neil the Ethical Werewolf | Dec 4, 2007 6:28:38 PM

It might be interesting to analyze what the Big Lies are in a screed associating intelligence and race.

What keeps the controversy alive, I suspect, is that the racists are making arguments, which coincide with the worldview of conservatives wishing for heriditary aristocracy.

Intelligence is one faculty ("g") -- one person is strictly smarter than another. The faculty is inherited, and there's little that society or the individual can or should do to alter where an individual is in the hierarchy. Richer and higher status individuals (and families, since this characteristic is heritable) are richer and have a higher status, because they are smarter. It is these propositions, which attract the attention of people already committed to a conservative point of view.

Posted by: Bruce Wilder | Dec 4, 2007 6:38:05 PM

Saletan is a stone cold racist. PERIOD.

Posted by: David Ehrenstein | Dec 4, 2007 7:12:14 PM

Imagine two wheat fields. Now imagine two genetically identical sets of seeds.

I'm not on one side or the other on this one, but this assumption that you start with identical seeds and that it is the enviornment, and not the seeds, is a political response to explain away this measureable difference, and not a scientific one.

I doubt if any real research on this can take place without the poltical claptrap trying to protect a favored group just as all scientific investigation into homosexuality has similarly been stymied because of the possible findings and protection of that group is also more important to detractors than a real answer.
It's amazing that the left would rather practice group identity politics and protect these groups at the expense of true scientific knowledge.

Posted by: El viajero | Dec 4, 2007 7:19:54 PM

Over at Slate, Stephen Metcalf has written a frankly devastating rebuttal to William Saletan's series of race, genes, and IQ.

Oh please. Metcalf doesn't know what he's talking about. He's a cultural critic. He makes claims that are directly contradicted by leading researchers in the field. He should stick to reviewing plays.

But at least he's not so stupid as to claim that a genetic component to black-white IQ differences has been discredited. The basic scientific consensus at this point is: There are substantial racial differences. We don't know what causes them. But genetic differences may be part of the cause.

Posted by: JasonR | Dec 4, 2007 7:21:19 PM

I might not be as absolute as Ehrenstein... but I've said all along I think the mistake here is going down this road to begin with - the notion that you can "prove" something scientific via IQ scores and race-based analysis. If nothing else, apologia aside, Saletan should have known better; that he didn't calls into question at least his judgment, if not, as David suggests, his prejudices.

Posted by: weboy | Dec 4, 2007 7:22:16 PM

It is sorta hilarious that Metcalf knows more about the science than Saletan.

I understand why, not wanting to complicate things, he didn't get into the underlying conceptual idiocy of the racist IQ claims. But it should be gone into more often, because a lot of the discussion operates on the assumption that there is some validity to the IQ test. That, of course, is a crock. There may be some sense in talking about IQs in terms of that set of behaviors needed in this or that society, but as far as an IQ that is related to a basic, universal human tool kit of behaviors, we simply don't have anything like that.
Any test for IQ that tested the human animal's natural intelligence would have to be constructed so as to be taken by all human animals, and would include things like social behavior as well as symbol manipulation. It's isn't the kind of thing you could test for by distributing individual pieces of paper to human beings in a classroom, of course - doing that would immediately cause your test to be invalid, since individual testing on pieces of paper in classrooms is, in itself, as much a social construct as, say, making one component of the test spearing fish with a bamboo spear. In other words, the racists mix up two things - genetics, which is about the range of natural behaviors for hominids, and IQ tests, which are products of the early 20th century, long before we had a good grasp of neurology, and at the height of the racist imperialist ethos. This, too, is why they cling to head size as a genetic target - because otherwise they are lost. Nobody knows what proper configuration of neural paths, chemical transmitters and the like go into the overall form of human intelligence. It is interesting, the hybrid of 19th century science with early 21st century racism - interesting how it brings into focus the kind of entrenched prejudices you see in the governing class.

Posted by: roger | Dec 4, 2007 7:50:43 PM

weboy,

What does Saletan claim has been proved "by IQ scores and race-based analysis?"

Posted by: JasonR | Dec 4, 2007 7:53:40 PM

roger,

But it should be gone into more often, because a lot of the discussion operates on the assumption that there is some validity to the IQ test. That, of course, is a crock. There may be some sense in talking about IQs in terms of that set of behaviors needed in this or that society,

IQ scores are correlated with academic performance, job performance, crime and many other life outcomes, and have obvious value for that reason, among others. Most intelligence researchers seem to believe that IQ tests effectively measure "general intelligence," the aspect of cognition associated with general problem-solving ability, rationality and abstract thought.

Your whole post is just kind of rambling and confused.

Posted by: JasonR | Dec 4, 2007 8:06:21 PM

Well if that's worth a read, then perhaps this is too.
Would require some actual synthesis.

Volume 82, No. 4 December 2007 The Quarterly Review of Biology

RETHINKING THE THEORETICAL FOUNDATION OF SOCIOBIOLOGY

David Sloan Wilson [&] Edward O.Wilson

keywords

altruism,cooperation,eusociality, group selection, human evolution,
inclusive fitness theory, kin selection, major transitions, multilevel selection, pluralism, sociobiology

Worth working through without so much of the buzz-think.
But having to do with congregate selection.

Posted by: has_te | Dec 4, 2007 8:10:03 PM

Metcalf: What does my anger have to do with your crappy research? Bingo!

Saletan, looking for a podium to preach from, choose a topic that he believed could frame himself as the 'above the fray' wise analyst and debunker. But his works habits (or his IQ?) didn't include doing anything like the thorough job of research required on a topic of this scientific complexity and politcal fervor. He just made an ass of himself - and lays him open to a full barrage of charges of inherent racsim on his part. A really wise man wouldn't open that drawer unless he was prepared to open lots of other drawers as well - and he was far too lazy to do that.

So, regardless of his 'actual' race biases (or not), he puts himself in the company of people of which there is no doubt that they are racists.

In my mind, this equates to David E's judgment:

Saletan is a stone cold racist. PERIOD.

So, can we just forget Saletan and consign him to obscurity? I sure hope so.

Posted by: JimPortlandOR | Dec 4, 2007 8:16:18 PM

JimPortland,

So, regardless of his 'actual' race biases (or not), he puts himself in the company of people of which there is no doubt that they are racists.

What people?

Lots of the usual smears, innuendos and namecalling here, but no real substantive criticism.

Posted by: JasonR | Dec 4, 2007 8:29:34 PM

Jason R, most researchers is no researcher.

However, you do show that you don't understand what a social construct is. Here are a few social constructs: "academic performance, job performance, crime and many other life outcomes." I'm wondering if you are including, in the crime category, kidnapping people, murdering them, raping them, and making them labor for you - a pretty good description of the economy of the American South for 200 years.

Let's try to imagine what possible story could exist behind the idiotic notion that skull size has to do with human intelligence. We will, for the moment, grant that races are natural kinds, although of course they aren't. We will also pretend that the IQ test developed in the late 19th century corresponds to natural human intelligence, instead of ways of getting ahead in a particular society (which is what you seem to think it is). Now, here's the story you would have to need to believe, in Darwinian terms. You 'd have to believe that at some point around 30,000 years ago, a subpopulation of humans, distinguished only by their skin color, suddenly spread a genetic feature that allowed them to be smarter - as defined by the symbol manipulations of the IQ test - but that the phylogenetic effect of that change lagged behind by, oh, 25,000 years. And even then, those poor black folks seemed to be forging ahead in Egypt. In fact, for around 4,000 years, the symbol manipulation of, say, writing, was far advanced on the Nile, while Germanic tribes were making due with ooga booga. But what the hell - they were passing on an absolutely useless genetic mutation cause God loved them. Then they learned to symbol manipulate to the extent that they learned a writing system. But until around five hundred years ago, they were dwarfed, in terms of their technological knowledge, by the Chinese, while North African, African, Latin American, Indian and other civilizations held about the same stock of knowledge. But then, man, their head size really kicked in.

Obviously, this is a stupid story. You can't believe that story and hold to anything like Darwinian evolution. Luckily, most people who believe that story are doubtful about Darwin anyway. Wasn't he a communist?

So, to shrink this argument down so it is just big enough for your head size to wrap around it, Jason:
1. claims about the genetic mutation within homo sapiens that leads to higher IQs among white folk have to have a sense of IQ that relates to all human beings - otherwise, they are making a claim that some natural genetic mutation was specifically designed for a social product. Like, we evolved feet to press the brake and gas pedal on a car. Now, I don't know, you might believe that we did evolve feet for just that reason, Jason R. But most experts in the Most Expert university don't.
2. But hypthosize for a second that the IQ test we have isn't laughable. The second claim is either: blacks (and we will forget that this term is so inclusive as to be meaningless) had some genetic mutation that lead to a lesser expression of human intelligence, or that whities (again, forget that this is meaningless as a natural kinds category) had some genetic mutation that lifted them to a higher IQ capacity.
3. Well, then, this had to have happened at a certain date. If we date it 30,000 years ago, about when pigmentation differences were kicking in, we are claiming that this mutation occured and was inherited even though it had no reproductive effect. How do we know that? Because it had no phylogenetic affect whatsoever. It didn't until about 4,000 B.C.
4. If, however, we aren't claiming that this happened 30,000 years ago, when did it happen? 4,000 years ago? How do you explain the apparent lag in such things as symbol manipulation among whitie groups, then, until almost 400 A.D.?
5. The end run around this is, really, Lamarckian evolution, in which you just throw in the towel and go all the way with pseudoscience and claim that acquired traits can be inherited. At that point, you not only are a racist, but you are a kooky racist.
Which, it seems to me, might well be your category, Jason R.

Posted by: roger | Dec 4, 2007 8:52:45 PM

roger,

Jason R, most researchers is no researcher.

No, most researchers is lots of researchers.

We will, for the moment, grant that races are natural kinds

I don't know what this is supposed to mean. What's "natural kind" supposed to mean? "Race," as the term is generally used, refers to categories of human beings primarily defined by empirical characteristics such as skin and hair color, the shape of facial features, and the geographical distribution of ancestors.

We will also pretend that the IQ test developed in the late 19th century corresponds to natural human intelligence,

As I said, most intelligence researchers seem to believe that IQ tests effectively measure cognitive abilities associated with general problem solving and reasoning.

The hypothesis that reproductively isolated human populations evolved significant differences in cognitive abilities, just as they evolved significant differences in other biological traits, is not obviously false. Indeed, it may well be true. That is why most intelligence researchers believe that genetic differences may account in part for the racial differences in IQ scores.

Posted by: JasonR | Dec 4, 2007 9:15:31 PM

Oh, the siren song of the 90th-percentile SAT kid who feels threatened when people denigrate IQ. JasonR, give a gander at http://cscs.umich.edu/~crshalizi/weblog/523.html

It is abundantly footnoted.

Get back to us with evidence.

Posted by: wcw | Dec 4, 2007 9:21:40 PM

JasonR, give a gander at ...

What about it, wcw?

Get back to us with evidence.

Evidence of what?

Posted by: JasonR | Dec 4, 2007 9:25:28 PM

Don't know why I would bother feeding the troll, but if he were half as smart as he seems to think he is, he could do a fucking Google search and find out that natural kind has a definition pertinent to the discussion. But that would assume some sort of ability not in evidence.

Posted by: OneirosDreaming | Dec 4, 2007 10:10:05 PM

"'Race,' as the term is generally used, refers to categories of human beings primarily defined by empirical characteristics such as skin and hair color, the shape of facial features, and the geographical distribution of ancestors."

This illustrates such a profound incomprehension of science that I'll put aside your various logical fallacies (ad hominem against Metcalf, begging the question about the value of IQ*) to address it.

Populations in the genetic sense are groups of interbreeding organisms that are generally dilineated by specific genetic markers. Assuming (almost certainly incorrectly**) that Asians, Caucasians and Africans represent three distinct genetic populations, there are several reasons that differences might have developed. Selection, mutation, and genetic drift are the main ones.
Drift occurs only in very small populations that are cut off from others, and is highly unlikely to apply to human populations (except perhaps exceptionally secluded ones like pygmies, who have obviously developed significant differences, though those might also have been caused by selection.)
Assuming there is a g, it is clearly coded for by hundreds of alleles, not one or two like eye color is, meaning it is almost impossible that a single mutation would alter the intelligence of an entire population significantly.
Such a mutation could affect a population if there was significant selective pressure. Selection could also explain a difference in intelligence if there was selective pressure for certain "smart" alleles to bo advantaged. Selection clearly explains certain differences, such as those in skin color. But those differences rely on powerful selective pressures (need of vitamin D vs. cancer resistance) that vary geographically. African, Asian, American, Australian and European societies functioned nearly identically for the vast majority of human existence. What different selective pressures would have led to more complex brain development in Europe? At what point to you imagine "reasoning" became advantageous in Europe in a way that it wasn't in Africa? Even if you accept the most darwinistic view of society imagineable, capitalism has been around only 500 years, and for most of that time impacted a phenominally small part of the population.

The fact is, there is no scientific category that is more than tangentially connected to "race", and there is no known genetic marker that is more than tangentially connected to "IQ." Both of those, in the context you are using, are social constructions, not scientific ones.

* You argue that IQ correlates to various outcomes and has value for that reason. Given that race also correlates to outcomes, you might as well skip the middle man and claim that race is linked to intelligence because it is linked to outcomes. IQ's fatal flaw is that it carries scientific weight, but isn't necessarily measuring anything beyond the same old environmental factors that it presumes to make irrelevant.

** Genetic data indicates that there is more genetic variation between populations within Africa than between almost any other "different races."

You're welcome to italicize various lines and force me into tortured explanations based on poor phrasing or random scientific studies you google up, but the fact is it's phenomenally clear that you have no scientific understanding of the genetics side of this debate.

Posted by: Sam L. | Dec 4, 2007 10:19:52 PM

Sam L,

This illustrates such a profound incomprehension of science

I'm afraid it's your comprehension of science that is profoundly flawed. Let's go through it...

Populations in the genetic sense are groups of interbreeding organisms that are generally dilineated by specific genetic markers. Assuming (almost certainly incorrectly**) that Asians, Caucasians and Africans represent three distinct genetic populations,

The boundaries are fuzzy, but racial categories are defined primarily by traits determined by genes, such as skin color and the shape of facial features, and by genetic heritage.

there are several reasons that differences might have developed. Selection, mutation, and genetic drift are the main ones.

You are confused. Mutation is a mechanism by which genetic change occurs in an organism. Selection and drift are processes that influence the incidence of a mutation in a population.

Assuming there is a g, it is clearly coded for by hundreds of alleles, not one or two like eye color is, meaning it is almost impossible that a single mutation would alter the intelligence of an entire population significantly.

This assertion is just scientific nonsense. Large numbers of genes may be involved in intelligence, but a small number of mutations, or even a single mutation, may cause significant (and even very severe) cognitive impairment. The severe cognitive impairment associated with Williams Syndrome, for example, is believed to be caused by variation in a small number of genes. A mutation in the single gene GDI1 is associated with an IQ below 70. In any case, such a mutation would not need to alter the intelligence of an entire population significantly to produce a significant difference in average intelligence between populations, so your assertion here would be irrelevant to the issue of differences in average intelligence between racial groups even if it were true.

Posted by: JasonR | Dec 4, 2007 10:51:02 PM

this assumption that you start with identical seeds and that it is the enviornment, and not the seeds, is a political response to explain away this measureable difference, and not a scientific one

You're denying that we are all descended from common ancestors?


As I said, most intelligence researchers seem to believe that IQ tests effectively measure cognitive abilities associated with general problem solving and reasoning.

Well, they sort of have to believe that, don't they? Someone who didn't believe there was such a thing as "general intelligence" would be unlikely to pursue a career researching it.

Doesn't mean it exists. There are people who devote their lives to researching ESP and poltergeists too.

Posted by: Jason C. | Dec 4, 2007 10:53:14 PM

You did this on purpose, didn't you Ezra? In fact, you and your buddy Matt Y.

Posted by: El Cid | Dec 4, 2007 11:08:15 PM

Not that I really want to defend JasonR, but I think it's pretty self-evident that 1) IQ is a social construct and that 2) IQ measures a type of intelligence that is casually correlated with academic performance and a number of other things. Those aren't contradictory things, right? Insofar as the causes of high or low academic performance are worth investigating, then there is a useful function for IQ tests to play in society.

Posted by: Korha | Dec 4, 2007 11:13:58 PM

"You are confused. Mutation is a mechanism by which genetic change occurs in an organism. Selection and drift are processes that influence the incidence of a mutation in a population."

Genetic change doesn't occur in a single organism, ever. Bio 101. Mutation is also by nature random, not affected by selection or drift. Selection determines the incidence of specific alleles (genes) within a population, and does not rely on mutation. Darwin's observed changes in the size of finches beaks were not the results of mutations in the genetic code, but the result of certain alleles becoming more or less common based on changing selective pressures. Drift can occur when a population is small enough that a mutation becomes common by chance, but more often also has nothing to do with it. It is generally caused by the loss of an allele through chance in an exceedingly small population, not a mutation.

Your final point is interesting. Certain mutations such as Downs syndrome or Williams syndrome can dramatically lower cognitive function.
If these conditions were much more common among African Americans, it could be argued that the "average" intelligence of African Americans was lower based on a small percentage suffering from defined ailment. But people with these conditions rarely reproduce. With both syndromes there is no half way. You either have the condition or you don't. Given that there is variation in intelligence within whole populations of whites and blacks, a large disabled group of African Americans can't explain it away. Besides, if there were a single mutated allele that made some African Americans dumber, it would be fairly easy to identify genetically.

I shouldn't bother explaining it, because you clearly haven't taken genetics (at least in the last 10-20 years) but the statistical results you're trying to justify wouldn't come from random mutations, they'd come from selection for certain "dumb" alleles (which would also exist in white populations) or selection against certain "smart" alleles. Over time, this would lead to higher instances of the "dumb" alleles in African Americans and higher instances of the "smart" alleles among whites. That would account for the possibility of very smart blacks and very dumb whites while still explaining a statistical advantage among whites. It's what you're looking for. The only problem is that there isn't any evidence of why it would have happened, especially considering how poorly racial delineations match up to genetic populations.

I encourage you to keep responding, if only because you're illustrating so effectively that you have no idea what the fuck you are talking about. Maybe you can get a job at Slate?

Posted by: Sam L. | Dec 4, 2007 11:24:08 PM

Sam L,

Such a mutation could affect a population if there was significant selective pressure. Selection could also explain a difference in intelligence if there was selective pressure for certain "smart" alleles to bo advantaged.

Yes. You got this right, at least. Well done.

Selection clearly explains certain differences, such as those in skin color.

As far as I'm aware, there is no scientific consensus on the cause of variation in human skin color. It could be the result of selection pressures, or it could random. We don't know yet. If you think you have evidence that this is a settled question, please produce it. I suspect this is just more guessing-presented-as-fact on your part.

But those differences rely on powerful selective pressures (need of vitamin D vs. cancer resistance) that vary geographically. African, Asian, American, Australian and European societies functioned nearly identically for the vast majority of human existence. What different selective pressures would have led to more complex brain development in Europe?

There was obviously significant variation in the ancestral natural environments of modern human racial groups, including significant variations in climate, predators, food supplies, and diseases. There were also significant differences in social organization. We simply don't know what, if any, selection pressures on cognitive ability these environmental differences between populations may have produced.

The fact is, there is no scientific category that is more than tangentially connected to "race", and there is no known genetic marker that is more than tangentially connected to "IQ."

Wrong on both. I just cited a specific example of a "genetic marker" (a particular mutation) that is very strongly correlated with IQ. And as I already said, skin color and the shape of facial features, which are both "scientific categories," are obviously strongly associated with racial categories.

You argue that IQ correlates to various outcomes and has value for that reason. Given that race also correlates to outcomes, you might as well skip the middle man and claim that race is linked to intelligence because it is linked to outcomes.

As I said, race is correlated with IQ scores and IQ scores are correlated with life outcomes. So, yes, race is correlated with outcomes. Are you disputing this? Most intelligence researchers believe that IQ scores measure a general factor of intelligence, but the correlation of IQ and life outcomes is independent of that hypothesis.

IQ's fatal flaw is that it carries scientific weight, but isn't necessarily measuring anything beyond the same old environmental factors that it presumes to make irrelevant.

I don't even know what this statement is supposed to mean. Racial differences in IQ scores may be caused entirely by environmental differences, but most intelligence researchers seem to believe that they are caused by a combination of environmental and genetic differences. I'm not aware of a single reputable intelligence researcher who believes it has been clearly established that environmental differences alone are the cause.

** Genetic data indicates that there is more genetic variation between populations within Africa than between almost any other "different races."

I see this statement a lot from people like you who think they understand the science but don't. The fact that genetic variation within a race may be greater than genetic variation between races tells us precisely nothing about average differences between races in particular genetic traits. Racial categories aren't associated with differences in quantity of genetic variation but with particular types of genetic variation (variation in the genes that code for skin color, facial features, etc.).

Posted by: JasonR | Dec 4, 2007 11:29:26 PM

JasonR and el viajero,

Your comments indicate that you are missing one or more fundamental points about the method of measuring genetic vs. environmental influences. So, for that matter, do most people on the opposite side of the fence who argue that genes account for little or none of the variation in performance or social standing.

Two basic points, the first of which I think everyone will accept:

1. There is no such thing as a complex trait of a living being which is controlled entirely by a gene or entirely by the environment. Genes always require environments for their expression, and without genes there is no complex trait of an animal that gets expressed in the environment. When we are talking about causes and not simply correlations among populations, there is no such thing as being caused 100% by genes or 0% by genes or even 50% by genes. This percentage language has no place. You can manipulate any trait (assuming you were powerful and knowledgeable enough) through either changing genes or changing environments.

Take a trait that you think is entirely or almost entirely caused by genes. Brown hair? Having two eyes? We know full well how things can go awry in the womb or through the wonders of modern chemicals so that people with genes normally expressed one way don't get the usual result. When we are talking about causes and causal mechanisms, we are not using the statistical language that genes explain X% and environment explains Y%, as though the role of one detracted from or operated independently of the other.

2. Where percentage of influence language has a place is in the domain of statistical correlations between variables. That's a problem, because most people who talk about how much of the variation in intelligence (or whatever) is "explained by" or "due to" genes really want to talk about causes and not mere correlations. Not only are we necessarily talking about statistical correlations and not causes, but the correlations are variable. That is, they do not rise to the status of statistical laws as you might find in quantum physics or chemistry.

Using the statistical methods of IQ researchers (regression, factor analysis, etc.) it turns out that the % of variation "explained" by genes depends on how you specify the environment in which the trait is measured. The more narrowly you specify the environment, the more of the remaining variation in a trait will be due to genes. This follows trivially from the concepts involved. In general, the more widely you allow an environment to vary, the less of the variation in traits will be "explained" by genes.

The field example cited by Ezra is a good one. If you very rigorously remove (or assume away) variation in the environment (as is the case within field A and field B), then any remaining variation in plant traits must be from their genes. Or, if you allow quite a lot of variation in conditions across the environments (as in field A compared to field B) but no variation in genes, then you can get a smaller or even non-existent "role" for genes. You could get the same result within field A if half of the seeds in that field were in a sunny area and the other half of identical seeds were in a shady area.

The response is often: well, we need to create a homogenous environment to do real science, so get rid of those confounding factors to get the right answer about what percent of variation in the real world genes "explain." But this is a fool's game for several reasons.

First, it's a cheat to insist on shrinking variation in the environment without shrinking variation in the genes. You are tilting the field in favor of finding a larger genetic "influence" using statistical methods.

The more environmental variation you remove, the more genes must "explain" what remains of necessity. Getting a result less than 100% only means you haven't removed all the variation in the environment. Getting a result of 75%, or whatever, isn't any deep discovery about human beings but instead says more about how much environmental variation you are allowing in your study.

Whether you come up with a genetic "influence" of 50% or 100% doesn't mean that environment was anything less than essential for the outcome. You could manipulate the outcome by manipulating the environment just as well as if your statistical methods had given you a genetic "influence" of 10%.

Moreover, to say that lower intelligence is associated with a particular gene in some circumstance has no implications for how that gene would be expressed in different circumstances. Some variations of a gene may do better in shade, others in sun. Some plants may require a little extra water to flourish. The only way to find out is to water it. But even if we were to learn this, it would not mean that genes "explained" or "accounted for" 50%, 80% or whatever of the variation as though this were some general law and not a highly specific statistical correlation.

People who need to get things done in the world don't use these nonsensical contrivances. They use confirmable laws and causal mechanisms.

Posted by: jd | Dec 5, 2007 12:02:01 AM

The comments to this entry are closed.