RSS
From ScottBarryKaufman.com:
STUDY ALERT: Meta-analysis of the heritability of human traits based on fifty years of twin studies
Tinca J C Polderman, Beben Benyamin, Christiaan A de Leeuw, Patrick F Sullivan, Arjen van Bochoven, Peter M Visscher & Danielle Posthuma
Despite a century of research on complex traits in humans, the relative importance and specific nature of the influences of genes and environment on human traits remain controversial. We report a meta-analysis of twin correlations and reported variance components for 17,804 traits from 2,748 publications including 14,558,903 partly dependent twin pairs, virtually all published twin studies of complex traits. Estimates of heritability cluster strongly within functional domains, and across all traits the reported heritability is 49%. For a majority (69%) of traits, the observed twin correlations are consistent with a simple and parsimonious model where twin resemblance is solely due to additive genetic variation. The data are inconsistent with substantial influences from shared environment or non-additive genetic variation. This study provides the most comprehensive analysis of the causes of individual differences in human traits thus far and will guide future gene-mapping efforts. All the results can be visualized using the MaTCH webtool.
The PDF can be read here.
Here’s James Thompson’s write-up of the findings.
English
Arabic
Chinese (Simplified)
Dutch
French
German
Hebrew
Hindi
Italian
Japanese
Korean
Persian
Polish
Portuguese
Russian
Spanish
Swedish
Turkish
Vietnamese
@Jayman Release the hounds!
Imagine that ! Mankind , men , in all their arrogance and conceit cannot fathom the the mysteries of creation . The eye cannot see itself .
A meta-analysis can be dodgy: many in medicine seem not to be above suspicion.
Just because the results of this one seem roughly consistent with my intuition is no reason to accept it uncritically.
Still, it starts off on the right foot by using “data” as a plural – always encouraging, that.
P.S. Why only fifty years of twin studies? They go back much further than that.
I made a living handling data for 30+ years. I never heard anyone say unselfconsciously "the data are..."
It is true of course that the Latin word data is a plural. If it's the Latin word you're using, though, please be so good as to put it in italics.Replies: @Foreign Expert, @dearieme
Way more than 0%.
From a public policy point of view — even from a parenting policy point of view — the lack of substantial effects from shared environment might actually be regarded as the key finding. No matter how much or little genes contribute, it’s only shared environment that’s manipulable by governmental or parental intervention.
If such interventions are doomed to failure, why pursue them?
For the same reason you look under the lamp post?
That’s the generous–toward the nurture folks–interpretation.
What it actually finds is most traits are around 50% heritable. But the other 50% isn’t really “nurture” it’s mostly … noise. Interactions with the environment that we don’t really understand and certainly can’t control.
~~~
Adequate nutrition, suppressed disease effects and adequate education are obvious bennies, but they’ve mostly been done. (What makes education inadequate now is the students you’re stuck with.) If we want to improve “social outcomes” now, what we need is … better people! Eugenics.
You got it wrong. Noise is not environment. The only source of noise are errors of measurement. What is the precision/repeatability of IQ test? If you test two twins one may perform better than usual and one may perform worse than usual. If the "smarter" twin performs worse than usual the estimate of heritability will be overestimated...and so on. When ∆IQ different of IQ's of two twins is smaller heritability is larger. If the set of twins is large and the span of IQ's is wide the effect of noise will have minimal impact on correlation because it will average out. So the variance that is not explained by correlation is due to non shared environment and not noise. While the variance explained by correlation is due to genes and shared environment. W/o doing studies on twins separated at birth you cannot decouple genes from shared environment which means you always end up with over estimated heritability. There are not that many separated twins at birth available, so most studies overestimate heritability. No studies can decouple intrauterine shared environment effect from heritability estimate.Replies: @res
I was about to make the same point. What jumps out to me is that there is little data here or elsewhere to support the shared environment claims. Sometimes, disproving an idea is just as important, maybe more important, than proving something. While none of us will see a day when people reject magic as the answer, hard data countering the claims of environment at least helps blunt some of it.
Just because the results of this one seem roughly consistent with my intuition is no reason to accept it uncritically.
Still, it starts off on the right foot by using "data" as a plural - always encouraging, that.
P.S. Why only fifty years of twin studies? They go back much further than that.Replies: @John Derbyshire
The English word “data: is not a plural. It is a mass noun, like “rice,” “sand,” and “grass.” We say “the rice is cooked,” not “the rice are cooked.”
I made a living handling data for 30+ years. I never heard anyone say unselfconsciously “the data are…”
It is true of course that the Latin word data is a plural. If it’s the Latin word you’re using, though, please be so good as to put it in italics.
I am in no way advocating for any specific “interventions” here, but a 17% variance due to environmental factors is hardly trivial.
I made a living handling data for 30+ years. I never heard anyone say unselfconsciously "the data are..."
It is true of course that the Latin word data is a plural. If it's the Latin word you're using, though, please be so good as to put it in italics.Replies: @Foreign Expert, @dearieme
And the reference point of a map or chart is called a datum. Two are called datums.
You pasted this form the abstract. That’s about as far you went, right? I do not think you understand it. By inconsistence they mean that correlation for MZ twins is not twice the correlation for DZ twins. That is all. This happens when gene effects are not really additive and/or when there are shared environment effects. Twin studies cannot control for the effect of shared environment. Only studies of twin reared separately (there are only few of them) can isolate gene from the shared environment effects, however still they cannot eliminate the effects of shared intrauterine environment. The bottom line is that the correlation of a given trait strength (it can be IQ) between twins has error and thus it is not directly heritability. Shared environment usually will increase the correlation, meaning that heritability will appear larger, i.e., overestimated. Non shared environment (providing it is random) always lowers the correlation and the actual (not apparent) heritability. If the actual heritability is X then 1-X fraction of variant is explained both by shared and non shared environments. But you can never measure X exactly. Usually you end up with X+∆. Both environments (shared and non shared) can be manipulated to maximize or minimize a given trait. Your pronouncement about public policy and governmental or parental intervention is nonsense.
What it actually finds is most traits are around 50% heritable. But the other 50% isn't really "nurture" it's mostly ... noise. Interactions with the environment that we don't really understand and certainly can't control.
~~~
Adequate nutrition, suppressed disease effects and adequate education are obvious bennies, but they've mostly been done. (What makes education inadequate now is the students you're stuck with.) If we want to improve "social outcomes" now, what we need is ... better people! Eugenics.Replies: @utu
What it actually finds is most traits are around 50% heritable. But the other 50% isn’t really “nurture” it’s mostly … noise. Interactions with the environment that we don’t really understand and certainly can’t control.
You got it wrong. Noise is not environment. The only source of noise are errors of measurement. What is the precision/repeatability of IQ test? If you test two twins one may perform better than usual and one may perform worse than usual. If the “smarter” twin performs worse than usual the estimate of heritability will be overestimated…and so on. When ∆IQ different of IQ’s of two twins is smaller heritability is larger. If the set of twins is large and the span of IQ’s is wide the effect of noise will have minimal impact on correlation because it will average out. So the variance that is not explained by correlation is due to non shared environment and not noise. While the variance explained by correlation is due to genes and shared environment. W/o doing studies on twins separated at birth you cannot decouple genes from shared environment which means you always end up with over estimated heritability. There are not that many separated twins at birth available, so most studies overestimate heritability. No studies can decouple intrauterine shared environment effect from heritability estimate.
http://slatestarcodex.com/2016/03/16/non-shared-environment-doesnt-just-mean-schools-and-peers/Replies: @utu
This is from 2015. Why post now?
I made a living handling data for 30+ years. I never heard anyone say unselfconsciously "the data are..."
It is true of course that the Latin word data is a plural. If it's the Latin word you're using, though, please be so good as to put it in italics.Replies: @Foreign Expert, @dearieme
“I made a living handling data for 30+ years. I never heard anyone say unselfconsciously “the data are…” Obvs you were working in the wrong places.
If such interventions are doomed to failure, why pursue them?
For the same reason you look under the lamp post?
You got it wrong. Noise is not environment. The only source of noise are errors of measurement. What is the precision/repeatability of IQ test? If you test two twins one may perform better than usual and one may perform worse than usual. If the "smarter" twin performs worse than usual the estimate of heritability will be overestimated...and so on. When ∆IQ different of IQ's of two twins is smaller heritability is larger. If the set of twins is large and the span of IQ's is wide the effect of noise will have minimal impact on correlation because it will average out. So the variance that is not explained by correlation is due to non shared environment and not noise. While the variance explained by correlation is due to genes and shared environment. W/o doing studies on twins separated at birth you cannot decouple genes from shared environment which means you always end up with over estimated heritability. There are not that many separated twins at birth available, so most studies overestimate heritability. No studies can decouple intrauterine shared environment effect from heritability estimate.Replies: @res
Really? You do understand that things like measurement noise show up as unshared environment in the analysis, right?
http://slatestarcodex.com/2016/03/16/non-shared-environment-doesnt-just-mean-schools-and-peers/
is there any deep reason why it should be a 50/50 universe? Some kind of balance or evolutionary steady state etc.?
http://slatestarcodex.com/2016/03/16/non-shared-environment-doesnt-just-mean-schools-and-peers/Replies: @utu
Yes, really. You can’t be serious.
Why was gravity not so strong the universe didn’t collapse right after the Big Bang but not so weak it didn’t drift off into a thin gruel?
Interesting stuff, like the universe, tends to be around the dividing line of various continuums.
Did you read the link? Do you have an argument (beyond the handwaving single case in your earlier comment) or are we back to your opinion being sufficient proof?
In case it isn’t obvious. Yes, I am serious.
The errors of IQ measurements, from what I read, seem to be negligible in comparison with the effect of non-shared environment. So they can be ignored. Are you saying that if heritability was derived from twin studies to be, say 70% that the remaining 30% are all measurement errors? If so, you must have very low confidence in the reliability of IQ testing.
I wrote "Noise is not environment." Are you saying this is not so? Noise will lower the measured correlation/heritability and thus increase the fraction attributed to non-shared environment but it does not make noise environment or does not make environment noise.
All other examples (Luck of the draw...) this guy looks at do not apply to heritability of IQ via twin studies. Though if some twins are actually not exactly identical genetically, then yes this is an error, noise that will reduce correlation that cannot be corrected w/o knowledge of fraction of shared genes.
Perhaps you should concentrate on the opposite issue of how the shared environment including the intrauterine phase should be dealt with to obtain true heritability. Until you do you do not know how much your heritability results are overestimated.
I’ve read the link. He doesn’t make point that you think he makes. Actually he says this:
I said this:
Is it much different?
The errors of IQ measurements, from what I read, seem to be negligible in comparison with the effect of non-shared environment. So they can be ignored. Are you saying that if heritability was derived from twin studies to be, say 70% that the remaining 30% are all measurement errors? If so, you must have very low confidence in the reliability of IQ testing.
I wrote “Noise is not environment.” Are you saying this is not so? Noise will lower the measured correlation/heritability and thus increase the fraction attributed to non-shared environment but it does not make noise environment or does not make environment noise.
All other examples (Luck of the draw…) this guy looks at do not apply to heritability of IQ via twin studies. Though if some twins are actually not exactly identical genetically, then yes this is an error, noise that will reduce correlation that cannot be corrected w/o knowledge of fraction of shared genes.
Perhaps you should concentrate on the opposite issue of how the shared environment including the intrauterine phase should be dealt with to obtain true heritability. Until you do you do not know how much your heritability results are overestimated.