Professor Steve Hsu, a quantum physicist, did a lot of genetics work and modeling in China and the USA...He concluded that thousands of genes have some effect on intelligence, and that heritability is somewhere between 70 and 80%...A key finding is that health is strongly correlated with intelligence, implying a lower level of adverse mutations..(Forgive me Professor Hsu if I have oversimplified)...all of which is consistent with your conclusions (and common sense)...And these twin studies all have the flaw that adverse factors such as head injuries or other life threatening events are generally unknown to the researchers...
Forget inheritance. The real question - does it travel with you when reincarnated? If not, then Kamala Harris could in fact, be the reincarnation of Leonard DaVinci. But, if so, then the specter of Kamala Harris will torment future generations and civilizations until the end of time. I don't know what outcome is scarier.
Reported average IQ scores vary across different studies and geographical locations, but general patterns in the United States and globally suggest the following approximate averages:
East Asians/Asian Americans: Average IQ around 106
White/European Americans: Average IQ around 100-103
Hispanic/Latino Americans: Average IQ around 89-94
Black/African Americans: Average IQ around 85-92
Sub-Saharan Africans (globally): Average IQ around 70
“
Rationale I saw was that Neanderthals didn’t interbreed with Sun-Saharan Africans and therefore didn’t pass on their intelligence genes to that group.
I think the paper in question has proven not much about IQ but has demonstrated that genetics is a false science. How on earth did anyone claim "heredity" of IQ without examining IQs of parents??? The study started with an assumed (but false) premise, that IQ is somehow "coded" by imaginary "genes" and oh well, that was fun...
The IQ test that is the best is the one I did best on, obviously. Such studies need to be vetted carefully, since the belief that intelligence is heritable is obvious racism and any study that seems to validate that belief means the authors are racists and need to be sent to that place of weeping and teeth gnashing.
I recall my first IQ test was in 8th grade. But they never gave us the results. These things are silly anyway. IQ has nothing to do with surviving in the real world. The higher the IQ the greater the arrogance, one-up-manship and general stupidity in practical affairs. I wonder if at age 75 my IQ has changed.
Took a quick look at the paper, it appears a study of studies. Couple of questions (more rhetorical than expecting any answer) : 1) are all IQ tests identical?, 2) does one's IQ test results vary across multiple tests, 3) were the twins given IQ tests at the same age?
Differences in tests, differences in scores for a single person retaking any test at the same age, and differences in scores on the same test at different ages are all sognificant confounding factors that make the meta-analysis uninterpretable for all but big signals, especially with the small number of subjects.
IQ tests correlate typically 0.8 to 0.9, but it can be rather less for near-ceiling or floor scores, which makes it surprisingly unlikely to meet a high cut-off score on both. (See ultra hogh IQ society psychometrician Grady Towers' on using multiple tests for admission: http://miyaguchi.4sigma.org/gradytowers/societies.html , though the specific numbers are wrong, the diagram gives an intuitive explanation. His followup is allegedly more accurate, but still does not account for ceiling effects or falling correlation between tests in the tails: http://miyaguchi.4sigma.org/gradytowers/followup.html ; Grady also sparked my interest in Rasch measures of intelligence over 20 years ago: http://miyaguchi.4sigma.org/gradytowers/about_rasch.html)
The standard error of measurement on the WJ-IV tests (a top individual IQ test), IIRC, gives a 95% confidence interval of nearly 1 s.d. (+/- 0.5 s.d.) for a single test score.
The WJ and SB tests' Rasch-measure (ability-criterion, "absolute", not age-normed) scores show that the size of the s.d.s and hence the size of IQ points varies substantially with age, with youngest and oldest having s.d.s up to 60% bigger than the middle-school minimum. Further, the right and left tails always have different s.d.s, the left being larger. (Riverside, the tests' publisher, keeps track of both , but the data released is only overall s.d.s) (See my post: https://enonh.substack.com/p/real-measures-of-intelligence for more on Rasch as practically applied to compare intelligence at different ages, also has link to WJ tech. manual with data, which is a better education in psychometrics than any textbook I've seen.)
So "If environment did nothing, and inheritance was everything (for taking IQ tests), then all scores would fall on the one-to-one line; " is not quite correct, it will have dispersion away from the line, and how much likely can only be crudely approximated.
The discussion about education being nurture likely has some truth, but to the extent IQ tests have questions with answers learned from education, then they are more testing education than intelligence (although sometimes related).
Education does not raise intelligence or IQ long-term, though it may temporarily raise scores on some types of tests a small amount (0.2 s.d. state achievement test improvement for teacher +1s.d. in improving test scores, much less for ex ante predictions of teacher ability). With typical ranges of environments in US, about none of the difference in performance is due to shared environmental effects: “unshared environment” is mostly not environment but unknown and perhaps unknowable - random idiosyncratic effects, epigenetic effects, even inborn soul.
The most g-loaded intelligence test questions are not the “culture fair” ones (Raven's is pretty lousy, actually) but things like general knowledge, vocabulary, and analogies. High-intelligence people learn the specific knowledge required regardless of education or race or other environment. Any question that special knowledge will allow a generally unintelligent person to get right gets weeded out in the validation process of test construction, as are questions that otherwise smart people too often get wrong.
I applaud the authors for their restraint for steering clear of the p-hacking trap that plagues so much modern research. Let's be real…an IQ score is nothing more than a number associated with a specific test. It measures precisely one thing, how well someone performs on that test. Full stop. It doesn't capture "intelligence," whatever that even means. And who gives a damn? Zoom out and look at the big picture: we humans aren't exactly paragons of smarts. We're barely near the bottom on the cosmic intelligence scale. Just look around and check out the wreckage of our world…the endless wars, the environmental suicide, the clown-show politics, the viral stupidity online. If anything, we've devolved into dumber creatures, not evolved into wiser ones. We're the species that invented the smartphone... and then uses it to spew hate and outrage while the planet burns. So bravo to the authors for not inflating their claims. The rest of us could use a heavy dose of that humility.
Mr. Briggs is always a good read; highly recommended.
As he extrapolates,
"IQ, as I often remind us, is a score on a test. The score itself is either a crude measure of one part of intelligence, or the crude smoothing (mushing together) of more than one part, conjoined with the natural variability found in test taking (underdone bits of potato, hidden cultural clues, etc.). IQ is not intelligence."
IQ, in my understanding, seems to be a tool for control and a substitute for self-respect. :)
I usually joke that my IQ ranges between plus/minus 360 (please, don't ask me why not 365 :) ):
And no, Jared C. Horvath don't know each other, and are not related, at least not that I know of, and we are certainly not identical twins. :)
Joking aside, as the article also insinuates, comparing IQ figures makes a little sense only in traditional cultures in identical environments, in which case I can't see how problem-solving abilities can be separated as born or bred. If they differ (and traditional cultures are mostly things of the past), the variables become uncontrollable for analytical purposes.
To make it a bit more clear, when it comes to two self-improving AI of the same computer architecture, and their initial programming assigns them to learn about the same type of problem-solving, their learning environments inescapably determine their differences. In the same controlled learning environment, their "IQ" decisions will NOT diverge...
Another factor is that in "normal" cases, the parent(s) demonstrate ways to handle conflicts and solve problems, and the child will imitate them, until (s)he doesn't. Is there going to be a switch or not, who knows? Still, inherited and acquired skills become inseparable.
Please, allow me to ask: Why is it important to differentiate between the two?
Apologies, if you believe my comment is too long. I can take it off and publish it in an article to which I would only leave a link here with a brief intro to the way it's related to your post.
You left out an undigested bit of beef, a blot of mustard, and a crumb of cheese. Sorry, it will have to be redone with these additional variables.
All,
Fixed the missing picture.
Professor Steve Hsu, a quantum physicist, did a lot of genetics work and modeling in China and the USA...He concluded that thousands of genes have some effect on intelligence, and that heritability is somewhere between 70 and 80%...A key finding is that health is strongly correlated with intelligence, implying a lower level of adverse mutations..(Forgive me Professor Hsu if I have oversimplified)...all of which is consistent with your conclusions (and common sense)...And these twin studies all have the flaw that adverse factors such as head injuries or other life threatening events are generally unknown to the researchers...
Forget inheritance. The real question - does it travel with you when reincarnated? If not, then Kamala Harris could in fact, be the reincarnation of Leonard DaVinci. But, if so, then the specter of Kamala Harris will torment future generations and civilizations until the end of time. I don't know what outcome is scarier.
“Average IQ Scores by Group
Reported average IQ scores vary across different studies and geographical locations, but general patterns in the United States and globally suggest the following approximate averages:
East Asians/Asian Americans: Average IQ around 106
White/European Americans: Average IQ around 100-103
Hispanic/Latino Americans: Average IQ around 89-94
Black/African Americans: Average IQ around 85-92
Sub-Saharan Africans (globally): Average IQ around 70
“
Rationale I saw was that Neanderthals didn’t interbreed with Sun-Saharan Africans and therefore didn’t pass on their intelligence genes to that group.
I think the paper in question has proven not much about IQ but has demonstrated that genetics is a false science. How on earth did anyone claim "heredity" of IQ without examining IQs of parents??? The study started with an assumed (but false) premise, that IQ is somehow "coded" by imaginary "genes" and oh well, that was fun...
I love that you liberate your readers from the obligation of wading through the minutiae.
The IQ test that is the best is the one I did best on, obviously. Such studies need to be vetted carefully, since the belief that intelligence is heritable is obvious racism and any study that seems to validate that belief means the authors are racists and need to be sent to that place of weeping and teeth gnashing.
I recall my first IQ test was in 8th grade. But they never gave us the results. These things are silly anyway. IQ has nothing to do with surviving in the real world. The higher the IQ the greater the arrogance, one-up-manship and general stupidity in practical affairs. I wonder if at age 75 my IQ has changed.
Took a quick look at the paper, it appears a study of studies. Couple of questions (more rhetorical than expecting any answer) : 1) are all IQ tests identical?, 2) does one's IQ test results vary across multiple tests, 3) were the twins given IQ tests at the same age?
1) No. 2) Yes. 3) I don't know, but doubt it.
Differences in tests, differences in scores for a single person retaking any test at the same age, and differences in scores on the same test at different ages are all sognificant confounding factors that make the meta-analysis uninterpretable for all but big signals, especially with the small number of subjects.
IQ tests correlate typically 0.8 to 0.9, but it can be rather less for near-ceiling or floor scores, which makes it surprisingly unlikely to meet a high cut-off score on both. (See ultra hogh IQ society psychometrician Grady Towers' on using multiple tests for admission: http://miyaguchi.4sigma.org/gradytowers/societies.html , though the specific numbers are wrong, the diagram gives an intuitive explanation. His followup is allegedly more accurate, but still does not account for ceiling effects or falling correlation between tests in the tails: http://miyaguchi.4sigma.org/gradytowers/followup.html ; Grady also sparked my interest in Rasch measures of intelligence over 20 years ago: http://miyaguchi.4sigma.org/gradytowers/about_rasch.html)
The standard error of measurement on the WJ-IV tests (a top individual IQ test), IIRC, gives a 95% confidence interval of nearly 1 s.d. (+/- 0.5 s.d.) for a single test score.
The WJ and SB tests' Rasch-measure (ability-criterion, "absolute", not age-normed) scores show that the size of the s.d.s and hence the size of IQ points varies substantially with age, with youngest and oldest having s.d.s up to 60% bigger than the middle-school minimum. Further, the right and left tails always have different s.d.s, the left being larger. (Riverside, the tests' publisher, keeps track of both , but the data released is only overall s.d.s) (See my post: https://enonh.substack.com/p/real-measures-of-intelligence for more on Rasch as practically applied to compare intelligence at different ages, also has link to WJ tech. manual with data, which is a better education in psychometrics than any textbook I've seen.)
So "If environment did nothing, and inheritance was everything (for taking IQ tests), then all scores would fall on the one-to-one line; " is not quite correct, it will have dispersion away from the line, and how much likely can only be crudely approximated.
Thanks for the detail.
The discussion about education being nurture likely has some truth, but to the extent IQ tests have questions with answers learned from education, then they are more testing education than intelligence (although sometimes related).
Education does not raise intelligence or IQ long-term, though it may temporarily raise scores on some types of tests a small amount (0.2 s.d. state achievement test improvement for teacher +1s.d. in improving test scores, much less for ex ante predictions of teacher ability). With typical ranges of environments in US, about none of the difference in performance is due to shared environmental effects: “unshared environment” is mostly not environment but unknown and perhaps unknowable - random idiosyncratic effects, epigenetic effects, even inborn soul.
The most g-loaded intelligence test questions are not the “culture fair” ones (Raven's is pretty lousy, actually) but things like general knowledge, vocabulary, and analogies. High-intelligence people learn the specific knowledge required regardless of education or race or other environment. Any question that special knowledge will allow a generally unintelligent person to get right gets weeded out in the validation process of test construction, as are questions that otherwise smart people too often get wrong.
I applaud the authors for their restraint for steering clear of the p-hacking trap that plagues so much modern research. Let's be real…an IQ score is nothing more than a number associated with a specific test. It measures precisely one thing, how well someone performs on that test. Full stop. It doesn't capture "intelligence," whatever that even means. And who gives a damn? Zoom out and look at the big picture: we humans aren't exactly paragons of smarts. We're barely near the bottom on the cosmic intelligence scale. Just look around and check out the wreckage of our world…the endless wars, the environmental suicide, the clown-show politics, the viral stupidity online. If anything, we've devolved into dumber creatures, not evolved into wiser ones. We're the species that invented the smartphone... and then uses it to spew hate and outrage while the planet burns. So bravo to the authors for not inflating their claims. The rest of us could use a heavy dose of that humility.
This wasn't the authors; only me. They did have a regression in their paper, which I thought best to ignore.
I liked what you did so I downloaded the paper and read it.
Mr. Briggs is always a good read; highly recommended.
As he extrapolates,
"IQ, as I often remind us, is a score on a test. The score itself is either a crude measure of one part of intelligence, or the crude smoothing (mushing together) of more than one part, conjoined with the natural variability found in test taking (underdone bits of potato, hidden cultural clues, etc.). IQ is not intelligence."
IQ, in my understanding, seems to be a tool for control and a substitute for self-respect. :)
I usually joke that my IQ ranges between plus/minus 360 (please, don't ask me why not 365 :) ):
https://rayhorvaththesource.substack.com/p/are-you-intelligent
And no, Jared C. Horvath don't know each other, and are not related, at least not that I know of, and we are certainly not identical twins. :)
Joking aside, as the article also insinuates, comparing IQ figures makes a little sense only in traditional cultures in identical environments, in which case I can't see how problem-solving abilities can be separated as born or bred. If they differ (and traditional cultures are mostly things of the past), the variables become uncontrollable for analytical purposes.
To make it a bit more clear, when it comes to two self-improving AI of the same computer architecture, and their initial programming assigns them to learn about the same type of problem-solving, their learning environments inescapably determine their differences. In the same controlled learning environment, their "IQ" decisions will NOT diverge...
Another factor is that in "normal" cases, the parent(s) demonstrate ways to handle conflicts and solve problems, and the child will imitate them, until (s)he doesn't. Is there going to be a switch or not, who knows? Still, inherited and acquired skills become inseparable.
Please, allow me to ask: Why is it important to differentiate between the two?
Thanks.
Your work is exemplary, always worth reading.
Apologies, if you believe my comment is too long. I can take it off and publish it in an article to which I would only leave a link here with a brief intro to the way it's related to your post.
Sorry, you didn't answer my question: Would you like me to publish my text separately or leave well enough alone?
No it's fine.
Thank you. I'll keep reading and popularizing your work! :)