Differences, within and without

Suppose that the narrow-sense heritability of IQ is 0.7 [ in typical western circumstances: no ball-peen hammers), and the non-genetic variation is almost all caused by mysterious unshared-environment effects – not the school you go to or the books in the house, but something essentially random, like somatic mutation, or randomness in development.

Then while a big fraction of variance in IQ is caused by genetic differences, quite a bit is not.

But now look at the difference between two groups. It’s entirely possible that those random forces – somatic mutation, noise in development, etc – are close to the same in both groups.

If so, the difference in the averages of the two groups would be almost entirely genetic, since the non-genetic factors would average out.

This entry was posted in Uncategorized. Bookmark the permalink.

55 Responses to Differences, within and without

  1. kot says:

    Bite sized troll post. Zzzz.

  2. whyteablog says:

    Would you say that, in truth, this situation might be pretty black and white?

  3. Garvan says:

    The post seems self evident, so perhaps you are surprised by the numbers, or expect us to be?

    Talking of ball hammers, in Cambodia male children are more likely to have learning disabilities related to head trauma than female children. The speculation I heard is that in Cambodia, society protects female children from domestic violence better than male children.

    So how do the numbers pan out when you take into account that perhaps 15% of children have learning disabilities in under developed countries (i.e. something as simple as a chronic ear infections so they cant hear the teacher causes children to drop out of school) compared to the west where it is what? About 5%? None of these numbers are published, are they? And in the west the disabled are not included in national IQ statistics, but in countries here there are no diagnostic facilities for mental disabilities, everybody is included.

    • Obviously he is referring to 1st world countries, esp. USA, where blacks share herd immunity advantage with whites.
      Effect of disabled is easily removed by picking median instead of mean.

  4. MawBTS says:

    It makes me think of the Minnesota Twin Study. The basic concept is that we study identical twins raised apart. If they are similar in adulthood, then genes = destiny. If they are different, then environmental factors = destiny.

    A common argument against the MST is “well, if you move a child from Generic White Family A to Generic White Family B, aren’t they probably still getting the same basic upbringing? How do you know you’re introducing any substantial change to their environment?”

    To an extent, this criticism is fair. Psychological research focuses a lot on America, maybe there’s interesting effects in specific populations, twin studies always have a low N, and more literature is always needed.

    But it also seems like Greg’s ball peen hammer example. You could construct a twin study that “proves” genes have zero influence on outcomes – raise Twin A in a normal environment, fire Twin B out of a cannon.

    But what would this hypothetical study teach us? Obviously humans are flesh and blood, and our environment ultimately rules us to the extent that it might kill us or permanently harm us. And it might be that if we raised one twin in Cleveland, Ohio and the other in a barrel of fire ants they’d have highly divergent outcomes, but that’s not the kind of environmental difference we need to think about for most problems.

    • “You could construct a twin study that “proves” genes have zero influence on outcomes – raise Twin A in a normal environment, fire Twin B out of a cannon. . . And it might be that if we raised one twin in Cleveland, Ohio and the other in a barrel of fire ants they’d have highly divergent outcomes.”

      Since the National Research Act of 1974, that’s more fun than a geneticist is allowed to have.

    • Jim says:

      Yes, traditionally it was supposed that even small differences in nurturing environment had a lot of effect. Type of toilet training, weaning age, magazines around the house, how much parents read to children, etc. At least it is pretty clear now that all of these things make little difference. So in any case the influence of nurturing environment, while it may not be zero, is much less than what was once commonly supposed.

    • Toad says:

      Is a childhood in a rubbled city in WW2 as good as fire ants? Did they all grow up to become gangster rappers?

    • Sterling Sorbet says:

      Most twin studies compare monozygotic to dizygotic in the same household, so one can assume roughly equal environmental conditions. The only other factor would be unshared environment.

  5. anon says:

    Ah. I see you saw that (mis)statement of Plomin as well…

  6. Isabel says:

    If these factors are random. I don´t think so. Without long term randomized controlled trials there is no way to tell it. And such study will never happen because it would violate human rights.

    • DevOps Dad says:

      While it is complicated with humans to compare genotypes and phenotypes, a simplified, yet convincing test can be made with dogs, and the test could cost around $100,000.

      Dog breeders have already done our work for us. The left tail of the IQ normal distribution of the smartest breeds of dog (Border Collie, Poodle) overlaps the right tail of the IQ normal distribution of the dumbest breeds of dog (Bulldog, Afghan). Take 20 just weaned puppies from the smartest and dumbest breeds and provide all with identical complex training and record everything.

      You can even provide the dumbest breeds with 50% more training, special foods or a more comfortable environment. When the dogs are two years old, test them on what they were taught.

      • iffen says:

        If you teach them how to fight other dogs, the BCs, Poodles and Afghans might form a united dog front against the Bulldogs.

  7. Either way, for the individual it’s all about how lucky or unlucky you are (counting genes as luck).

    • JMS says:

      The argument from the sanctimonious Left is that re-distributive taxing (esp. inheritance tax) allows gifted individuals from lower socio-economic backgrounds to achieve their potential.
      But this only boils down to a preference for genetic privilege over inherited privilege.
      Neither is earnt

  8. Anonymous says:

    “It’s entirely possible that those random forces … are close to the same in both groups”: Cochran’s Conjecture.

  9. Asher says:

    I have been saying exactly this for at least 10 years.

    “Environment” is just very localized variation within the larger structure of evolution of a type of life.

    IF humans are a product of evolution THEN everything human, including local variations, is a product of evolution. Thats about all laypeople need to know.

  10. Jonah says:

    Is this logic relevant in any way to Lewontin’s 15%/85% numbers?

  11. James James says:

    “unshared-environment effects – not the school you go to”

    According to Stuart Ritchie’s book “Intelligence: All That Matters”, unshared environment does include school. Because you and your brother might go to a different schools.

  12. RCB says:

    How much do the heritability studies worry about representativeness of, say, the nation? Everyone knows heritability is a population statistic that doesn’t necessarily generalize to another population. If the twins in your sample mostly come from higher socio-economic backgrounds, then you’re necessarily missing out on a lot of variation in family environment – like all the stuff in Charles Murray’s class bubble survey. So the shared-environment variance would be biased downward. On the other hand, high socio-economic families will also be more uniformly high IQ, which means that component of variance will also be underestimated. If both are underestimated to the same degree, then there’s no problem, since we ultimately compute a percentage (h^2) anyway. But that seems unlikely.

    Can anyone assure me that people worry about this?

    • JayMan says:

      National databases like the Add Health are fairly representative of the U.S. population.

    • Yudi says:

      Eric Turkheimer does. People have tried harder to do research on poorer social subgroups since his 2003 paper. In the US (but not any other First World country, apparently), family environment does matter more for intelligence as you go down the social scale, but it’s only really strong in the most deprived 2% of the population. I’d like to see more information gathered about heritability of IQ on the left side of the bell curve.

      Note that these results do not change the fact that, for all social groups, genes matter far more than politicians, public intellectuals, etc. acknowledge.

  13. JayMan says:

    “If so, the difference in the averages of the two groups would be almost entirely genetic, since the non-genetic factors would average out.”

    Indeed, and that works for a quite a few other traits in addition to IQ.

  14. Yudi says:

    Greg, how could this be proven or disproven?

    • Ziel says:

      You’d think the fact that these relative differences in intelligence have remained stable over generations, exist everywhere you look, are consistent across methods of measurement, and have been resistant to every intervention anyone has ever tried should be sufficient to out the matter to rest.

      Yet here we are.

      • Ziel says:

        “put” the matter to rest, of course.

      • Yudi says:

        That convinces precisely no one save the wicked. Results supporting the hereditary an hypothesis have to be fifty times stronger than for anything else in social science. Frustrating, but good for the quality of the research.

        • ziel says:

          Agree but my point is that It really wouldn’t matter what new evidence – and how much of it – were uncovered = the existing evidence is already overwhelming. The reality will never be accepted into the conventional wisdom.

          • gcochran9 says:

            Never is a long time.

            • ziel says:

              True, but I can’t imagine that it will be any “evidence” per se that will bring about a change in acceptance – I think it would have to be the result of some radical societal shift on the order of what happened in the 60’s.

              • gcochran9 says:

                As far as I can see fundamental attitudinal changes come along for no real reason at all. Contrast what a left-winger thought in 1948 with what they think today. And it’s not people changing hats: artsy-fartsy types were left-wing then as now, Jews were left-wing then as now, but back then they talked about the working man [and occasionally the Red Army] while now they talk about the glory of cutting someone’s dick off, or the hijab as the mark of true feminism.

                Sometimes I wonder about people.

              • Jim says:

                Perhaps the common theme is debunking whoever or whatever is considered most prestigious at a given time and validating that which is not so prestigious. Maybe the constant is just opposition to whatever is established opinion regardless of any rational arguments for or against it.

                The symbolic importance of stuff like hijabs is curious. Some people for example become extremely upset by women wearing burkas and want to outlaw wearing them. Most of these same people would probably not be upset by the traditional nun’s habit which is not much different. When I was a child it was still common to see nun’s attired that way.

                Personally I don’t much care whether a women wears a headscarf or not. I definitely think though that it would be unmanly for a man to wear one. I’m firm on that point.

                At any rate I don’t think that wearing hijabs is the problem. The incompatibility of Modern Western Culture with Islam is far deeper than women’s headdress fashions.

            • iffen says:

              Not if you run out of time.

  15. bob sykes says:

    The nonheredity component includes the error term as well as environmental causes, and the error term might be much larger than the environmental causes.

  16. As the average IQ in Africa is 0.5 – 1.0 SD below that of African-Americans (excluding the Igbo and perhaps a few others), I wonder how that happened. Did the brutal selection of slavery select for intelligence somehow? Do constant factors such as disease and nutrition depress development in a great percentage of sub-Saharan Africans, but when removed result in improved cognition, as picking up a board from the grass after a few days results in quick recovery?

    • vuzqk says:

      African-Americans have ~20% European ancestry on average. That could explain some portion of the gap (probably 20%) and rest could be nutrition, parasite load, and Flynn-style effects.

    • MawBTS says:

      I suspect it’s environmental. Why? Because of height.

      African Americans are taller than the west African populations they originate from, sometimes by several inches. Average African American height is 5’10 (I believe), versus 5’7 for Mauritania, 5’4 for Nigeria, 5’7 for Mali, 5’7 for Cote d’Ivoire, 5’6 for Ghana, etc. In NO African country does the national average equal 5’10 (although some African tribes do).

      Height, like IQ, is suppressed by poor nutrition/health. A standard deviation for height is 3.5 inches, so this fits a hypothesis of Africa “depressed” beneath its genetic potential by 0.5-1.0 SD across a range of quantitative traits.

      Either way, 3-4 inches is a lot of height and I don’t see a genetic explanation. I assume there’s no reason to suspect selection bias (meaning: no reason to think tall Africans were more likely to become slaves). Do you guys know of anything?

      Did the brutal selection of slavery select for intelligence somehow?

      At risk of being quoted out of context, slavery just wasn’t brutal enough.

      The Middle Passage killed about 15% of slaves…versus apparently 10% of the crew. If this had a selective effect, white Americans would be smarter than their counterparts in Europe. And the Middle Passage couldn’t affect IQ because the things that killed slaves (diarrheoia, etc) don’t give a fuck how smart or dumb you were. It’s not like Entamoeba histolytica gives you an IQ test before deciding to infect you. Slaves died basically at random.

      Once the slaves got to land…there’s some evidence that slaves were bred, but given what we know about race relations at the time (in 18th century Carolina it was illegal to teach a slave to write) I doubt breeding slaves for intelligence was in the game plan.

      • TWS says:

        If I remember correctly everybody gets taller in America. You’re no longer marrying Susy next door but Jane from two hundred miles away. Even if you are from the same country originally. At least those from the British Isles did.

      • DataExplorer says:

        If the slaves coming in had a mean IQ of 70, which is equivalent to mental retardation, you can imagine the slave owners wanting to breed that up to about 80 just to get more efficient workers. Not literally by selecting the most intelligent, but at least selecting the most efficient, which probably correlates with IQ.

    • Jim says:

      Lynn estimates the average IQ of SSA under optimal conditions to be about 80 basing that in part on the average IQ of 85 in African-Americans. In the US Virgin Islands with a largely African population but with environmental conditions generally far superior to SSA the average IQ is 78.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s