4 ms·

Do you believe in the evidence of partial heritability of cognitive function that we see in twin studies like TEDS? https://www.teds.ac.uk/about-teds https://w

by strken 3y ago

Do you believe in the evidence of partial heritability of cognitive function that we see in twin studies like TEDS? https://www.teds.ac.uk/about-teds https://www.teds.ac.uk/about-teds I think it can be simultaneously true that we suffer from believing ability is partly inherited, and also that we live in a world where that's actually the case. Though I also think it's cruel to tell a man with an IQ of 87 who's working as a chicken processor that the only reason he's not a Harvard professor is a bad education, bad parents, and his own failings.

em-bee 3y ago
of course both can be true. i already mentioned that but you are drawing the wrong conclusion. how would telling them that the reason they are not a harvard professor is because they don't have the talent for it be any better? what's cruel is that we tell them that they have no hope because their circumstances are beyond their control instead of supporting them with the resources they would need to develop the potential that they do have. all that statement is doing is looking down on them, quite regardless of what factors determine their potential. the problem is that people look for evidence of talent when we still don't know how to measure it. we simply have no idea whether the work in this competition is the result of talent or other factors. but we decide that these kid have talent and that other kids don't, when it's possible that others could create a similar work if they just had been given the opportunity. the discussion about talent gets in the way of helping people to develop. see also the discussion here: https://news.ycombinator.com/item?id=38287294 https://news.ycombinator.com/item?id=38287294 btw: i am skeptical about twin studies being enough to give us the answer: https://en.wikipedia.org/wiki/Twin_study#Criticism https://en.wikipedia.org/wiki/Twin_study#Criticism
- tptacek 3y ago
  The extent to which heritability "proves too much" seems like a particularly fertile source of objections.
- strken 3y ago
  I agree with the criticisms but don't agree that they invalidate twin studies to the extent that it's reasonable to believe cognitive traits are 100% nurture. Yes, some older studies used odd statistical modelling. Yes, the specific number they put on heritability might be biased by a few factors. Yes, people lie on self-report surveys. Yes, adoption studies are biased by who agrees to sign up. Yes, in utero conditions falsely appear as heritability in twin studies. These are all valid criticisms in that they affect the precision of the measurement, but they're not enough to fully write off heritability as a whole. We've identified SNPs that correlate with IQ and education attainment, too, so there's an obvious mechanism. I think it's better to tell people the truth: achievement is a murky blend of nature and nurture interacting together, the factors seem to be important in different proportions for different people, and we only understand it imprecisely and at a population level. Further complicating the matter, ability at birth can be broken down into heritability and environmental factors before birth. Our chicken processor probably tried his hardest and can feel good. His daughter can do better if she's lucky and works hard. There's no need to take an extreme view where "talent" (meaning something like potential at birth) must be 0% or 100% of achievement, or to tell that to people, because from what we know it's somewhere in the middle.
tptacek 3y ago
Two fun reads: https://sci-hub.wf/10.1002/hast.1008 https://sci-hub.wf/10.1002/hast.1008 http://bactra.org/weblog/523.html http://bactra.org/weblog/523.html
- strken 3y ago
  Probably goes without saying, but the g factor is so widely discredited as to be a strawman of IQ. Cars don't have an underlying s factor that determines how fast they go in all conditions with perfect correlation, and yet a measurement of their average performance on a set of tracks is still a useful measure for predicting performance on other tracks. So too for IQ and g factor. Neither am I arguing for total genetic determinism, just that the statement "it may well be that there are genetic differences. but until we have proof..." is ignoring the very real proof that cognitive traits (IQ and Big 5) are heritable to some extent. Even Turkheimer grudgingly admits that if you grab a bunch of IQ-correlated SNPs they can "predict about 10 percent of the variance". Weighing in on the 0% genetic side is an oddly strong position to hold, just as 100% would be.
  - tptacek 3y ago
    Heritability isn't remotely the same thing as genetic determinism. Things can easily be heritable with virtually no direct biological causative mechanism at all (wearing lipstick, for instance). SES is heritable! Check the first article out. Turkheimer, for what it's worth, is a very big name in this research field.
    - strken 3y ago
      Of course SES is heritable! If you believe IQ is heritable to a degree, genetic to a degree, and contributes to SES, then it's reasonable to expect SES to be both heritable and genetic to a degree. It would be surprising if there was some complicating factor that nullified the advantage from IQ but only in the case where the IQ-correlated SNPs were present.

by strken 3y ago

em-bee 3y ago

of course both can be true. i already mentioned that but you are drawing the wrong conclusion. how would telling them that the reason they are not a harvard professor is because they don't have the talent for it be any better? what's cruel is that we tell them that they have no hope because their circumstances are beyond their control instead of supporting them with the resources they would need to develop the potential that they do have. all that statement is doing is looking down on them, quite regardless of what factors determine their potential. the problem is that people look for evidence of talent when we still don't know how to measure it. we simply have no idea whether the work in this competition is the result of talent or other factors. but we decide that these kid have talent and that other kids don't, when it's possible that others could create a similar work if they just had been given the opportunity. the discussion about talent gets in the way of helping people to develop. see also the discussion here: https://news.ycombinator.com/item?id=38287294 https://news.ycombinator.com/item?id=38287294 btw: i am skeptical about twin studies being enough to give us the answer: https://en.wikipedia.org/wiki/Twin_study#Criticism https://en.wikipedia.org/wiki/Twin_study#Criticism

tptacek 3y ago
The extent to which heritability "proves too much" seems like a particularly fertile source of objections.
strken 3y ago
I agree with the criticisms but don't agree that they invalidate twin studies to the extent that it's reasonable to believe cognitive traits are 100% nurture. Yes, some older studies used odd statistical modelling. Yes, the specific number they put on heritability might be biased by a few factors. Yes, people lie on self-report surveys. Yes, adoption studies are biased by who agrees to sign up. Yes, in utero conditions falsely appear as heritability in twin studies. These are all valid criticisms in that they affect the precision of the measurement, but they're not enough to fully write off heritability as a whole. We've identified SNPs that correlate with IQ and education attainment, too, so there's an obvious mechanism. I think it's better to tell people the truth: achievement is a murky blend of nature and nurture interacting together, the factors seem to be important in different proportions for different people, and we only understand it imprecisely and at a population level. Further complicating the matter, ability at birth can be broken down into heritability and environmental factors before birth. Our chicken processor probably tried his hardest and can feel good. His daughter can do better if she's lucky and works hard. There's no need to take an extreme view where "talent" (meaning something like potential at birth) must be 0% or 100% of achievement, or to tell that to people, because from what we know it's somewhere in the middle.

tptacek 3y ago

Two fun reads: https://sci-hub.wf/10.1002/hast.1008 https://sci-hub.wf/10.1002/hast.1008 http://bactra.org/weblog/523.html http://bactra.org/weblog/523.html

strken 3y ago
Probably goes without saying, but the g factor is so widely discredited as to be a strawman of IQ. Cars don't have an underlying s factor that determines how fast they go in all conditions with perfect correlation, and yet a measurement of their average performance on a set of tracks is still a useful measure for predicting performance on other tracks. So too for IQ and g factor. Neither am I arguing for total genetic determinism, just that the statement "it may well be that there are genetic differences. but until we have proof..." is ignoring the very real proof that cognitive traits (IQ and Big 5) are heritable to some extent. Even Turkheimer grudgingly admits that if you grab a bunch of IQ-correlated SNPs they can "predict about 10 percent of the variance". Weighing in on the 0% genetic side is an oddly strong position to hold, just as 100% would be.
- tptacek 3y ago
  Heritability isn't remotely the same thing as genetic determinism. Things can easily be heritable with virtually no direct biological causative mechanism at all (wearing lipstick, for instance). SES is heritable! Check the first article out. Turkheimer, for what it's worth, is a very big name in this research field.
  - strken 3y ago
    Of course SES is heritable! If you believe IQ is heritable to a degree, genetic to a degree, and contributes to SES, then it's reasonable to expect SES to be both heritable and genetic to a degree. It would be surprising if there was some complicating factor that nullified the advantage from IQ but only in the case where the IQ-correlated SNPs were present.