One thing that has always bugged me about IQ is that it isn't clear to me exactly what it measures. I'm not talking the difference between intellectual and "emotional" IQ. I'm talking about how intelligence works in practical life vs intellectual life.
My late ex husband was at the top of his very arcane field, a small subset of Applied Mathematics. As a matter of practical life, he couldn't find his way out of a paper bag. He (luckily after our divorce) simply forgot to file his tax returns for two years. He didn't think to ask basic important questions of the vet when picking up the cat (eg. has it eaten or should we feed it?). He could never remember what day garbage day was. On and on and on.
I do very well on IQ and SAT type tests and performed well academically both undergrad, grad, and lawschool. But my spatial and math skills leave something to be desired in comparison with verbal use and reasoning. I could never perform STEM tasks with any degree of facility. My math scores on the SAT were good--well above 700 under the system in use back in 1962. But my verbal skills were very close to perfect.
In law school there was a student in my class who got 800 on the LSATs. But it was clear from his class participation that he wasn't really "getting" what the professors were saying. I don't remember where he placed in the class on graduation, but it was well below me. Law is, at root, more verbal than STEM.
What I have looked at when dealing with people, particularly less educated ones (with whom I dealt a lot in my legal career) is a quality I've called "brightness." That has to do with being quick to see connections between things that are relevant to their way in the world, as well as a curiosity about all things in their world and a willingness to go beyond their immediate world if that curiosity leads them there. I have NO IDEA whether this is measured by IQ. I can't see that it is measured well by the SAT and its ilk, which are only intended to predict academic success.
Often the high-school only educated contractors I encountered so often in my practice were WAY "brighter" than their lawyers. I had the same experience in grad school teaching "educationally disadvantaged" students who didn't have the grades OR test scores to get into the state University where I taught, but were selected because of their "promise." I have no idea how that promise was measured, but it really worked, because those classes had more students in the "bright" range than any other I taught, even where they didn't know how to put a sentence together. Those classes SANG with "brightness."
ARE there standardized tests aimed at "brightness" without worrying about academic achievement or the spatial/math parts of the tests?
Yup. The entire concept of a single trait that can be defined and measured as IQ is suspect at best. IQ tests measure the ability to do well on IQ tests.
Don, just a quick note: it is technically against the official guidelines of TIP to bring up politics unnecessarily, and so in the future try to avoid the "same mindset that gave us Trump" kind of thing (unless it directly bears on the subject). This goes for everyone else as well, so it's more of a general reminder, especially as other people already violated some comment guidelines in this thread (concerning direct insults) and were banned. You can read more on comment guidelines in the ABOUT page: https://www.theintrinsicperspective.com/about
Thanks. I will take a pass on the parapsychological: I'm totally agnostic on the topic but it's one of the fields of knowledge I don't feel curious enough about to pursue. But the rest of what you say makes sense.
And I will certainly add "non-linear" to my word-cloud around "bright." I was also going to add "flexible." And "open to picking up what it being said about fields other than their own" as with your PTSD-shy lawyers.
I like the "book smart" idea and how it is just one strain of intelligence. Important, of course, but not in isolation. I think what's that what I was getting at in my examples of many of the academics and lawyers I encountered.
(I've known a lot of academics. It took me 3 other degrees before I got my JD at 38, and I was a faculty wife for 10 more. By no means all of them were "book smart" only. I think the proportion is higher in the STEM and business school types, compared to the English faculty where I got to All But Dissertation before switching to law. )
The SATs only CLAIM to measure ability in college (as far as I remember). If in turn it is correlating with IQ, does that say something about IQ as a measure of a quality known as intelligence? Or is there something more? That is my question. Or more specifically, how is that "something more" measured? It it FOUND in the things that go into formal IQ, the types of questions asked?
Seems to me that the really deep measure of intelligence is a MIX of SAT/IQ measured stuff and "brightness." The ability to switch between them when need arises. That doesn't mean that everyone who mixes is a genius. But both QUALITIES are what measure how well one is able to navigate life, which is what smarts are generally for, after all.
A final note. One of my students in that "preferentially admitted" class met me on the street about 6 years later. I remembered him specifically because he was one of the ones who had not clue about how to write sentences or paragraphs, much less essays. And I watched him improve drastically. He told me he was now in his second year of grad school in architecture. Whoever saw "promise" in him in high school certainly knew what they were doing.
“But while different researchers give extremely wide estimates of the heritability of IQ, all the way from low single-digit percentages to upwards of half the variance,”
Nobody who is knowledgeable as far as I know puts IQ heritability at “low single-digits.” That link is to what one can explain from specific genes, that’s not an estimate of heredity, which needs to be derived from twin or adoption studies.
This gets into the whole "missing heritability" argument, in that twin studies show much higher heritability than modern genetic analyses do. But you can either think that the missing heritability is captured by twin studies and it's the genome-wide associations that are wrong, or you can think that twin studies are overestimating the effects. Personally, I think genomics is good enough at this point that if the heritability were *as strong* as twin studies initially showed, it would be higher. So while I personally doubt it's as low as single-digit percents, I also think twin studies are overestimating the upper bound. That still might mean it's pretty high though!
Yes, Alexander Young would place it higher, although I'll note his early research is about how twin studies overestimate heritability.
Regardless, according to the Introduction of the linked paper "PGS of intelligence is thought to predict 4-10.6% of the variance in intelligence (Davies et al., 2018; Hill et al., 2019; Lee et al., 2018)." I personally think that's an underestimation of heritability, but my point was only that there is "wide disagreement." It depends very strongly on where you draw the line between the lower bound of genetic studies and the upper bound of twin studies. Maybe rare variants will be found, but it would be a lot clearer if GWAS-type research was consistently creeping upwards in variance explanation, but that's not really what's been seen (so far).
But PGS are by definition only taking into accounts the genes we've identified as counting towards intelligence. They're (last I checked) never exhaustive in their nature.
So they would put a lower bound to heritability rather than a measure of it (however imperfect)
Establishing a lower bound is still a measure, since the actual heritability might be at that lower bound (e.g., if twin studies were massively overestimating). I think that's very unlikely, but it would then be the correct measure.
Correct me if I'm wrong, but GWAS has often been SNP focused and therefore neglects CNV's, insertion, deletion, inversion. Maybe this are better accounted for by whole genome sequencing, but could still neglect some other architectural aspects of genetics, such as epistasis, any epigenetic modifications, trans acting influences.
If such limitations were true, and that is the explanation for the missing heritability of intelligence, there would have to be a very surprising biological reason (at least, it would be surprising to me!) why you can get to something like 80% explanation of heritability of height through modern methods (although you need like ~100,000 SNPs to get there) but you get far lower, like 8-15%, explanations for intelligence.
Erik, this is pretty frustrating. I’ve published a GWAS in Science (last author) and many twin studies. You aren’t understanding these methods at all. The amount of variance explained by polygenic scores depends on the size/power of the discovery sample among many other things other than the heritability of the trait (hence the much greater variance accounted for by height polygenic scores). It is categorically NOT a measure of the heritability of a trait. There are no heritability estimates that suggest a heritability of IQ in the low single digits. SNP heritability estimates for intelligence are far higher, but even these are only based on heritability from SNPs, which are a very small percentage of the genome. They are not an estimate of total heritability. Heritability from rare variants is not captured. You said elsewhere you’re planning on writing an article on missing heritability. If so, please read up thoroughly before you do, it’s frustrating to read science writing so confident yet ill-informed.
(I would suggest you read Turkheimer's stuff very sceptically. He cites Plomin and von Stumm in his 10% SNP-heritability claim, but that paper itself says "For intelligence, SNP heritability is about 25%".)
Some of it may be attributable to height being a well defined, objective measurement whereas intelligence is much noisier of a construct, and as you mentioned much more variable and harder to measure. Additionally, seems like it could be a much more straightforward pathway of enactment (final common pathway of growth factors etc) as compared to complexity of brain and cognition. Complexity could also create more gene environment interactions, further contributing to noise. These are just conjectures tho!
OK, but what really gets our goat up is the 'isolated demand for rigour'. How many of the IQ sceptics are consistent and will say that, if IQ is uncertain, most of psychology is literally worthless? Hell, let's not even focus on middle-level stuff like big 5 personality traits. Most IQ critics believe in absolute trash like stereotype threat.
You're right that most people aren't consistent. Personally, I'm willing to even go further and say that most science is worthless, as the majority of scientific papers are never cited (or only self-cited)!
Although I do think the issues with what is attributed to the SMPY are of a different kind. It's not a problem of the statistical analysis or sample size, it's what claims it gets trotted out for.
Can appreciate this perspective, Mascilbinah: all research in psychology and other social sciences merits our skepticism and scrutiny, not just "those psychology takes any of us find distasteful." (And as you note, Erik, this also extends to a much wider sphere of scientific publishing.)
Maybe one takeaway is that human behavior is a notoriously difficult thing to study?
Yet I'm not entirely certain there's support for "most of psychology is literally worthless," without more clarity on what "most" means here. Challenging all of us to think about this, in turn: what findings in psychological research (categories of research, specific theories, or specific papers) do we yet believe have merit and promise, even if we're still early in our understanding?
And also this, given Erik's background: to what extent might physiologically-focused tools – including physical instrumentation, chemicals, and means of interpretation – particularly around (but not limited to) neurology, help us develop a more robust approach to psychology? (As a complement to making inferences around the causes and nature of human behavior mostly via observation by other humans, that is.)
This makes me sad because my own child has autism and an intellectual/learning disability, and I feel like she’s doomed, but I got a perfect score on my verbal SAT, so I often just sit around and wonder wtf happened and imagine how much better my life and hers would be if genetic IQ inheritance had freaking worked. 😢
You dig in deeply on one, possibly the best, applications of IQ and similar instruments that measure intelligence: identifying young people with great potential and supporting their education. I'm glad you drew the connection to “aristocratic tutoring.” Of course, there is the obverse effect of some kid being labeled below average, taking it to heart, and never rising to their potential.
The scientific racism that drove early research into IQ still animates some of the discourse. It would be great if Mismeasure of Man and other better histories had countered nefarious uses of measures of intelligence testing, but there are still people out there lying and making up data to support white supremacy and their racist fantasies. I get that it can be hard to distinguish between legitimate scientific questions and BS when some of the scientists involved are racist cranks, but I wish journalists and editors would make the effort instead of lumping it all together.
There are genuine disagreements about using factor analysis to identify entities that are not observable except through collecting data in tests and surveys. Specific examples of intelligence are easy to observe, but testing them in situations removed from environments where humans live and breath and actually use their cognitive abilities makes concepts like general intelligence or grit or whatever tricky to apply.
Your point that the "best" measure is not necessarily a "good" measure is really important. More effort needs to be made in explaining that such measures are "a noisy and weak signal" instead of marveling at the significance of a result arrived at by pounding a data set of test scores with factor analysis.
"Other[,] better histories" presumably meaning "better than Mismeasure of Man", right?
Interesting to note, though, that IQ research was opposed by all the nefarious groups it is now associated with. The Soviets didn't like it; the Nazis didn't like it; the Klan didn't like it; just about no one with an axe to grind has been a fan of the field, for some reason.
Perhaps that's why MoM & co. ended up not including the list you wish for: it's not actually very long. By and large, IQ testing has been benign.
(I often hear reference to "lying and making up data" by "crank" scientists, as you mention — but I've never actually *seen* it. Indeed, from what I've seen, there's no need to: most of the ingredients are fairly well-established. But I'm eager to see these fabricators and liars when and if available... I may be sadly out-of-date!)
Oh, I was in SMPY and never realized it! I entered through what I knew as the Duke (as in the university) Talent Identification Program. I took the SAT in 7th grade, did good, and spent the next 5 years getting course catalogs to go to enrichment programs we couldn’t afford. Maybe if I’d done better or been in a different location, there would have been more local or low cost options. As it is, I wonder about the feedback loop between kids who’d been able to get tutoring to know the geometry and Alg 2 on the SAT (I’d gotten neither), and the ability to pay for the interventions afterwards.
Believing that population-level traits are close to destiny does not require believing the same of individuals.
I don't think that the taller person will win every basketball 1v1, but if you split a classroom, city, or country in half based on height the team that could be fielded by the taller half will have increasingly greater chances to win, approaching 100%.
Most "IQ truthers" would say IQ is roughly as heritable as height, which seems like an apropriate bayesian prior. Maybe you should start lower since intelligence is likely more polygenetic. I don't think it makes sense to say they are as wrong as NYT columnists.
As far as I can tell, the genetics of IQ and height are pretty similar from a layman’s perspective. Obviously, environment can affect things downwards, but genes set potential — outside of things like stretching devices or training for an IQ test enough to raise a score. (Though my understanding is that for standardized testing there is good evidence that training/tutoring has limited impact on raising scores).
But height is much easier to measure and harder to deny, and has many fewer political controversies associated with it.
Leaving aside the genetics of it all, IQ and its proxies so reliably correlate with real-world outcomes that it’s unavoidable. You can nitpick any particular test or conception of G, but obviously there is something there, and inasmuch as intelligence has utility it should positively correlate with positive outcomes.
The blank slate position is so ridiculously wrong on so many issues that you’re going to get accused of bothsidesism for pointing out that the IQ science isn’t always perfect. There is massive social desirability bias against IQ such that people have been fired for even pointing out the “achievement gap,” which has been long recognized and used as a justification for policies to reduce it!
Recently, this research paper was withdrawn for being controversial because if you look at penises and IQ, you will also find racial correlations.
Just cursory stuff, comes from a weird field, authors names likely fake (There is a Cuntong Wang at the university of Illinois, but I couldn't find the co-author outside that particular paper), I don't think anyone today uses "Negroid" or "Caucasoid" for serious research work, they definitely don't use "Mongoloid," they have four overly broad racial categories despite the data ostensibly being multinational. I think it's a joke.
Edit: also, I doubt the specifications around penis measurement enumerated in the paper would be included if it were an actual study lol
1. The paper received a fair amount of criticism, which is why it was withdrawn.
2. The authors are Chinese and, in addition to being oblivious to Western Culture War issues, probably aren’t up to speed on politically correct terminology. (Pretty sure the University of Illinois scholar is not the coauthor here, given the email address listed on the paper.)
All this to say that if it were a troll, somebody would have figured that out by now. You were attentive enough to notice the possibility, as were many others.
Edit: actually it is the guy at the University of Illinois
I haven’t paid attention to any recent IQ discourse online, but I’m not surprised by any of this considering that it’s considered a “smart” thing to say dumb things like, “I like to run because it raises my endorphins!” Mostly by people who don’t have the slightest notion of biochemistry to begin with.
I never thought about IQ, I thought it was a dopey and outdated test even as a teenager, and I don’t remember any of my incredibly intelligent classmates at Harvey Mudd College ever really talking about it amidst our much more interesting and much weirder conversations, like hey Lucas, how did you get expelled from living on campus again because you decided to make thermite in your dorm room because you were interested in the chemistry of it?
I did know a bunch of people who went to CTY, and didn’t know that was the origin story, sounds about right. All this type of extreme inequality and aristocratic tutoring is really one of the biggest dark sides of America, speaking as someone who went on to work in education for many years in the country (going completely against the grain of the expected successful trajectory, speaking of such, of becoming a professor, working in aerospace, at Google, or a biotech company, like most of my classmates). The degree of excessive privilege and selective tutoring is indeed great in many communities, but as a teacher, on the whole, the most intelligent students I’ve usually encountered were not the ones locked in fierce tests and programs in a fight to the death, but the ones who were more blissfully unaware of all that and preferred to compose their own songs while learning abstract algebra on side, just for fun, with the support of loving and well intentioned (and usually well off) parents.
All in all, it’s a pretty strange dynamic, especially since I taught mostly minority students at one point who had mostly never even been to a major park 3 miles from their neighbourhood. That’s America. Now in Canada, you don’t tend to see the extremes much, and nowhere near the racial segregation inherent in them. The approaching and then going beyond aristocratic system in America seems pretty bizarre and blatantly unequal by comparison.
IQ discourse is not happening in a vacuum - it is largely a result of our current racial turmoil, and the debate over the cause of the racial gaps. IQ may not be the absolute most reliable measurement in the world, but the IQ gaps between groups are large and consistent enough that they hint at a genetic cause. Outside of this context, I don't think there would even be an IQ discourse.
Interesting essay! Unusually balanced for a discussion of IQ, and I appreciate that you pointed out that kids who are identified as having "high iQs" are singled out for more opportunities and training. I never took a formal IQ test or the SAT (I'm Canadian), but I got perfect scores on the math and spatial reasoning sections of my school district's "gifted" test and 800/800 on the quantitative section of the GRE. (I scored very high on the verbal too, but I'm definitely weaker there). I'm also weirdly good at Tetris. But in so many other ways I'm unexceptional. The kind of person who doesn't notice her dress is inside out and backwards for hours or notices halfway through the day I only shaved one of my legs. A little gullible too. I'm only fluent in one language and mispronounce words. A sociologist from the London School of Economics, Christopher Badcock, argues in one of his books that IQ tests only capture "mechanistic" cognitive abilities ... plus conformity to "Western" ways of thinking / Western culture, You might like his book, The Diametric Mind. He also argues that IQ tests give an inaccurate picture of intelligence for autistic people in particular as they capture autistic strengths but fail to capture cognitive weaknesses associated with the diagnostic label.
I did some research on the link between childhood giftedness (scoring high on an IQ test in childhood), high functioning autism (Aspergers), and rapid onset gender dysphoria (other mental health issues loosely addressed too). You might find it interesting, as I dig into some issues with childhood IQ testing in particular (e.g. precocity as a confounding factor).
>Christopher Badcock, argues in one of his books that IQ tests only capture "mechanistic" cognitive abilities ... plus conformity to "Western" ways of thinking / Western culture<
Unfortunately, that's nonsense.
Why? Take a moment and try to decide a method to investigate whether or not this might be true.
Suppose, for example, that East Asians score higher on IQ tests than Westerners.
That would make his thesis look sort of dumb, wouldn't it, if *literally the least-"Western" peoples on the planet* scored better than the most-Western?
But let's suppose obtaining the title "sociologist" *isn't* the academic equivalent of winning a Participation Trophy™, and one wouldn't make such an obvious and embarrassing misstep. Is there another way to see if there's just a big ol' cultural bias that somehow no one noticed?
Well, what if we gave IQ tests to a bunch of people from all around the world, and then compared which questions various groups tended to get wrong moreso than others?
I.e.: if there was a large cultural component to the tests, we would expect to see correlations between which questions are missed and what cultural background one has. These, then, would be the questions unfairly penalizing those with this background.
If, on the other hand, there *isn't* a correlation, and it appears random, we could say that the tests aren't just awarding higher scores to Western (say) takers but that, rather, IQ is a universally-applicable construct that's probably a pretty good proxy for g.
Before we look at the research on this, would you care to make a prediction? Will we see this pro-Western missed-question correlation?
(1) East Asia has significantly "Westernized" over the past century alongside industrialization. One could argue that the high performance on cognitive tests is indicative of conformity to "Western" ways of thinking, as well as the retention of various advantages from their own culture / gene pool.
(2) If you're looking at test scores given in the West, East Asian immigrants to the West are not, on average, representative of all East Asians. They are significantly more likely to be from elite families (presumably a bias toward higher intelligence), from urban centres, and significantly more likely to be drawn to Western ways of living and Western values (i.e. "Westernized"). You could also use the word "WEIRD" here (Western, Educated, Industrialized, Rich, Democratic), which is race-neutral; Joseph Henrich (The WEIRDest People in the World) points out that non-Westerners who are more "WEIRD"ly-inclined are more likely to immigrate to the West. Biased sample.
(3) East Asians eat significantly healthier diets than "white" people (on average, yes yes I know the Mediterranean diet is the healthiest in the world, I'm part Southern Italian). Diet (both maternal and the child's) makes a huge difference to brain development. An interesting observation is that East Asians (on average) consume less dairy than white people; consuming dairy products with iron-rich meat impairs iron absorption, and iron deficiencies are associated with poorer cognitive development. Jewish kosher rules prevent the consumption of dairy with red meat. Just one example, but generally speaking white people feed their kids more junk.
(4) It is possible for cognitive tests to both be biased toward a particular group (in this case Westerners / WEIRD populations, historically "white" people) *and* for that group to not be the highest performers on them once the tests and the ways of thinking involved in them go global.
(5) If by "least Western" you mean geographically, then okay, I guess we could hash this out with a map or something, but since I'm referring to cognitive configuration, culture, education, and economic factors, then it's silly to say East Asians (in particular urbanized ones) are the least Western when there are literal hunter-gatherer tribes still out there.
For an example of what I mean by "Western" ways of thinking, you can actually view one of Badcock's books online. The relevant pages are 87-135 and 280-283.
Smart people will typically do well on IQ tests; however, as I pointed out, many autistic people also do well despite having significant cognitive impairments because their talents are captured by the tests while their impairments are missed (I linked to an essay, you probably should have read it). They're probably the best we can do in terms of an objective, standardized assessment of "g"; I certainly have no idea how to design a multiple choice test that accurately captures, say, the creative, flow ability to improvise jazz music, social-emotional intelligence, or gut-level intuition. But IQ tests have significant limitations -- that was my point. They miss many cognitive abilities. They are better at identifying people who are cognitively impaired (low intelligence) versus average or above average. They are pretty useless at identifying and distinguishing between people of higher intelligence ... after all, it's quite unlikely that an intelligence test would be able to accurately assess the intelligence of someone who is smarter than the test designer ;-)
1.) None of this obviates the point that East Asians (or any other groups, to my knowledge) don't, as a group, miss different questions than e.g. Europeans, as a group. This is not what you would expect to see if indeed they were biased.
2.) Fair point re: Westernization — but not a killer one, I think. That is, if we consider Japan or China to be Western countries, the term is becoming meaningless; what, then, might we call "Eastern ways of thinking", if this isn't to be found in the Far East?
IMO — and I recognize this is informed by my politics, but I think it's almost certainly true regardless — the phrasing "they only test for Western intelligence" is chosen not because that's the clearest and most apt way to say it (modernized? urbanized? globalized? all better pointers to the element you speak of)... but because it's fashionable to critique "The West."
3.) The score difference is not due to a selection effects on immigrants, as far as I've ever seen — it persists despite attempts to find more representative samples, and AFAIK only a few real die-hards (in the field; the public is largely completely confused on the state of research on intelligence) still try to come up with ways the canonical "group ranking" could be an artifact.
4.) I think nutrition almost certainly can't be the cause, or we would see very different-looking data (e.g. one might expect, in that case, that low SES has a strong casual effect — but it doesn't).
I was going to say that I never said that IQ tests *only* test for Western intelligence, but upon review I can see how I could have been more careful with my words to prevent confusion:
Badcock argues that IQ tests *only* capture *mechanistic* cognitive abilities (as opposed to what he calls *mentalistic* cognitive abilities), with the exception that they also detect (to a lesser extent and not exclusively) conformity to "Western" (i.e. WEIRD, which I agree is the better term, but not the one he uses) ways of thinking. Again, read the essay I linked to, because my toddler just woke up and I don't have the time or desire to type out arguments I explained in significant detail, with sources, in a long-form essay.
SES and SAT scores are positively correlated ... but regardless, too many affluent white families feed their children crappy processed convenience foods and brain-rot like soy formula (and are putting cheese on their hamburgers and drinking mild with their steaks! lol) for this argument to negate my point about cultural differences in nutrition. My observation growing up in a city with a high Asian population is that lower income Asian families feed their kids healthier food *on average* than high-income white families.
I wouldn't expect East Asians to get different questions incorrect on average than Westerners. But, again, a test can *biased* toward a certain group and that group can still not be the highest performers. My husband quite cheekily put it, "being given a handicap in golf doesn't mean you win the game."
→ If you don't see differential performance by question, between two groups, then those questions are not biased in favor of either group.
That is, this *is* the mechanism by which a test is culturally biased: it has questions for which the skills involved appear to be unevenly distributed among the groups in question. No variation-by-group necessitates that variation-by-individual is all that remains.
I remain unconvinced that the tests detect conformity to WEIRD thinking in any meaningful way, unless 1) there are specific questions on the tests for which people from WEIRD nations do better, on average, as a group; and 2), unless Japan and China are just as WEIRD as Europe, and Japanese and Chinese people in the U.S. are no WEIRD-er than Japanese and Chinese people in their home countries.
It is true that bias can still not be enough to obviate another advantage, but conformity in scores and rank-ordering, and with East Asians at the top, would be a weird pattern to observe if the thesis were true. It isn't impossible — but there's no reason to believe it's true, either, and this is not what you'd *predict* you'd see.
(That said, if the entire modern, developed world counts as WEIRD by whatever metric we're using to say the tests detect WEIRDness... then sure, I'm fine with the tests being described this way!)
But perhaps the essay you linked will convince me.
Huh? Okay, say Group A is significantly more talented than Group B across 9/10 measures, with Measure 10 being the exception. Group B designs a cognitive aptitude test that includes 5 of these measures, including Measure 10. This would create a slight bias toward Group B, where their tests scores would indicate higher relative cognitive aptitude than if the test included all 10 measures. However, Group A might still outperform Group B on the test.
I am not saying this is *literally* how it works, I just made up an example. But it draws parallels to what I'm saying about autistic people scoring well on IQ tests. To quote Hans Asperger:
“[T]he Binet test, especially at older age levels, involves above all logical, abstract thinking. Since this is what autistic children often find so congenial, they may achieve a high score, which would give a false picture of their intelligence.”
Badcock (along with other writers) argues that the Western mind is slightly more "autistic" on average. There are good historic reasons for this. Iain McGilchrist's The Master and His Emissary is a better book related to this topic than Badcock's works, however.
I am a researcher of creative intelligence and I can tell you the other ways to study and cultivate intelligence. I went into this because my family is savants and we do not fit into the school system at all. We are destroyed by it. Generation after generation.
My children are all prodigies. I have a math and a science prodigy, an acting prodigy and my youngest is rolling her to prodigious giftedness now. Everything we think about intelligence is half wrong. You are missing the somatic intelligence and that eliminates the savants the ones who lean on their own. Our system's hyper focus on competition also eliminates the savants. The most gifted must create their own thing or go insane.
Right now we are driving the most gifted into insanity. I hate our systems and am glad they are failing. They destroy people just all other natural resources.. I will build new ones rather than try to change the old ones. There is nothing in them worth saving. They are set up to exploit instead of cultivate. Everyone working on them is too average and lack the sensitivity to really see and serve anyone that is a true brilliant outlier.
I think the telling problem is giving an LLM an SAT. When it scores near perfect, what is being measured or correlated? Certainly not genetics. Nor future potential. Neither cranial size, nor household income, certainly not ancestry. What is measured is in fact the entropy of the LLM model, the number of states it can take on, and move through to settle on a stable state relative to an input symbol string.
The lower the entropy, the worse the answers will be. It’s quite easily measurable.
Let’s speculate with that a tiny bit.
The correlate in humans will be close to the identification of neural density of interconnects, over the volume of the brain. It varies a lot over the animal kingdom - birds have much higher density brains than mammals, but are not so big. White matter in the brain has lower density from more sheathing.
Up to the point of synaptic pruning at puberty, these interconnects keep growing based on inputs - children in wealthy families have more words, images, and concepts shared by adults with children than poorer families (more occupied with survival) by count (recorders placed on infants and laboriously counted) leading to fewer interconnects in less environmentally rich contexts. Entropy is related to inputs. Children with high entropy singled out for even more inputs relative to a general population will get… more entropy. More variations of states. More “learning”. A child exposed to a blank sheet of paper in silence for 15 years will of course not be as smart as one immersed in a matrix of tutors on a world tour of music, science, literature and math for 15 years, as synaptic pruning kicks in.
Some genetics will make synaptic density higher or lower, or more rapid induction (autism filters sensory inputs less, psychosis more, autism spectrum children seem brighter…) paternal age at conception may induce mutations in neurotransmitters which increase brain entropy. Older parents may have high entropy children, a genetic trait but may look like it correlates to wealth due to older paternal age linking to more resources for children.
Entropy, “perplexity” (the inverse), how entropy is created, how it is genetically potentiated (neurotransmitters), how it is managed (synaptic pruning at puberty), how it alters over time (wisdom), and how it fails (Alzheimer’s, under-reinforcement over time), all related.
Testing for it with word problems is like rating heart function by measuring sprint times on a track.
I was a gifted student, until an accident at thirteen, which changed my cognitive life in ways I am now starting to understand at age 69. My fall from grace was profound and I never understood why, nor did “they” ever bother to explain.
This is a very bad take on the SMPY, and on the heritability of IQ in general. Your main thesis is that 11-13 year olds were tested. And those that scored very high were selected for extra tutoring and coaching. That's why those kids grow up to be very successful.
So the extra tutoring worked on SMPY kids. But somehow extra tutoring doesn't work on the millions of other kids who were not part of the SMPY? Do you realize how many billions of dollars have been spent on programs like Head Start and it doesn't work? The most recent example is Mark Zuckerberg blowing $100M on Newark schools and it did NOTHING.
Spending money on extra program to help kids who are not doing well do not work. Pupil spending do not correlate with test score! You really should spent more time reading the literature on this. At least read some articles from Freddie deBoer on this. He has written extensively on why "education doesn't work". And it's probably genes.
Is there anything in Head Start about individualized tutoring? Isn't it just preschool? And did Zuckerberg's donation to Newark schools really get used well, or was it all spent on improving the facilities and buying the kids ipads? I can easily see burning a 100 million and reducing the student : teacher ratio of a big school system by only a tiny fraction.
I've read Freddie's writings on education before. I think it's just good evidence that we are incredibly ineffective at education at a mass scale. But because within the US switching from public school A to public school B has few effects doesn't mean "education doesn't work."
Starting from the 1960s, the US has spent billions of dollars trying to improve educational outcomes. Other countries like Australia have tried this also with their aboriginal population. Almost everything you can think of was tried. The most invasive programs involved forced adoptions in Australia! Nothing has worked long term. There are big "fade out" effects that researches try to hide. Here are just two programs for you to chew on:
Your bringing up "individualized tutoring" and "ipads" shows that you don't know the literature. The fact that money doesn't change educational outcomes is a "HATE FACT" for liberals. This conflicts with your foundational belief that you can improve intelligence through education. The evidence from behavior genetics shows that is simply not true. If you weren't beaten or starved as a child (as it is the case for most children in Western countries), you will reach your genetic potential which was determined at conception.
MRI scan and machine learning technology will continue to improve. We can already measure cranial volume very accurately using MRIs. And cranial volume correlates with intelligence. In the future, MRI scans will be able count the billions of neurons cells in the brain and their connections individually.
How will you cope when there's visual evidence of intelligence in MRI scans? How will you cope when we find out that the number of neurons differs greatly in different populations? How will you cope when we can map the regions in the brain that causes poor impulsive control and leads to greater propensity to violence?
Will you try to ban this type of research because you don't like the conclusions?
I think this is a really great example of the massively overconfident pop-IQ worldview I'm discussing here. It is just the inverse of the New Yorker editor I poked fun of. Just as it's absurd for her to say "genes have zero influence on intelligence" so is it absurd to say "If you weren't beaten or starved as a child you will reach your genetic potential which was determined at conception." That's not true, both at a scientific level, and at a personal level (you probably don't act out this belief in your personal life, just like the New Yorker editor doesn't act her, equally untrue, belief). The first things you reach to are the same sort of older now-myth studies like the SMPY. Like the MTAS, which consisted of a mere 130 individuals, all in Minnesota, and happened in the 1960s (even now Minnesota is overwhelming white, and has one of the largest racial success gaps in the nation, if I'm not mistaken). You're taking poorly-controlled circumstantial studies with low numbers and tons of potential co-founders and drawing really strong causal conclusions when there are a ton of alternative hypotheses. This same sort of selective credulity extends to neuroscience, where there is no consensus explanation for IQ (I know this because I have a PhD in neuroscience). If anything, the central finding of neuroscience for the last 30 years is that the brain is extremely extremely plastic and that measuring things like size of cortical areas is not very relevant (see, for instance, how when people become experts at something the neuroimaging markers *shrink*). You're not talking about population genetics or neuroscience as it exists, you're talking about how it's portrayed on the fringes at places like X, blogs, and Reddit.
One thing that has always bugged me about IQ is that it isn't clear to me exactly what it measures. I'm not talking the difference between intellectual and "emotional" IQ. I'm talking about how intelligence works in practical life vs intellectual life.
My late ex husband was at the top of his very arcane field, a small subset of Applied Mathematics. As a matter of practical life, he couldn't find his way out of a paper bag. He (luckily after our divorce) simply forgot to file his tax returns for two years. He didn't think to ask basic important questions of the vet when picking up the cat (eg. has it eaten or should we feed it?). He could never remember what day garbage day was. On and on and on.
I do very well on IQ and SAT type tests and performed well academically both undergrad, grad, and lawschool. But my spatial and math skills leave something to be desired in comparison with verbal use and reasoning. I could never perform STEM tasks with any degree of facility. My math scores on the SAT were good--well above 700 under the system in use back in 1962. But my verbal skills were very close to perfect.
In law school there was a student in my class who got 800 on the LSATs. But it was clear from his class participation that he wasn't really "getting" what the professors were saying. I don't remember where he placed in the class on graduation, but it was well below me. Law is, at root, more verbal than STEM.
What I have looked at when dealing with people, particularly less educated ones (with whom I dealt a lot in my legal career) is a quality I've called "brightness." That has to do with being quick to see connections between things that are relevant to their way in the world, as well as a curiosity about all things in their world and a willingness to go beyond their immediate world if that curiosity leads them there. I have NO IDEA whether this is measured by IQ. I can't see that it is measured well by the SAT and its ilk, which are only intended to predict academic success.
Often the high-school only educated contractors I encountered so often in my practice were WAY "brighter" than their lawyers. I had the same experience in grad school teaching "educationally disadvantaged" students who didn't have the grades OR test scores to get into the state University where I taught, but were selected because of their "promise." I have no idea how that promise was measured, but it really worked, because those classes had more students in the "bright" range than any other I taught, even where they didn't know how to put a sentence together. Those classes SANG with "brightness."
ARE there standardized tests aimed at "brightness" without worrying about academic achievement or the spatial/math parts of the tests?
Yup. The entire concept of a single trait that can be defined and measured as IQ is suspect at best. IQ tests measure the ability to do well on IQ tests.
Huh? I am asking questions. It bugs me that it isn’t clear—can anyone help? Here’s what I’ve noticed. Is there any research.
No, you say, you need to know intelligence research first. Before I guess I am allowed to notice things and wonder. Bravo you for defending the castle
Don, just a quick note: it is technically against the official guidelines of TIP to bring up politics unnecessarily, and so in the future try to avoid the "same mindset that gave us Trump" kind of thing (unless it directly bears on the subject). This goes for everyone else as well, so it's more of a general reminder, especially as other people already violated some comment guidelines in this thread (concerning direct insults) and were banned. You can read more on comment guidelines in the ABOUT page: https://www.theintrinsicperspective.com/about
Thanks. I will take a pass on the parapsychological: I'm totally agnostic on the topic but it's one of the fields of knowledge I don't feel curious enough about to pursue. But the rest of what you say makes sense.
And I will certainly add "non-linear" to my word-cloud around "bright." I was also going to add "flexible." And "open to picking up what it being said about fields other than their own" as with your PTSD-shy lawyers.
I like the "book smart" idea and how it is just one strain of intelligence. Important, of course, but not in isolation. I think what's that what I was getting at in my examples of many of the academics and lawyers I encountered.
(I've known a lot of academics. It took me 3 other degrees before I got my JD at 38, and I was a faculty wife for 10 more. By no means all of them were "book smart" only. I think the proportion is higher in the STEM and business school types, compared to the English faculty where I got to All But Dissertation before switching to law. )
The SATs only CLAIM to measure ability in college (as far as I remember). If in turn it is correlating with IQ, does that say something about IQ as a measure of a quality known as intelligence? Or is there something more? That is my question. Or more specifically, how is that "something more" measured? It it FOUND in the things that go into formal IQ, the types of questions asked?
Seems to me that the really deep measure of intelligence is a MIX of SAT/IQ measured stuff and "brightness." The ability to switch between them when need arises. That doesn't mean that everyone who mixes is a genius. But both QUALITIES are what measure how well one is able to navigate life, which is what smarts are generally for, after all.
A final note. One of my students in that "preferentially admitted" class met me on the street about 6 years later. I remembered him specifically because he was one of the ones who had not clue about how to write sentences or paragraphs, much less essays. And I watched him improve drastically. He told me he was now in his second year of grad school in architecture. Whoever saw "promise" in him in high school certainly knew what they were doing.
“But while different researchers give extremely wide estimates of the heritability of IQ, all the way from low single-digit percentages to upwards of half the variance,”
Nobody who is knowledgeable as far as I know puts IQ heritability at “low single-digits.” That link is to what one can explain from specific genes, that’s not an estimate of heredity, which needs to be derived from twin or adoption studies.
This gets into the whole "missing heritability" argument, in that twin studies show much higher heritability than modern genetic analyses do. But you can either think that the missing heritability is captured by twin studies and it's the genome-wide associations that are wrong, or you can think that twin studies are overestimating the effects. Personally, I think genomics is good enough at this point that if the heritability were *as strong* as twin studies initially showed, it would be higher. So while I personally doubt it's as low as single-digit percents, I also think twin studies are overestimating the upper bound. That still might mean it's pretty high though!
I talked to Alexander Young about this, and i don’t think anyone in this debate maintains low single digits is a good estimate.
https://www.cspicenter.com/p/getting-at-true-heritability-alexander
Yes, Alexander Young would place it higher, although I'll note his early research is about how twin studies overestimate heritability.
Regardless, according to the Introduction of the linked paper "PGS of intelligence is thought to predict 4-10.6% of the variance in intelligence (Davies et al., 2018; Hill et al., 2019; Lee et al., 2018)." I personally think that's an underestimation of heritability, but my point was only that there is "wide disagreement." It depends very strongly on where you draw the line between the lower bound of genetic studies and the upper bound of twin studies. Maybe rare variants will be found, but it would be a lot clearer if GWAS-type research was consistently creeping upwards in variance explanation, but that's not really what's been seen (so far).
For interest, can you point me at a summary of arguments/evidence that twin studies might be significantly overestimating heritability?
This paper I have in my notes is recent and has a lot on the different explanations from a philosophy of biology perspective (there's a section on twin overestimation you can source for citations). https://www.sciencedirect.com/science/article/abs/pii/S0039368122000632
And I personally will likely write something on the missing heritability problem at some point in the future, although I don't know when.
Did you ever write this up?
But PGS are by definition only taking into accounts the genes we've identified as counting towards intelligence. They're (last I checked) never exhaustive in their nature.
So they would put a lower bound to heritability rather than a measure of it (however imperfect)
Establishing a lower bound is still a measure, since the actual heritability might be at that lower bound (e.g., if twin studies were massively overestimating). I think that's very unlikely, but it would then be the correct measure.
Correct me if I'm wrong, but GWAS has often been SNP focused and therefore neglects CNV's, insertion, deletion, inversion. Maybe this are better accounted for by whole genome sequencing, but could still neglect some other architectural aspects of genetics, such as epistasis, any epigenetic modifications, trans acting influences.
If such limitations were true, and that is the explanation for the missing heritability of intelligence, there would have to be a very surprising biological reason (at least, it would be surprising to me!) why you can get to something like 80% explanation of heritability of height through modern methods (although you need like ~100,000 SNPs to get there) but you get far lower, like 8-15%, explanations for intelligence.
Erik, this is pretty frustrating. I’ve published a GWAS in Science (last author) and many twin studies. You aren’t understanding these methods at all. The amount of variance explained by polygenic scores depends on the size/power of the discovery sample among many other things other than the heritability of the trait (hence the much greater variance accounted for by height polygenic scores). It is categorically NOT a measure of the heritability of a trait. There are no heritability estimates that suggest a heritability of IQ in the low single digits. SNP heritability estimates for intelligence are far higher, but even these are only based on heritability from SNPs, which are a very small percentage of the genome. They are not an estimate of total heritability. Heritability from rare variants is not captured. You said elsewhere you’re planning on writing an article on missing heritability. If so, please read up thoroughly before you do, it’s frustrating to read science writing so confident yet ill-informed.
(I would suggest you read Turkheimer's stuff very sceptically. He cites Plomin and von Stumm in his 10% SNP-heritability claim, but that paper itself says "For intelligence, SNP heritability is about 25%".)
Some of it may be attributable to height being a well defined, objective measurement whereas intelligence is much noisier of a construct, and as you mentioned much more variable and harder to measure. Additionally, seems like it could be a much more straightforward pathway of enactment (final common pathway of growth factors etc) as compared to complexity of brain and cognition. Complexity could also create more gene environment interactions, further contributing to noise. These are just conjectures tho!
OK, but what really gets our goat up is the 'isolated demand for rigour'. How many of the IQ sceptics are consistent and will say that, if IQ is uncertain, most of psychology is literally worthless? Hell, let's not even focus on middle-level stuff like big 5 personality traits. Most IQ critics believe in absolute trash like stereotype threat.
You're right that most people aren't consistent. Personally, I'm willing to even go further and say that most science is worthless, as the majority of scientific papers are never cited (or only self-cited)!
Although I do think the issues with what is attributed to the SMPY are of a different kind. It's not a problem of the statistical analysis or sample size, it's what claims it gets trotted out for.
Can appreciate this perspective, Mascilbinah: all research in psychology and other social sciences merits our skepticism and scrutiny, not just "those psychology takes any of us find distasteful." (And as you note, Erik, this also extends to a much wider sphere of scientific publishing.)
Maybe one takeaway is that human behavior is a notoriously difficult thing to study?
Yet I'm not entirely certain there's support for "most of psychology is literally worthless," without more clarity on what "most" means here. Challenging all of us to think about this, in turn: what findings in psychological research (categories of research, specific theories, or specific papers) do we yet believe have merit and promise, even if we're still early in our understanding?
And also this, given Erik's background: to what extent might physiologically-focused tools – including physical instrumentation, chemicals, and means of interpretation – particularly around (but not limited to) neurology, help us develop a more robust approach to psychology? (As a complement to making inferences around the causes and nature of human behavior mostly via observation by other humans, that is.)
This makes me sad because my own child has autism and an intellectual/learning disability, and I feel like she’s doomed, but I got a perfect score on my verbal SAT, so I often just sit around and wonder wtf happened and imagine how much better my life and hers would be if genetic IQ inheritance had freaking worked. 😢
Thank god for that SAT/income correlation though. Thanking the universe for that one.
You dig in deeply on one, possibly the best, applications of IQ and similar instruments that measure intelligence: identifying young people with great potential and supporting their education. I'm glad you drew the connection to “aristocratic tutoring.” Of course, there is the obverse effect of some kid being labeled below average, taking it to heart, and never rising to their potential.
The scientific racism that drove early research into IQ still animates some of the discourse. It would be great if Mismeasure of Man and other better histories had countered nefarious uses of measures of intelligence testing, but there are still people out there lying and making up data to support white supremacy and their racist fantasies. I get that it can be hard to distinguish between legitimate scientific questions and BS when some of the scientists involved are racist cranks, but I wish journalists and editors would make the effort instead of lumping it all together.
There are genuine disagreements about using factor analysis to identify entities that are not observable except through collecting data in tests and surveys. Specific examples of intelligence are easy to observe, but testing them in situations removed from environments where humans live and breath and actually use their cognitive abilities makes concepts like general intelligence or grit or whatever tricky to apply.
Your point that the "best" measure is not necessarily a "good" measure is really important. More effort needs to be made in explaining that such measures are "a noisy and weak signal" instead of marveling at the significance of a result arrived at by pounding a data set of test scores with factor analysis.
"Other[,] better histories" presumably meaning "better than Mismeasure of Man", right?
Interesting to note, though, that IQ research was opposed by all the nefarious groups it is now associated with. The Soviets didn't like it; the Nazis didn't like it; the Klan didn't like it; just about no one with an axe to grind has been a fan of the field, for some reason.
Perhaps that's why MoM & co. ended up not including the list you wish for: it's not actually very long. By and large, IQ testing has been benign.
(I often hear reference to "lying and making up data" by "crank" scientists, as you mention — but I've never actually *seen* it. Indeed, from what I've seen, there's no need to: most of the ingredients are fairly well-established. But I'm eager to see these fabricators and liars when and if available... I may be sadly out-of-date!)
Oh, I was in SMPY and never realized it! I entered through what I knew as the Duke (as in the university) Talent Identification Program. I took the SAT in 7th grade, did good, and spent the next 5 years getting course catalogs to go to enrichment programs we couldn’t afford. Maybe if I’d done better or been in a different location, there would have been more local or low cost options. As it is, I wonder about the feedback loop between kids who’d been able to get tutoring to know the geometry and Alg 2 on the SAT (I’d gotten neither), and the ability to pay for the interventions afterwards.
Believing that population-level traits are close to destiny does not require believing the same of individuals.
I don't think that the taller person will win every basketball 1v1, but if you split a classroom, city, or country in half based on height the team that could be fielded by the taller half will have increasingly greater chances to win, approaching 100%.
Most "IQ truthers" would say IQ is roughly as heritable as height, which seems like an apropriate bayesian prior. Maybe you should start lower since intelligence is likely more polygenetic. I don't think it makes sense to say they are as wrong as NYT columnists.
As far as I can tell, the genetics of IQ and height are pretty similar from a layman’s perspective. Obviously, environment can affect things downwards, but genes set potential — outside of things like stretching devices or training for an IQ test enough to raise a score. (Though my understanding is that for standardized testing there is good evidence that training/tutoring has limited impact on raising scores).
But height is much easier to measure and harder to deny, and has many fewer political controversies associated with it.
Leaving aside the genetics of it all, IQ and its proxies so reliably correlate with real-world outcomes that it’s unavoidable. You can nitpick any particular test or conception of G, but obviously there is something there, and inasmuch as intelligence has utility it should positively correlate with positive outcomes.
The blank slate position is so ridiculously wrong on so many issues that you’re going to get accused of bothsidesism for pointing out that the IQ science isn’t always perfect. There is massive social desirability bias against IQ such that people have been fired for even pointing out the “achievement gap,” which has been long recognized and used as a justification for policies to reduce it!
Recently, this research paper was withdrawn for being controversial because if you look at penises and IQ, you will also find racial correlations.
https://www.researchsquare.com/article/rs-3816362/v1
Reality does not have a progressive bias.
You bought into a troll paper I think.
I didn’t buy into anything.
Do you have any evidence it was a troll paper?
Just cursory stuff, comes from a weird field, authors names likely fake (There is a Cuntong Wang at the university of Illinois, but I couldn't find the co-author outside that particular paper), I don't think anyone today uses "Negroid" or "Caucasoid" for serious research work, they definitely don't use "Mongoloid," they have four overly broad racial categories despite the data ostensibly being multinational. I think it's a joke.
Edit: also, I doubt the specifications around penis measurement enumerated in the paper would be included if it were an actual study lol
Consider:
1. The paper received a fair amount of criticism, which is why it was withdrawn.
2. The authors are Chinese and, in addition to being oblivious to Western Culture War issues, probably aren’t up to speed on politically correct terminology. (Pretty sure the University of Illinois scholar is not the coauthor here, given the email address listed on the paper.)
All this to say that if it were a troll, somebody would have figured that out by now. You were attentive enough to notice the possibility, as were many others.
Edit: actually it is the guy at the University of Illinois
https://www.researchgate.net/profile/Cuntong-Wang
Surprised spending time in the US didn’t sensitize him to issues he should steer clear of.
I haven’t paid attention to any recent IQ discourse online, but I’m not surprised by any of this considering that it’s considered a “smart” thing to say dumb things like, “I like to run because it raises my endorphins!” Mostly by people who don’t have the slightest notion of biochemistry to begin with.
I never thought about IQ, I thought it was a dopey and outdated test even as a teenager, and I don’t remember any of my incredibly intelligent classmates at Harvey Mudd College ever really talking about it amidst our much more interesting and much weirder conversations, like hey Lucas, how did you get expelled from living on campus again because you decided to make thermite in your dorm room because you were interested in the chemistry of it?
I did know a bunch of people who went to CTY, and didn’t know that was the origin story, sounds about right. All this type of extreme inequality and aristocratic tutoring is really one of the biggest dark sides of America, speaking as someone who went on to work in education for many years in the country (going completely against the grain of the expected successful trajectory, speaking of such, of becoming a professor, working in aerospace, at Google, or a biotech company, like most of my classmates). The degree of excessive privilege and selective tutoring is indeed great in many communities, but as a teacher, on the whole, the most intelligent students I’ve usually encountered were not the ones locked in fierce tests and programs in a fight to the death, but the ones who were more blissfully unaware of all that and preferred to compose their own songs while learning abstract algebra on side, just for fun, with the support of loving and well intentioned (and usually well off) parents.
All in all, it’s a pretty strange dynamic, especially since I taught mostly minority students at one point who had mostly never even been to a major park 3 miles from their neighbourhood. That’s America. Now in Canada, you don’t tend to see the extremes much, and nowhere near the racial segregation inherent in them. The approaching and then going beyond aristocratic system in America seems pretty bizarre and blatantly unequal by comparison.
IQ discourse is not happening in a vacuum - it is largely a result of our current racial turmoil, and the debate over the cause of the racial gaps. IQ may not be the absolute most reliable measurement in the world, but the IQ gaps between groups are large and consistent enough that they hint at a genetic cause. Outside of this context, I don't think there would even be an IQ discourse.
Interesting essay! Unusually balanced for a discussion of IQ, and I appreciate that you pointed out that kids who are identified as having "high iQs" are singled out for more opportunities and training. I never took a formal IQ test or the SAT (I'm Canadian), but I got perfect scores on the math and spatial reasoning sections of my school district's "gifted" test and 800/800 on the quantitative section of the GRE. (I scored very high on the verbal too, but I'm definitely weaker there). I'm also weirdly good at Tetris. But in so many other ways I'm unexceptional. The kind of person who doesn't notice her dress is inside out and backwards for hours or notices halfway through the day I only shaved one of my legs. A little gullible too. I'm only fluent in one language and mispronounce words. A sociologist from the London School of Economics, Christopher Badcock, argues in one of his books that IQ tests only capture "mechanistic" cognitive abilities ... plus conformity to "Western" ways of thinking / Western culture, You might like his book, The Diametric Mind. He also argues that IQ tests give an inaccurate picture of intelligence for autistic people in particular as they capture autistic strengths but fail to capture cognitive weaknesses associated with the diagnostic label.
I did some research on the link between childhood giftedness (scoring high on an IQ test in childhood), high functioning autism (Aspergers), and rapid onset gender dysphoria (other mental health issues loosely addressed too). You might find it interesting, as I dig into some issues with childhood IQ testing in particular (e.g. precocity as a confounding factor).
Link: https://thecassandracomplex.substack.com/p/the-drama-of-the-gifted-children
Glad this popped up on my feed!
>Christopher Badcock, argues in one of his books that IQ tests only capture "mechanistic" cognitive abilities ... plus conformity to "Western" ways of thinking / Western culture<
Unfortunately, that's nonsense.
Why? Take a moment and try to decide a method to investigate whether or not this might be true.
Suppose, for example, that East Asians score higher on IQ tests than Westerners.
That would make his thesis look sort of dumb, wouldn't it, if *literally the least-"Western" peoples on the planet* scored better than the most-Western?
But let's suppose obtaining the title "sociologist" *isn't* the academic equivalent of winning a Participation Trophy™, and one wouldn't make such an obvious and embarrassing misstep. Is there another way to see if there's just a big ol' cultural bias that somehow no one noticed?
Well, what if we gave IQ tests to a bunch of people from all around the world, and then compared which questions various groups tended to get wrong moreso than others?
I.e.: if there was a large cultural component to the tests, we would expect to see correlations between which questions are missed and what cultural background one has. These, then, would be the questions unfairly penalizing those with this background.
If, on the other hand, there *isn't* a correlation, and it appears random, we could say that the tests aren't just awarding higher scores to Western (say) takers but that, rather, IQ is a universally-applicable construct that's probably a pretty good proxy for g.
Before we look at the research on this, would you care to make a prediction? Will we see this pro-Western missed-question correlation?
A few things:
(1) East Asia has significantly "Westernized" over the past century alongside industrialization. One could argue that the high performance on cognitive tests is indicative of conformity to "Western" ways of thinking, as well as the retention of various advantages from their own culture / gene pool.
(2) If you're looking at test scores given in the West, East Asian immigrants to the West are not, on average, representative of all East Asians. They are significantly more likely to be from elite families (presumably a bias toward higher intelligence), from urban centres, and significantly more likely to be drawn to Western ways of living and Western values (i.e. "Westernized"). You could also use the word "WEIRD" here (Western, Educated, Industrialized, Rich, Democratic), which is race-neutral; Joseph Henrich (The WEIRDest People in the World) points out that non-Westerners who are more "WEIRD"ly-inclined are more likely to immigrate to the West. Biased sample.
(3) East Asians eat significantly healthier diets than "white" people (on average, yes yes I know the Mediterranean diet is the healthiest in the world, I'm part Southern Italian). Diet (both maternal and the child's) makes a huge difference to brain development. An interesting observation is that East Asians (on average) consume less dairy than white people; consuming dairy products with iron-rich meat impairs iron absorption, and iron deficiencies are associated with poorer cognitive development. Jewish kosher rules prevent the consumption of dairy with red meat. Just one example, but generally speaking white people feed their kids more junk.
(4) It is possible for cognitive tests to both be biased toward a particular group (in this case Westerners / WEIRD populations, historically "white" people) *and* for that group to not be the highest performers on them once the tests and the ways of thinking involved in them go global.
(5) If by "least Western" you mean geographically, then okay, I guess we could hash this out with a map or something, but since I'm referring to cognitive configuration, culture, education, and economic factors, then it's silly to say East Asians (in particular urbanized ones) are the least Western when there are literal hunter-gatherer tribes still out there.
For an example of what I mean by "Western" ways of thinking, you can actually view one of Badcock's books online. The relevant pages are 87-135 and 280-283.
Link to download: https://www.researchgate.net/publication/333816494_The_Diametric_Mind_sample
Smart people will typically do well on IQ tests; however, as I pointed out, many autistic people also do well despite having significant cognitive impairments because their talents are captured by the tests while their impairments are missed (I linked to an essay, you probably should have read it). They're probably the best we can do in terms of an objective, standardized assessment of "g"; I certainly have no idea how to design a multiple choice test that accurately captures, say, the creative, flow ability to improvise jazz music, social-emotional intelligence, or gut-level intuition. But IQ tests have significant limitations -- that was my point. They miss many cognitive abilities. They are better at identifying people who are cognitively impaired (low intelligence) versus average or above average. They are pretty useless at identifying and distinguishing between people of higher intelligence ... after all, it's quite unlikely that an intelligence test would be able to accurately assess the intelligence of someone who is smarter than the test designer ;-)
1.) None of this obviates the point that East Asians (or any other groups, to my knowledge) don't, as a group, miss different questions than e.g. Europeans, as a group. This is not what you would expect to see if indeed they were biased.
2.) Fair point re: Westernization — but not a killer one, I think. That is, if we consider Japan or China to be Western countries, the term is becoming meaningless; what, then, might we call "Eastern ways of thinking", if this isn't to be found in the Far East?
IMO — and I recognize this is informed by my politics, but I think it's almost certainly true regardless — the phrasing "they only test for Western intelligence" is chosen not because that's the clearest and most apt way to say it (modernized? urbanized? globalized? all better pointers to the element you speak of)... but because it's fashionable to critique "The West."
3.) The score difference is not due to a selection effects on immigrants, as far as I've ever seen — it persists despite attempts to find more representative samples, and AFAIK only a few real die-hards (in the field; the public is largely completely confused on the state of research on intelligence) still try to come up with ways the canonical "group ranking" could be an artifact.
4.) I think nutrition almost certainly can't be the cause, or we would see very different-looking data (e.g. one might expect, in that case, that low SES has a strong casual effect — but it doesn't).
I was going to say that I never said that IQ tests *only* test for Western intelligence, but upon review I can see how I could have been more careful with my words to prevent confusion:
Badcock argues that IQ tests *only* capture *mechanistic* cognitive abilities (as opposed to what he calls *mentalistic* cognitive abilities), with the exception that they also detect (to a lesser extent and not exclusively) conformity to "Western" (i.e. WEIRD, which I agree is the better term, but not the one he uses) ways of thinking. Again, read the essay I linked to, because my toddler just woke up and I don't have the time or desire to type out arguments I explained in significant detail, with sources, in a long-form essay.
SES and SAT scores are positively correlated ... but regardless, too many affluent white families feed their children crappy processed convenience foods and brain-rot like soy formula (and are putting cheese on their hamburgers and drinking mild with their steaks! lol) for this argument to negate my point about cultural differences in nutrition. My observation growing up in a city with a high Asian population is that lower income Asian families feed their kids healthier food *on average* than high-income white families.
I wouldn't expect East Asians to get different questions incorrect on average than Westerners. But, again, a test can *biased* toward a certain group and that group can still not be the highest performers. My husband quite cheekily put it, "being given a handicap in golf doesn't mean you win the game."
I'll read the link, sure, but one point:
→ If you don't see differential performance by question, between two groups, then those questions are not biased in favor of either group.
That is, this *is* the mechanism by which a test is culturally biased: it has questions for which the skills involved appear to be unevenly distributed among the groups in question. No variation-by-group necessitates that variation-by-individual is all that remains.
I remain unconvinced that the tests detect conformity to WEIRD thinking in any meaningful way, unless 1) there are specific questions on the tests for which people from WEIRD nations do better, on average, as a group; and 2), unless Japan and China are just as WEIRD as Europe, and Japanese and Chinese people in the U.S. are no WEIRD-er than Japanese and Chinese people in their home countries.
It is true that bias can still not be enough to obviate another advantage, but conformity in scores and rank-ordering, and with East Asians at the top, would be a weird pattern to observe if the thesis were true. It isn't impossible — but there's no reason to believe it's true, either, and this is not what you'd *predict* you'd see.
(That said, if the entire modern, developed world counts as WEIRD by whatever metric we're using to say the tests detect WEIRDness... then sure, I'm fine with the tests being described this way!)
But perhaps the essay you linked will convince me.
Huh? Okay, say Group A is significantly more talented than Group B across 9/10 measures, with Measure 10 being the exception. Group B designs a cognitive aptitude test that includes 5 of these measures, including Measure 10. This would create a slight bias toward Group B, where their tests scores would indicate higher relative cognitive aptitude than if the test included all 10 measures. However, Group A might still outperform Group B on the test.
I am not saying this is *literally* how it works, I just made up an example. But it draws parallels to what I'm saying about autistic people scoring well on IQ tests. To quote Hans Asperger:
“[T]he Binet test, especially at older age levels, involves above all logical, abstract thinking. Since this is what autistic children often find so congenial, they may achieve a high score, which would give a false picture of their intelligence.”
Badcock (along with other writers) argues that the Western mind is slightly more "autistic" on average. There are good historic reasons for this. Iain McGilchrist's The Master and His Emissary is a better book related to this topic than Badcock's works, however.
I am a researcher of creative intelligence and I can tell you the other ways to study and cultivate intelligence. I went into this because my family is savants and we do not fit into the school system at all. We are destroyed by it. Generation after generation.
My children are all prodigies. I have a math and a science prodigy, an acting prodigy and my youngest is rolling her to prodigious giftedness now. Everything we think about intelligence is half wrong. You are missing the somatic intelligence and that eliminates the savants the ones who lean on their own. Our system's hyper focus on competition also eliminates the savants. The most gifted must create their own thing or go insane.
Right now we are driving the most gifted into insanity. I hate our systems and am glad they are failing. They destroy people just all other natural resources.. I will build new ones rather than try to change the old ones. There is nothing in them worth saving. They are set up to exploit instead of cultivate. Everyone working on them is too average and lack the sensitivity to really see and serve anyone that is a true brilliant outlier.
I think the telling problem is giving an LLM an SAT. When it scores near perfect, what is being measured or correlated? Certainly not genetics. Nor future potential. Neither cranial size, nor household income, certainly not ancestry. What is measured is in fact the entropy of the LLM model, the number of states it can take on, and move through to settle on a stable state relative to an input symbol string.
The lower the entropy, the worse the answers will be. It’s quite easily measurable.
Let’s speculate with that a tiny bit.
The correlate in humans will be close to the identification of neural density of interconnects, over the volume of the brain. It varies a lot over the animal kingdom - birds have much higher density brains than mammals, but are not so big. White matter in the brain has lower density from more sheathing.
Up to the point of synaptic pruning at puberty, these interconnects keep growing based on inputs - children in wealthy families have more words, images, and concepts shared by adults with children than poorer families (more occupied with survival) by count (recorders placed on infants and laboriously counted) leading to fewer interconnects in less environmentally rich contexts. Entropy is related to inputs. Children with high entropy singled out for even more inputs relative to a general population will get… more entropy. More variations of states. More “learning”. A child exposed to a blank sheet of paper in silence for 15 years will of course not be as smart as one immersed in a matrix of tutors on a world tour of music, science, literature and math for 15 years, as synaptic pruning kicks in.
Some genetics will make synaptic density higher or lower, or more rapid induction (autism filters sensory inputs less, psychosis more, autism spectrum children seem brighter…) paternal age at conception may induce mutations in neurotransmitters which increase brain entropy. Older parents may have high entropy children, a genetic trait but may look like it correlates to wealth due to older paternal age linking to more resources for children.
Entropy, “perplexity” (the inverse), how entropy is created, how it is genetically potentiated (neurotransmitters), how it is managed (synaptic pruning at puberty), how it alters over time (wisdom), and how it fails (Alzheimer’s, under-reinforcement over time), all related.
Testing for it with word problems is like rating heart function by measuring sprint times on a track.
Super interesting article, thanks! But why the “even” on Lady Gaga?
Yes! Lol I wondered the same thing.
I was a gifted student, until an accident at thirteen, which changed my cognitive life in ways I am now starting to understand at age 69. My fall from grace was profound and I never understood why, nor did “they” ever bother to explain.
This is a very bad take on the SMPY, and on the heritability of IQ in general. Your main thesis is that 11-13 year olds were tested. And those that scored very high were selected for extra tutoring and coaching. That's why those kids grow up to be very successful.
So the extra tutoring worked on SMPY kids. But somehow extra tutoring doesn't work on the millions of other kids who were not part of the SMPY? Do you realize how many billions of dollars have been spent on programs like Head Start and it doesn't work? The most recent example is Mark Zuckerberg blowing $100M on Newark schools and it did NOTHING.
Spending money on extra program to help kids who are not doing well do not work. Pupil spending do not correlate with test score! You really should spent more time reading the literature on this. At least read some articles from Freddie deBoer on this. He has written extensively on why "education doesn't work". And it's probably genes.
https://freddiedeboer.substack.com/p/education-doesnt-work-20
https://freddiedeboer.substack.com/p/is-the-conventional-wisdom-on-educational
https://freddiedeboer.substack.com/p/education-commentary-is-dominated
Is there anything in Head Start about individualized tutoring? Isn't it just preschool? And did Zuckerberg's donation to Newark schools really get used well, or was it all spent on improving the facilities and buying the kids ipads? I can easily see burning a 100 million and reducing the student : teacher ratio of a big school system by only a tiny fraction.
I've read Freddie's writings on education before. I think it's just good evidence that we are incredibly ineffective at education at a mass scale. But because within the US switching from public school A to public school B has few effects doesn't mean "education doesn't work."
Starting from the 1960s, the US has spent billions of dollars trying to improve educational outcomes. Other countries like Australia have tried this also with their aboriginal population. Almost everything you can think of was tried. The most invasive programs involved forced adoptions in Australia! Nothing has worked long term. There are big "fade out" effects that researches try to hide. Here are just two programs for you to chew on:
https://en.wikipedia.org/wiki/Minnesota_Transracial_Adoption_Study
https://en.wikipedia.org/wiki/Stolen_Generations
Your bringing up "individualized tutoring" and "ipads" shows that you don't know the literature. The fact that money doesn't change educational outcomes is a "HATE FACT" for liberals. This conflicts with your foundational belief that you can improve intelligence through education. The evidence from behavior genetics shows that is simply not true. If you weren't beaten or starved as a child (as it is the case for most children in Western countries), you will reach your genetic potential which was determined at conception.
MRI scan and machine learning technology will continue to improve. We can already measure cranial volume very accurately using MRIs. And cranial volume correlates with intelligence. In the future, MRI scans will be able count the billions of neurons cells in the brain and their connections individually.
How will you cope when there's visual evidence of intelligence in MRI scans? How will you cope when we find out that the number of neurons differs greatly in different populations? How will you cope when we can map the regions in the brain that causes poor impulsive control and leads to greater propensity to violence?
Will you try to ban this type of research because you don't like the conclusions?
I think this is a really great example of the massively overconfident pop-IQ worldview I'm discussing here. It is just the inverse of the New Yorker editor I poked fun of. Just as it's absurd for her to say "genes have zero influence on intelligence" so is it absurd to say "If you weren't beaten or starved as a child you will reach your genetic potential which was determined at conception." That's not true, both at a scientific level, and at a personal level (you probably don't act out this belief in your personal life, just like the New Yorker editor doesn't act her, equally untrue, belief). The first things you reach to are the same sort of older now-myth studies like the SMPY. Like the MTAS, which consisted of a mere 130 individuals, all in Minnesota, and happened in the 1960s (even now Minnesota is overwhelming white, and has one of the largest racial success gaps in the nation, if I'm not mistaken). You're taking poorly-controlled circumstantial studies with low numbers and tons of potential co-founders and drawing really strong causal conclusions when there are a ton of alternative hypotheses. This same sort of selective credulity extends to neuroscience, where there is no consensus explanation for IQ (I know this because I have a PhD in neuroscience). If anything, the central finding of neuroscience for the last 30 years is that the brain is extremely extremely plastic and that measuring things like size of cortical areas is not very relevant (see, for instance, how when people become experts at something the neuroimaging markers *shrink*). You're not talking about population genetics or neuroscience as it exists, you're talking about how it's portrayed on the fringes at places like X, blogs, and Reddit.
An excellent example of how the reverse psychology effect leads to being wrong along every dimension.