Twins reared apart do not exist
Posted by tptacek 2 days ago
Comments
Comment by tptacek 2 days ago
Related, from last year: https://news.ycombinator.com/item?id=42200209
Comment by bell-cot 2 days ago
In other words - Junior should not be presumed to be smarter, fitter, more deserving, or destined for success, just because his parents did well. No matter how attractive that conclusion might sound, to people who consider themselves to be the "better" sort.
Comment by dragonwriter 2 days ago
Comment by bell-cot 2 days ago
The problem is most people's zealous desire to read socially self-serving conclusions into any data they can find on such subjects. And when they really like the Q.E.D. punchline, humans have very low standards for the "logic" used to reach it.
Comment by DaveZale 2 days ago
success has components of luck as in "right place, right time" for someone with the right qualities and connections, and many of the very successful are quick to admit this
I am a 3rd generation machinist along the paternal line, and although the machines I operate are in expensive labs, those my father and grandfather operated were probably just as challenging. Engineering also seems to often run in the family. How much is nature and nuture? "It varies" is a safe response
Comment by tptacek 2 days ago
Comment by DaveZale 2 days ago
Comment by tptacek 2 days ago
Comment by DaveZale 2 days ago
Comment by tptacek 2 days ago
Comment by DaveZale 2 days ago
Comment by tptacek 2 days ago
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
Comment by jagoff 1 day ago
"Whether or not you like MISTRA, “they left out the non-identical twins” is a side issue, and the broader evidence that IQ is substantially heritable is extremely strong.
Even if DZA were excluded, MZA alone provides a solid heritability estimate. The DZA sample was small and noisy, and MISTRA uses other twin and family data. Plus, meta-analyses confirm high IQ heritability, typically 0.5-0.8; Later structural-equation models applied to the full MISTRA cognitive dataset (MZA + DZA) estimated the heritability of general intelligence around 0.77, essentially the same ballpark as the original simple estimate. laplab.ucsd.edu. Intelligence is highly heritable, potentially reaching 80% in adulthood, supported by further studies like the Haworth et al. meta-analysis, showing age-related increases in heritability.
Strong evidence for polygenic and SNP heritability is shown from Plomin & von Stumm's 2018 research, showing how polygenic scores predict general intelligence.
Adoption studies further support the genetic influence, as adoptive siblings show weaker correlations compared to biological ones.
Environment is certainly NOT outweighed entirely or beyond merit, but evidence clearly shows the "uncomfortable" result having strong support despite desperate attempts to debunk.
In short; the op's story only works if including the DZA data actually drags the heritability estimate down into trivial territory. It does not.
Later analyses of the MISTRA sample that explicitly include both MZA and DZA twins and use full structural-equation models estimate heritability of general intelligence (g) at about 0.77 in adults. https://laplab.ucsd.edu/articles2/Lee2010.pdf
That is higher, if anything, than the original approximate 0.70. In other words:
The alleged “suppressed control group” does not turn the result into “no heritability”.
The more sophisticated models using that very same DZA data still say “IQ differences in this adult sample are heavily genetic.”
You just don't like this data, and don't want to accept it, because of the implications."
"For some people some subjects take on an irrational religious experience. Anything that is not a forceful agreement must be destroyed." Disagreement is not against the rules. Flagging a comment simply because you disagree with the content does appear to be against the rules, however, based on my reading.
Comment by davidbessis 1 day ago
Now the substance:
"The alleged “suppressed control group” does not turn the result into “no heritability”."
=> Of course not, did anyone claim there was no heritability? But
1/ It's not "alleged", it's printed black on white in the paper.
2/ There is no excuse for suppressing control group data (it's like suppressing the placebo arm of a drug study).
3/ It does turn the result into "junk", and it does establish a definite case of scientific malpractice among people arguing that IQ heritability is 0.70.
As for later analyses, they weren't the topic of my post, but that doesn't mean they're casher.
Comment by davidbessis 1 day ago
"I don't care if you find it fair. If you can't accept that genetics determines the entire organism (stress: entire) and does not stop at the neck, then you'd perceive my later criticisms as much worse than - gasp! oh great heavens! my pearls! - unfair. It is a bitter pill to swallow that some people were simply born with better hardware than yourself, one you are obviously railing against. Now, rush on and down vote this comment as well to lighten the burden of your cognitive dissonance. I'm also finding it difficult to reconcile your use of the flag/report on the parent comment versus the rules dictating and describing what is disallowed content. Disagreement is not against the rules. Perceived "fairness" is not in the rules."
Sorry to inform you that you don't understand the meaning of the verb "determine", as "genetics determines the entire organism" is scientifically wrong for obvious reasons: "influences", yes; "encodes proteins for", yes; but "determines", no.
And, no, I'm not railing against anyone's hardware as I'm pretty satisfied with mine.
Comment by jagoff 1 day ago
The evidence is overwhelming in that direction:
Twin & family designs: h² ~0.4–0.8, often ~0.7–0.8 in adult samples like MISTRA.
Adoption: people resemble their biological relatives more than their adoptive relatives in IQ, despite strong environmental differences.
Molecular genetics: polygenic scores and GCTA show real, replicable genetic signal for intelligence.
What it does not mean is that:
1. IQ is fixed at birth in a way that cannot be influenced by environment.
2. IQ differences justify any sort of discrimination or moral ranking.
Those two points are where people tend to clutch their pearls and panic. Just because we have noticed an uncomfortable truth does not mean that it is valid to use it in a discriminatory manner; the issue becomes that people will inevitably do so, thus those "in power" cripple the theory in the crib so as to avoid the fearful and uncomfortable implications.
Comment by tptacek 1 day ago
It's practically nobody's position that there's no linkage between genetics and intelligence (that would be weird indeed), but it's important for this comment for you to believe that's the counterargument --- otherwise the comment doesn't make sense.
Comment by jagoff 2 days ago
Comment by davidbessis 2 days ago
As it happens, I often run into trouble due to my conceptualist views on the foundations of mathematics (that is, I'm a hardcore physicalist & anti-Platonist cognitivist, which is quite rare among mathematicians), so I find your criticism particularly unfair.
Comment by jagoff 1 day ago
Comment by tptacek 1 day ago
Comment by red-iron-pine 1 day ago
created: 1 day ago
karma: -3
everything you need to know in a nutshell. username itself even has an implication.
Comment by tptacek 2 days ago
The truth is, for these kinds of studies, the whole enterprise might as well be metaphysical; people saying these kinds of things have formed a religious conviction about the heritability of behavioral traits, and their real objection is that science continues to be done on the topic at all. Ironic, given the frequency with which they complain that this science is suppressed.
(I too have a near-religious conviction about this subject, though in a different direction; I do not, however, pretend that conviction is itself a methodological critique!)
Comment by jagoff 1 day ago
Whether or not you like MISTRA, “they left out the non-identical twins” is a side issue, and the broader evidence that IQ is substantially heritable is extremely strong.
Even if DZA were excluded, MZA alone provides a solid heritability estimate. The DZA sample was small and noisy, and MISTRA uses other twin and family data. Plus, meta-analyses confirm high IQ heritability, typically 0.5-0.8; Later structural-equation models applied to the full MISTRA cognitive dataset (MZA + DZA) estimated the heritability of general intelligence around 0.77, essentially the same ballpark as the original simple estimate. laplab.ucsd.edu. Intelligence is highly heritable, potentially reaching 80% in adulthood, supported by further studies like the Haworth et al. meta-analysis, showing age-related increases in heritability.
Strong evidence for polygenic and SNP heritability is shown from Plomin & von Stumm's 2018 research, showing how polygenic scores predict general intelligence.
Adoption studies further support the genetic influence, as adoptive siblings show weaker correlations compared to biological ones.
Environment is certainly NOT outweighed entirely or beyond merit, but evidence clearly shows the "uncomfortable" result having strong support despite desperate attempts to debunk.
In short; the op's story only works if including the DZA data actually drags the heritability estimate down into trivial territory. It does not.
Later analyses of the MISTRA sample that explicitly include both MZA and DZA twins and use full structural-equation models estimate heritability of general intelligence (g) at about 0.77 in adults. https://laplab.ucsd.edu/articles2/Lee2010.pdf
That is higher, if anything, than the original approximate 0.70. In other words:
The alleged “suppressed control group” does not turn the result into “no heritability”.
The more sophisticated models using that very same DZA data still say “IQ differences in this adult sample are heavily genetic.”
You just don't like this data, and don't want to accept it, because of the implications.
Comment by tptacek 1 day ago