Brain size in the NY Times

There’s an article about brain size and hyenas in today’s NY Times. It turns out that the species of hyena with the greatest “social intelligence” also has the largest brains. How about that, a correlation between brain size and intelligence in non-human species?

When J. Philippe Rushton writes about the correlation between brain size and intelligence in humans, he doesn’t get an article in the NY Times. Nope, his research is called “pseudo-science” and he is called a “racist.” But this is what he says:

The preponderance of evidence demonstrates that brain size is correlated positively with intelligence and that both brain size and cognitive ability are correlated with age, sex, social class, and race.

Among other findings, people in high status occupations have bigger brains than people in low status occupations. Men have bigger brains than women. Asians have bigger brains than whites, who have bigger brains than blacks.

It makes sense that a bigger brain is a more powerful brain, all else being equal.

Comments

[Brandon Berg]

Asians have bigger brains than whites, who have bigger brains than blacks.

Really? In absolute terms, or relative to body size?

[Peter]

Notice that the article is entitled "Sociable, But Smart." When I saw that title on the Times' main page I thought the article was about some nerds who defy the odds and actually have normal social skills. Instead, it's about freakin' hyenas!

[Vim]

The NY times cites race and intelligence research fairly often, even really technical stuff that many of its readers probably can't follow. Three examples are:

1. "But some genes have more than one role, and some of these brain-related genes could have been selected for other properties."

http://www.nytimes.com/2007/06/26/science/26human.html

2. "Researchers Say Intelligence and Diseases May Be Linked in Ashkenazic Genes"

http://www.nytimes.com/2005/06/03/science/03gene.html?pagewanted=print

3. "Brain Size Is Linked to a Gene"

http://query.nytimes.com/gst/fullpage.html?res=9E06E0DB1439F937A1575AC0A9649C8B63

These are just a few. You can obviously find many more if you look harder.

Brain size is a crude proxy for brain morphology studies. I do not believe that brain size differences would mean much for humans without an accompanying relationship to brain morphology. As far as I know, there haven't been any studies saying that there were significant differences in gross brain morphology between races.

In the particular case of men versus women, we see that somewhat different architectures can still produce similarly intelligent groups.

By the logic of 'brain size' being equal to IQ, we would be concluding that the humans with the biggest biceps or the largest rackets are the best tennis players. Even in sports of raw power such as boxing or wrestling, this is not the case.

[godparticles]

Along these same lines, I viewed a NatGeo special last night called "Dog Genius" about the evolution of dogs from wolves with the main selective pressure coming from interaction with humans. Fascinating stuff for anyone interested in heritable (behavioral) traits. There were several standout stories, one where Russian(!) geneticists bred foxes to be less aggressive and found that the foxes "became" more like dogs (both physically and behaviorally), and a second about the capacity for dogs to learn using fast mapping which was long thought a humans-only domain.

[Vim]

godparticles:

Interestingly, dogs have smaller brains than wolves.

[keypusher]

By the logic of 'brain size' being equal to IQ, we would be concluding that the humans with the biggest biceps or the largest rackets are the best tennis players.

No one says that, though. People make the more modest claim that brain size is positively correlated with intelligence, which does appear to be true.

[John Smith]

What a larger brain means is that a greater evolutionary pressure was exerted thousands of years ago towards favoring intelligence and thinking as opposed to physical size. The brain consumes a tremendous amount of calories and is also dead weight. If there was NOT a corresponding increase in intelligence, the larger brain would be a tremendous liability.

In Race Differences in Intelligence by Lynn, he cites several studies that finds a .44 correlation between brain size and IQ.

Asians have bigger brains than whites, who have bigger brains than blacks.

Really? In absolute terms, or relative to body size?

In absolute terms. See beginning at pp. 20 in the link below.

http://psychology.uwo.ca/faculty/rushtonpdfs/PPPL1.pdf

[Vim]

Can we cite anybody (Rushton, Lynn, Lynn-Rushton, Rushton-Lynn) not related to the Pioneer fund?

[Joe Canada]

"Can we cite anybody (Rushton, Lynn, Lynn-Rushton, Rushton-Lynn) not related to the Pioneer fund?"

Here's one:

http://brain.oxfordjournals.org/cgi/content/abstract/awh696v1?maxtoshow=&HITS=10&hits=10&RESULTFORMAT=&fulltext=witelson&searchid=1135349423969_1109&FIRSTINDEX=0&journalcode=brain

[bobvis]

By the logic of 'brain size' being equal to IQ, we would be concluding that the humans with the biggest biceps or the largest rackets are the best tennis players. Even in sports of raw power such as boxing or wrestling, this is not the case.

Frankly, I believe this *is* the case. The average boxer is bigger than the average non-boxer. The average wrestler is bigger than the average non-wrestler. This is because there is a correlation between size and boxing and wrestling performance. Yes, this does not mean that the biggest boxer or wrestler will *always* win, but isn't what is meant by a correlation.

As more evidence, note that boxing and wrestling both have strict rules that disallow people above a certain weight from competing in certain weight classes. The reason for this is to prevent unfair (and therefore uninteresting) matchups. There is a reason the heavyweight champion of the world is considered to be the best boxer in the world...

[Peter]

There is a reason the heavyweight champion of the world is considered to be the best boxer in the world...

Interestingly, right now the heavyweight division is in a major slump, with no superstars but plenty of boring, mediocre fighters. People were booing at last month's much-hyped heavyweight "unification" bout in Madison Square Garden and some were so bored they got up early and left. Middleweight and welterweight is where the action is.

[The Engineer]

he cites several studies that finds a .44 correlation between brain size and IQ.

Social science is crap. In my work, if I got an R^2 value of .44, I would discard it as NOT correlated!

[Vim]

Joe Canada:

1. Your study confounds differences between families in a way that a study such as the following does not:

http://www.pnas.org/cgi/content/abstract/97/9/4932

2. We know that brain morphology responds to longterm cognitive habits:

http://www.pnas.org/cgi/content/abstract/070039597

3. A final problem is that as people become more specialized in a task, they use parts of the brain that others do not typically use, for example in:

http://www.mathematicalbrain.com/pdf/PRODIGY.PDF

[Vim]

The Engineer:

"Social science is crap. In my work, if I got an R^2 value of .44, I would discard it as NOT correlated!"

Haha. Yes, me too. Unfortunately, if they did this, they would have nothing to publish!

[JewishAtheist]

Vim:

The NY times cites race and intelligence research fairly often...

Clearly, you don't know how to play the "Waahhh the NYT is liberal!!" game. When they write something "liberal" it's business as usual. When they just as often write the opposite, it's "OMG the NYT wrote this!"

[Armed & Dangerous]

A much more interesting NYT article:

http://www.nytimes.com/2008/03/04/world/americas/04feet.html?ref=americas

[John Canada]

"http://www.pnas.org/cgi/content/abstract/97/9/4932"
Vincent Sarich is one of the authors of the study.

I read the Vincent Sarich/Miel book published after that study and I thought that i read that he said there were problems with the above study (i'll look it up when i get home) but he never explained why there were problems with the study which i thought was suspicious.

[godparticles]

Shout out to my man Carl Zimmer who wrote the NYT article. You all should check out his "diavlogs" on Bloggingheads.tv. He was on with Mike Gazzaniga this past Saturday. They talked about how neuroscience may one day require changes in law.

[bobvis]

Social science is crap. In my work, if I got an R^2 value of .44, I would discard it as NOT correlated!

Do you mean, your ability to deal with uncertainty and ambiguity is crap? :)

[Just teasing!]

Can we cite anybody (Rushton, Lynn, Lynn-Rushton, Rushton-Lynn) not related to the Pioneer fund?

Is there a list of "approved" institutions that fund studies?

[star]

>Social science is crap. In my work, if I
>got an R^2 value of .44, I would discard
>it as NOT correlated!

Please stick to engineering, you would be lost in the field of cognitive research and apparently wouldn't know why.

[John Canada]

"I read the Vincent Sarich/Miel book published after that study and I thought that i read that he said there were problems with the above study"

He goes into it from pp.221 - 223. Between families there was a positive r of .44 but within families 0. Then goes on to explain that processes happening over evolutionary time would be hard to replicate in real world lab time. Goes on to explain that IQ and brain size are "crude measurements".

Yet earlier in the book he avers that the correlation between Brain size and IQ is robust and that two groups differing in brain size by 1 std. would show a difference of 6 points in IQ since r is .40 (pp. 217).

I need someone with a higher g to explain it to me since i don't get what he's trying to say since within family differences are zero and the bit about evolutionary time vs. real-lab time. I would have thought that within family differences would be the end of it if replicated.

From Race by V. Sarich and Frank Miele.

[trumwill]

People were booing at last month's much-hyped heavyweight "unification" bout in Madison Square Garden and some were so bored they got up early and left. Middleweight and welterweight is where the action is.

I'm not a boxing fan, so help me out here. Is it that the fighters are bad or just that the match was boring? I was under the impression was that the problem with heavyweight boxing was that the fights were slower, not that the fighters weren't as good. Whenever I've seen fights in the various classes, I figured that while the smaller fighters were more fun to watch, the bigger fighters would win in a contest.

It happens that way in college football sometimes. Tulsa is a really fun team to watch and Ohio State is kind of dull, though I'd bet on the latter over the former to win 9 times out of 10 if they played each other.

[SFG]

I wouldn't say social science is crap just because they have low correlations. If you are dealing with a system with 20 input variables, each with an equal effect on the independent variable, they will each have an r-squared of .05. A person's income is affected by their race, sex, genes for personality(conscientiousness and neuroticism come to mind), genes for intelligence, height, effects due to parental environment on personality and intelligence, effects due to school environment on personality and intelligence, parents' SES, and what else have I forgotten?

Or let's look at IQ. Something as complex as the brain is no doubt affected by tens of genes, few of which are going to have an r-squared over .05 with IQ (think about it). And these no doubt interact with environment in weird ways. Do you think molecular biology and genetics are crap?

[SFG]

Furthermore, the r-squared goes down for variables with a nonlinear relationship. HS himself has shown (or at least argued) that intelligence has a nonmonotonic effect on income: good up to a point, then not so good.

Sorry. I'm all for throwing out this blank-slate silliness but I certainly don't believe the social sciences are crap because you can't predict a person's income with a one-line equation the way you predict the path of an electron in a magnetic field.

[SFG]

Hey Engineer: I'd also like to point out that y=x^2 from x=-1 to x=1 will have a R-squared of ZERO, so this nonlinear thing isn't as silly as you think.

[tommy]

Sorry. I'm all for throwing out this blank-slate silliness but I certainly don't believe the social sciences are crap because you can't predict a person's income with a one-line equation the way you predict the path of an electron in a magnetic field.

The problem is that the social sciences are crap because social scientists not only refuse the throw out the blank slate crap, but, dogmatic leftists that most are, they insist upon it at the expense of better approaches. For example, most sociologists still seem to prefer Marx to Darwin.

[SFG]

Yeah, I can't wait until the generation of 68 dies off and we can accept biology plays a role in sociology.

[Peter]

People were booing at last month's much-hyped heavyweight "unification" bout in Madison Square Garden and some were so bored they got up early and left. Middleweight and welterweight is where the action is.

I'm not a boxing fan, so help me out here. Is it that the fighters are bad or just that the match was boring? I was under the impression was that the problem with heavyweight boxing was that the fights were slower, not that the fighters weren't as good.

Not to bring the thread off-topic, so here is a pretty decent description of last month's (in)action. To sum up, one guy had a big size and skill advantage, but instead of fighting to win fought not to lose (there's a difference); meanwhile, the lesser fighter was content to avoid being knocked out even though he knew he'd lose by decision. Most people watching the fight were lucky to avoid death by boredom.

[Turambar]

The question is not who puts on the best show but who is the best fighter.

Middles have some very impressive speed and there is no way I could stand up to it before I got to haul off on them. But the question is more complicated when you factor to adversaries that are professional fighters. Could a middle/cruiser make a show of it in the heavy weight division? Roy Jones did take a HW title.

Over a certain size I think the heavy weights are equivalent. Fedor is the best fighter in the world and he isnt an incredible physical specimen- no Hong-Man Choi.

And this may not be a valid analogy to brain size at all....

[The Engineer]

Hey SFG, why the hell are you using linear regression for a nonlinear system. Your argument is... crap.

Vin is right. Linear regression is the only tool social scientists have, and if they didn't write papers about research with laughably low R^2, they would have nothing to publish.

But in terms of the social sciences actually KNOWING anything, they're crap.

Also, regarding outputs with lots of multiple inputs, good research would be able to control for the multiple variables. Then, controling the other variables and only varying the one of interest, you could show that it has statistically significant influence. The R^2 for the group of variables should be quite high.

star's reply was just PRICELESS. That's the level of thinking in the social sciences, I guess.

[bobvis]

But in terms of the social sciences actually KNOWING anything, they're crap.

If by knowing, you mean knowing with absolute precision, you're right. If you reject all knowledge that isn't arbitrarily precise though, you will be throwing out a lot of useful knowledge. If a guy laughing manically runs at you with a chainsaw, you won't run away because you don't *know* that he intends on killing you. He may in fact only want to kill you 20% of the time. The other 80% he may be just having fun getting attention. R^2 ~ 20%.

[The Engineer]

If people weren't trying to ram liberalism down our throats using social science "studies" with these bogus R^2, I really wouldn't care that it's all crap.

But they are.

Social Science is politics by another name. They use statistics to give themselves an air of authority, but when you really check their numbers, their research only confirms that they don't know anything.

[bobvis]

If people weren't trying to ram liberalism down our throats using social science "studies" with these bogus R^2, I really wouldn't care that it's all crap.

Would you be ok with social science if it was trying to ram some other -ism down our throats?

Social Science is politics by another name.

The social science *you* pay attention to probably is. If you use HS as your source for studies, you can bet on it. Is your argument against liberalism or low R^2 though? If the R^2 was .98, would you then accept the liberal agenda?

[Joshua Holmes]

Men have bigger brains than women. Asians have bigger brains than whites, who have bigger brains than blacks.

It makes sense that a bigger brain is a more powerful brain, all else being equal.

Which is certainly the case among blacks, whites, and Asians, as well as between men and women. *snort*

Middleweight and welterweight is where the action is.

There's no action left in boxing. No man in his 20s even watches it anymore; scandals killed it and MMA is burying the corpse.

[Vim]

Social science is politics on both sides. Who here really believes that republican/libertarian politics does not play into the 'race realist' view? In general, that's why I have no interest in work connected with the Pioneer fund. It obviously has political affiliations. Charles Murray of the Bell Curve is a devoted libertarian. Even some writers on 'race realism', who are a little closer at hand, are prone to rants about not only the intellectual inferiority of some minorities but also to the inferiority of minority cultures.

Are the hyena who do not have complex social organization inferior to the ones that don't? What would that even mean? No serious biologist would suggest such a silly idea.

Perhaps I am an old fashioned type of scientist but I think the technical questions of what IQ is and why humans differ in IQ and the reaction to that answer are two different issues, one is amenable to scientific inquiry and the other is not.

On the topic of statistics in social science, it's often quite bad. There is no reason to believe that this is less true among 'race realists'.

This review of IQ and Weath of Nations has a very good discussion of the bizarre things 'race realists' do in their desperation to generate data:
http://www.nature.com/hdy/journal/v92/n4/full/6800418a.html

I do not see a world where races differ in IQ as dramatically different. There is already lots of variation in IQ within a single race. The question of what a group of people of IQ 100 should do in relation to a group of IQ 85 is quite similar (I think) to what a group of IQ 180, should do in relation to a group of IQ 100. (The answer isn't disenfranchize and economically enslave the lower IQ group.)

[The Engineer]

If the R^2 was .98, would you then accept the liberal agenda?

It would be more difficult to dismiss out of hand, yes. Or, perhaps, one would have to get into more of the details as to what the control variables are, why they were chosen, etc.

Social science is politics on both sides.

Seeing as Social Science departments have literally NO conservatives (what's the R^2 on that?), it is hard to justify this statement.

Social Science is a bastion of leftists. Citing Charles Murray validates nothing.

I'll have to crack my copy of "The Bell Curve" and see what the R^2 is of the studies that Murray cites. I'm guessing they're more like 0.6 than 0.4.

[scottynx]

SFG writes: "If you are dealing with a system with 20 input variables, each with an equal effect on the independent variable, they will each have an r-squared of .05"

The Engineer never refuted this point. That head-size doesn't overwhelmingly dominate as a predictor of IQ doesn't mean it is "crap". Someday, many years from now, maybe researchers will be able to understand humans with an engineering precision to the point that we can get an R^2 of .98 for the head-size/IQ relationship because we know tons more variables to control for. It is a much harder slog getting there if you throw out all moderate correlations as crap.

[The Engineer]

The Engineer never refuted this point.

I didn't feel the need to, because I thought that it was awfully speculative.

Rather than "someday" discovering all the possible control variables to get an overall R^2 of .98 for IQ, what is more likely is that there is simply a ton of variation in IQ that is nothing more than randomness. It can't be controlled with more data.

Bryan Caplan recently presented some data on returns to education for African Americans, and his regression had an R^2 of .11, which he took as showing a correlation. That's even more craptistic than the IQ data!

[The Superfluous Man]

Here's some similar and very indirect evidence: http://www.economist.com/science/displaystory.cfm?story_id=E1_VNSRTVS&CFID=8047003&CFTOKEN=f9ec05617d6170bb-852EC3FC-B27C-BB00-0143C0838F0C5300

Title: Bats and balls
Subtitle: Bigger testes mean smaller brains

Basically, bigger balls means smaller brains, and bigger balls are a result of greater promiscuity, likely because bigger balls means more sperm which allows more insemination attempts. The pattern is also found in primates.

Sub-Saharan Africa is known to be a bit more than monogamous than say, Japan. Put two and two together, and you have a hunch(I wouldn't say evidence) that brain size, inversely to promiscuity, runs, in descending order, East Asian > White > Black

[SOL]

"Perhaps I am an old fashioned type of scientist but I think the technical questions of what IQ is and why humans differ in IQ and the reaction to that answer are two different issues, one is amenable to scientific inquiry and the other is not."

That statement is the product of an incurious or frightened mind. For "old fashioned scientists", nothing is off the table. Why should the fact that there is a difference in intelligence among races and the responses to that be forbidden(if that is indeed your belief)?

"I do not see a world where races differ in IQ as dramatically different"

Dramatically different from what? Does not Africa differ dramatically from Japan? In what way? How about Detroit vs. Grosse Point, MI? I've been to all 4 places and I can say that there is a difference. Here is a hint, 2 of the 4 places are "better" by any conceivable criteria(except of you are testing the efficacy of body armor).

"The question of what a group of people of IQ 100 should do in relation to a group of IQ 85 is quite similar (I think) to what a group of IQ 180, should do in relation to a group of IQ 100. (The answer isn't disenfranchize and economically enslave the lower IQ group.)"

No, it is most certainly not similar. You should know that.

[The Engineer]

Hey guys, I'm surprised that no one's posted hat sizes yet. 7-3/4 here. Ginormous head over here. You should see my kids' heads. Hard to believe they were the products of natural childbirth.

[scottynx]

"what is more likely is that there is simply a ton of variation in IQ that is nothing more than randomness. It can't be controlled with more data."

Engineer, when you get down to it, the human brain is just physics and by extension chemistry. Do you have some religious belief in it's perpetual mysteriousness? I doubt that, so you making that statement doesn't make much sense to me. Since it is physical as opposed to spiritual, sooner or later we will understand the underlying physics that leads to human thoughts and we will understand the physical differences between people's brains that leads to intelligence differences. We will also understand the genetic and environmental effects that lead to those physical differences as well.

It is profoundly unintellectual to see systematic trends and differences in such an important phenomenon like intelligence (what else has driven human evolution as importantly as intelligence?) but just throw up your hands and say it's all probably random.

[The Engineer]

Since it is physical as opposed to spiritual, sooner or later we will understand the underlying physics that leads to human thoughts and we will understand the physical differences between people's brains that leads to intelligence differences.

Can you explain "regression to the mean" with respect to IQ if IQ is totally dependent on non-random factors?

[scottynx]

"Can you explain "regression to the mean" with respect to IQ if IQ is totally dependent on non-random factors?

It is worth noting that the same exact question can be asked about regression to the mean in height and that the answer to both questions are the same.

Here is Rushton on regression to the mean:
"Basic genetic theory predicts that the IQ of offspring will regress towards the mean IQ of the population group from which the parents come. This has been amply documented for a number of physical traits in humans and in other species.

Regression to the mean is seen, typically, when individuals with high IQ scores mate. Their children tend to show lower scores than their parents. The converse happens for low IQ parents; they have children with somewhat higher IQs. This is because the parents pass on some, but not all, of their exceptional genes to their offspring. It is analogous to rolling a pair of dice and having them come up two 6’s or two 1’s. The odds are that on the next roll, you will get some value that is not quite as high (or as low)."
http://www.vdare.com/misc/051207_rushton_fallacy.htm

Here is Steve Sailer explaining the same thing:
"We each carry two sets of genes. You might have gotten lucky and gotten dominant genes that granted you a huge amount of some desirable trait. But your recessive genes are also a random selection from the average of your ancestors' genes, weighted by their closeness to you on the family tree. At the moment of your child's conception, you and your mates' four sets of genes are completely reshuffled. Thus, the children of the highly intelligent tend to have kids who aren't as bright as they are. That's why royal dynasties are founded by usurpers with exceptional talents, but quickly recede to nothing-specialness. In merciful contrast, the exceptionally dim tend to have children who are a little smarter than they are."
http://www.vdare.com/sailer/feminist_eugenics.htm

So highly intelligent people are likely to have hit the jackpot in the exact way their parents chromosomes were divied up and reshuffled (not just merely lucky in that their parents had a lot of good chromosomes, though this matters a lot too) Since their chromosomes will be divied up and reshuffled again in their children, it is not likely that their children will get another shuffle just as awesome.

Note that the higher the average genetic intelligence of the population, (ex: Ashkenazic Jews) the less astoundingly lucky the chromosome shuffling has to be to get the same level of intelligence, because more cards in the parents decks are likely to already be good cards. Children of higher Askenazic Jews thus regress towards the jewish population mean that is higher than gentile whites.

This probably goes a long way towards explaining why the children of blacks in the top income quintile do worse in school than the children of whites and asians in the bottom income quintile.

[The Engineer]

I need to think about this more, but I wonder if the difference between something that has a high degree of randomness and something that is just probalistically unlikely is any difference at all.

I mean, if there are genes for intelligence, but you only have a certain probability of getting them, how is that different than saying that there is a degree of randomness in intelligence heritability?

[anonymous 1]

You might be open- minded based on your last comment, Engineer.

You are cleary intelligent. I would ask you to consider the idea of non- mendelian inheritance (which scientists strongly think is the basis for IQ transmission and the explanation of regression to the mean in IQ inheritance):

http://en.wikipedia.org/wiki/Non-Mendelian_inheritance

This is a good genetics primer for basic definitions:

http://64.233.167.104/search?q=cache:eq4cC9kAh3kJ:www.uwstout.edu/faculty/parejkok/Intro/mendelian_and_non.htm+non+mendelian+inheritance+iq&hl=en&ct=clnk&cd=3&gl=us

Key statement:
"Almost all traits in humans are non-Mendelian."

This is a more simplified way of saying exactly what was being explained to you regarding regression to the mean in the previous comment.

[Vim]

Regression to the mean would apply in terms of extraordinary environments producing extraordinary people. So this is no great argument for genetics.

In terms of the Engineer's original question, of course there is no way to explain regression to the mean without there being some randomness. So far we have elucidated two sources of randomness, the procedure of inheritance and the effect of the gene once it has been inherited.

"if there are genes for intelligence, but you only have a certain probability of getting them, how is that different than saying that there is a degree of randomness in intelligence heritability?"

It's not. It's exactly as you said, there would be large amount of randomness or error in prediction that would be difficult to overcome. From a purely statistical perspective, you would never have very good prediction of individual IQs. However, for large populations the error in prediction would decrease.

Getting back to the original point of correlation, correlation is not causation. When one points this out, people claim they know this, but then they go back to talking about correlated variables as if one causes the other. Not only could A cause B, B cause A or some C cause both in a mechanistic sense, but also we must worry that the cause C could be related to the sampling procedure for observations.

In particular, there is a bias in terms of what type of environmental conditions each ethnic group lived under in their starting country and what kind of environments they were forced into upon moving to the West.

[star]

Engineer:

> I mean, if there are genes for intelligence, but you
> only have a certain probability of getting them, how
> is that different than saying that there is a degree
> of randomness in intelligence heritability?

Intelligence is about 80% heritable in adults. Like height, the heritability measurement increases with age. The 20% that is due to environmental causes is unrelated to the shared environment and, as such, does not reflect such social factors as family, school, neighborhood, or other non-biological factors. The environmental factors are largely due to exposure to toxins, disease history, and the intrauterine environment.

The degree of randomness in the IQs of children is such that it produces a Gaussian distribution around the regression point. The regression point is half way between the mean for the race of the parents and the mean IQ of the parents. This fact has been demonstrated in numerous longitudinal studies.

The number and identities of intelligence genes is unknown, but the leading researcher in this area (Robert Plomin) says that he does not expect to find more than 1% of the total variance associated with any single gene. He has already identified a small number of genes and knows that his scanning methodology would not miss any large contributors. For that reason (he has already done a lot of scanning), he is confident that no single gene will be found that accounts for a large amount of variance, except for those associated with serious retardation.

Assuming that you understand statistics, I would like to suggest that you do some independent reading so that you will not make silly comments about what is a large or small correlation, as applied to cognitive research. You are not up to speed.

[scottynx]

Since we are on the subject, here are some interesting comments about regression to the mean at Econlog by razib of gnxp.com:

[Bryan Caplan Says:]"After enough generations, all heritable characteristics should even out. But they"
don't. What actually happens?"

[Razib responds:]"huh? i know you know genetics dude, so you must be asking something more sophisticated than i can figure out (i can't think like an economist). populations regress back to the breeding population value, right? but the new breeding population value doesn't go all the way back to the old mean, and unless you apply more selection it won't. in terms of the biophysical substrate, genetics is discrete and not blending, so the part of the variance that is additive genetic alters the population's architecture. if you assume heritability remains constant you can just iterate the breeder's equation to model the population (mean value after selection = heritability X selection differential). in terms of assortative mating you start to increase heritability and the population starts to develop substructure so that there may be multiple breeding values to which individuals may regress. that is, the population turns multi-modal.

hope i didn't make myself look stupid ;-)"
http://econlog.econlib.org/archives/2007/11/is_this_a_swind.html

One more thing, I talked about "chromosome" shuffling in my comment above, but "gene" shuffling is actually correct because of the phenomenon of chromosomal crossover: "matching regions on matching chromosomes break and then reconnect to the other chromosome".
http://en.wikipedia.org/wiki/Chromosomal_crossover