Highlights From The Comments On Lynn And IQ
...
Shaked Koplewitz writes:
Doesn't Lynn's IQ measure also suffer from the IQ/g discrepancy that causes the Flynn effect?
That is, my understanding of the Flynn effect is that IQ doesn't exactly measure g (the true general intelligence factor) but measures some proxy that is somewhat improved by literacy/education, and for most of the 20th century those were getting better leading to improvements in apparent IQ (but not g). Shouldn't we expect sub Saharan Africans to have lower IQ relative to g (since their education and literacy systems are often terrible)?
And then the part about them seeming much smarter than a first worlder with similar IQ makes sense - they'd do equally badly at tests, but in their case it's because e.g. they barely had a chance to learn to read rather than not being smart enough to think of the answer.
(Or a slightly more complicated version of this - e.g. maybe they can read fine, but never had an education that encouraged them to consider counterfactuals so those just don't come naturally).
Yeah, this is the most important factor that I failed to cover in the post (I edited it in ten minutes later after commenters reminded me, but some of you got the email and didn’t see it).
For the second effect mentioned in the post - the one where Malawians are obviously smarter than intellectually disabled people - you could attribute it to any of:
Lynn’s data and analysis were bad.
Lynn’s data and analysis were fine as far as they go, but the tests he based his work off of were trivially culturally biased - for example, they asked about English vocabulary in non-English speaking countries, or math problems to people who had never learned math.
The tests weren’t trivially biased, but the concept of IQ itself breaks down once you try to extend it to extremely under-educated populations, and it no longer predicts things as well as you would expect.
The concept of IQ is fine, but you are personally miscalibrated about what low IQ means because the only very-low-IQ people in your training set had developmental disorders.
I think these probably explain 5%, 5%, 40%, and 50% of the effect respectively, and I should have been more careful to emphasize (3), which I think explains 40% of the effect.
The particular way I would flesh out 3 would be something like - if you’re illiterate and (somewhat) innumerate, you probably don’t have enough practice with symbols and complex mental operations to do even a “culture fair” IQ test like Raven’s Matrices. This doesn’t necessarily mean that your IQ is higher than the Raven’s Matrices says - the person who underperforms on Ravens for this reason will also underperform on a wide variety of other abstract/intellectual/symbolic tasks, and this is part of what IQ means. But it means that Raven’s IQ won’t predict concrete tasks as well as you would expect.
Fujimura writes:
The other major factor that I think should be reassuring about Lynn's estimates (and other cross-national IQ estimates) is that when you look at "non-problematic" sources that seem like proxies for IQ (e.g. World Bank data, educational performance), you see the same pattern as Lynn and others' IQ data.
It's easy for people to quibble about each and every IQ measure (and so people do), but that we see the same pattern of results using otherwise uncontroversial data sources should be reassuring.
Yeah, many people tried to gotcha me with claims that Lynn did this or that or the other thing wrong. Lynn tries to defend his methodology here, but I think (and tried to argue in the post) that at this point, that debate is of historical interest only - there’s too much confirmation now. One commenter brings up World Bank Harmonized Learning Outcomes as an example. Another points me to this preprint, which tries to update Lynn’s numbers using all modern standardized testing data and correlations with social development index and GDP. They find mostly similar numbers to Lynn: Malawi goes from 60 → 66, and new last place goes to Sao Tome & Principe at 62. This is by people affiliated with Lynn and scientific racism, and you can choose not to trust their judgment either, but I think at least the SDI correlations are an extremely simple regression that it would be hard to fake. This kind of stuff is why I think simple failures of data collection and analysis are unlikely to explain more than 5% of the gap with our common sense. There’s definitely something weird about these numbers, but it’s got to be more complicated than just “racist people screwed up the test”.
But continuing on this subject - if IQ has two components, why would World Bank education data and GDP track the abstract/symbolic component of IQ, rather than the practical component of IQ? Or, rather, it’s obvious why this would happen in education. But why would GDP track abstract/symbolic rather than practical?
One possible answer is that the causal pathway is high GDP → lots of education → lots of practice with abstract reasoning → high abstract/symbolic IQ. I don’t think this can be the whole story, because some countries that “cheated” to get high GDP (eg oil sheikhdoms) can’t translate it into IQ points at the same rate as everyone else. I’m stuck with the boring basic explanation that maybe you need to do a lot of abstract reasoning tasks to get high GDP.
Harzerkatze writes:
[Your claim that blacks everywhere should have the same genes] is far from true. While "white" may be a descriptor for a group of somewhat similar genetic backgrounds, having common ancestors not too far in the past, "black" is different, grouping populations of similar skin color, but common ancestors diverging way further back in time.
Yeah, I didn’t want to get into all of this on the post, but I agree the way I phrased it was misleading.
Lynn and other national IQ estimates find very low IQs for all sub-Saharan African countries - I mentioned Malawi at 60 in the post, but Nigeria, on the other side of Africa, is 69. Whatever is going on there is a pan-African problem, such that I don’t think differences between African groups are very relevant.
US blacks are mostly descended from people in west Africa, eg Nigeria. Some people also brought up that US blacks have significant white admixture. This is true but it’s still not enough to be relevant to this discussion. If we assumed everything was genetic and US blacks with their ~20% white admixture had genetic IQ of 85, we would still expect African blacks to have IQ in the low 80s. However you parse it, there’s got to be some kind of health/education/environment effect going on there.
Africa is extremely genetically diverse, but I think most of the countries measured in the paper, including Malawi, are some variety of Niger-Congo speakers, who I don’t think are that much more diverse than white people or anyone else. The really interesting African ethnicities, like the Khoi-San, don’t show up as much at a national level.
Andrew Clough writes:
Speaking of charity and IQ, the lowest of low hanging fruit is putting iodine in salt. You can donate to the Global Iodine Network like I do for the long term benefit of poorer countries without worrying you're just delaying Malthus's reemergence. Givewell calls Salt Iodization "slightly below the range of cost-effectiveness of the opportunities that we expect to direct marginal donations to" which in the grand scheme of things is quite good.
Yeah, salt iodization is great. I had always heard of iodine related problems being concentrated in central Asia and especially Afghanistan, but looking at the map…
… sub-Saharan Africa is also a hot spot.
I wonder what’s wrong in Cuba - this is exactly the sort of easily gameable metric I would usually expect them to be good at, or at least carefully faking.
If you’re interested, you can donate to Iodine Global Network here.
Bob Jacobs writes:
> His opponents pointed out both his personal racist opinions/activities
That's the mildest possible way you could've put it. He wasn't someone who had "personal racist opinions" that he kept as "personal racist opinions". He was the editor-in-chief of Mankind Quarterly, a white supremacist journal that was founded by people like:
Henry Garrett an American psychologist who testified in favor of segregated schools during Brown versus Board of Education, Corrado Gini who was president of the Italian genetics and eugenics Society in fascist Italy, and Otmar Freiherr von Verschuer who was director of the Kaiser Wilhelm Institute of anthropology human heredity and eugenics in Nazi Germany. He was a member of the Nazi Party and the mentor of Josef Mengele, the physician at the Auschwitz concentration camp infamous for performing human experimentation on the prisoners during World War 2. Mengele provided for Verschuer with human remains from Auschwitz to use in his research into eugenics.
It's funded by the pioneer fund, an organization he was a board member of and that has been classified as a white supremacist hate group, with one of its first projects being to fund the distribution in US churches and schools of "Erbkrank", a Nazi propaganda film about eugenics.
He's not just called racist, he *is* racist, he even describes *himself* as a racist.
No contesting any of this.
MM writes:
I spent 18 months in a country where people are supposed to have an iq of about 70, according to the map. My neighbors and friends were mostly non-literate. They did not seem less intelligent than the people I know in my current (US) neighborhood or the people I grew up with (in the US). Most of them would not have performed well on IQ tests, though. They'd never attended school and had no familiarity with puzzle-solving. This was 35 years ago and most people had not seen movies or even photographs. I remember sitting with one older woman and helping her interpret a black-and-white photograph: this is the arm, here's where it connects to the body, etc. It's hard for people from literate societies with tons of exposure to text & graphical representations to see the extent of the gap.
Calvin writes:
I have a decent amount of experience with the intellectually disabled, and saying "cognitive issues are only responsible for a small part of the [communication] deficit" is so wrong that it makes me question everything else in this essay. Trust me, even making allowances for poor hearing or difficulty forming words, the cognitive issues are responsible for 90% of the deficit. An IQ of 60 is really low and it's a significant handicap.
I was concerned to hear this - I have a little experience with the intellectually disabled, but it didn’t involve knowing people’s exact IQ, so I’m not very well-calibrated here.
Looking for more information, I found https://www.hrw.org/reports/2001/ustat/ustat0301-01.htm, which purports to describe the characteristics of very low IQ people, mostly in the context of criminal justice (where lawyers often try to use a client’s low IQ as a mitigating factor - ie maybe he didn’t truly understand that crime is wrong). The report says things like:
Although all persons with mental retardation have significantly impaired mental development, their intellectual level can vary considerably. An estimated 89 percent of all people with retardation have I.Q.s in the 51-70 range. An I.Q. in the 60 to 70 range is approximately the scholastic equivalent to the third grade […]
Although mental retardation of any degree has profound implications for a person's cognitive and social development, it is a condition which in many cases is not readily apparent. While some of the mentally retarded, such as those whose retardation is caused by Down's syndrome or fetal alcohol syndrome, have characteristically distinctive facial features, most cannot be identified by their physical appearance alone. Unless their cognitive impairment is unusually severe (e.g. an I.Q. below 40), persons with mental retardation may be thought of as "slow" but the full extent of their impairment is often not readily appreciated, particularly by people who have limited contact with or knowledge of them, including police, prosecutors, judges, and other participants in the criminal justice system. Many capital offenders with mental retardation did not have their condition diagnosed until trial or during post-conviction proceedings.
And gave some examples (slightly out of order for this list):
Oliver Cruz, who was executed in Texas on August 9, 2000, had an I.Q. that was measured variously at 64 and 76. Cruz nonetheless insisted to reporters that, although he was perhaps "slow in reading, slow in learning," he was not mentally retarded.
Mitigation specialist Scharlette Holdman recalled a client who so successfully hid his retardation from his attorneys that he allowed them to sign him up for college-level calculus classes, which he could not comprehend. He had gone through much of his schooling allowing his younger sister to complete his homework for him. When he was given papers to read in connection to his case, he would carefully stare at them. If he was asked a substantive question, he usually responded, "I don't recall." Only when experts in retardation evaluated him and investigators reviewed his school records and spoke to his family did lawyers discover he had mental retardation and had been considered "slow" since his early childhood.
Another capital defendant "hid his mental retardation for most of his life by working at a very repetitive job as a switcher on the railroad. He lied about finishing high school. He was actually in special education classes and did not finish the sixth grade. He was drafted into the army and discharged because of his mental retardation. He lied about his service record. He often made things up so that people would not suspect mental retardation."
Morris Mason, whose I.Q. was 62-66, was executed in 1985 in Virginia after being convicted of rape and murder. Before his execution, Mason asked one of his legal advisors for advice on what to wear to his funeral
As one psychiatrist testified about a capital defendant with an I.Q. of between 35 to 45: "[People with mental retardation try] to go along with people that they suspect are in authority. For example, I asked [the defendant] where we were when I saw him, and he obviously didn't know, so I asked him if we were in Atlanta and he said `Yes, we are in Atlanta.' In fact, we were in Birmingham, Alabama. I could have said New York and he would have said `Sure, New York'
These people are obviously not going to win Nobels anytime soon. But even the guy with IQ 35 - 45 was still talking to people. I think this supports the thesis that intellectually disabled people without specific syndromes can seem pretty normal most of the time.
(though keep in mind that anything from the court system should be treated with a grain of salt - defense attorneys have an incentive to exaggerate the intellectual disability of their clients in the hopes that it gets them a lighter sentence)
Lyman Stone writes:
Emil's post isn't correct, however.
We know from the recent Reich lab paper on long-run genetic selection that there was strong selection for IQ in the neolithic revolution, which implies agriculture strongly selects for IQ and ability to plan.
Malawians are 60-80% subsistence farmers.
Even a "normal" low-IQ person cannot do the implied math and long-term planning involved in this kind of farming. And in fact, economists routinely find that African small-plot subsistence agriculture is actually highly optimized; farmers make very precise choices about where to plant which seeds, which fertilizer to use, etc. Key point is basically: it really isn't true that an IQ 60 person can run a farm functionally.
Moreover, mean IQ of 60 implies large shares even lower, at ranges that are uniformly nonverbal even without specific disability. And this is why in the actual record-level NIQ database, they truncate estimates below 60, because even the database managers realize these estimates are crazy.
See my post here: https://substack.com/home/post/p-154757665
We know that people with extremely low IQs in the Flynn sense must be capable of subsistence agriculture, because pre-Flynn Effect, most of the West had extremely low IQs, and they were all doing subsistence agriculture.
How is this possible? Responding to Lyman’s comment, I wrote:
I stick to the claim in this post - that our estimates for what a very low IQ means are poorly-grounded, and that people with low IQs can do some pretty impressive things, especially if they're concrete and part of a cultural transmission package. Maybe this is the Joseph Henrich "Secret Of Our Success" thing. We know that Malawians get poor test scores in school, so it seems like there's some disconnect between do-well-on-tests intelligence and run-a-subsistence-farm intelligence, and the abstract/concrete and novel/cultural distinctions are the best explanation that I can think of.
You say that "the phenotype that arises from a given tested IQ in America is clearly vastly worse than the phenotype arising from the same tested IQ in Africa", which I basically agree with. I think part of it is the syndromes issue raised above, and part of it is that maybe Malawians have zero contact with the culture of abstraction that IQ tests come out of whereas even very uneducated Westerners have some contact with it, and maybe another part of it is that whatever health/nutrition issues the Malawians have preferentially harm faculties responsible for more abstract tasks rather than more concrete ones.
For an opposite data point, when I was in Haiti, my boss told me (secondhand, no personal experience) of extreme difficulties working with Haitians, like that they couldn't alphabetize files even when that was explained to them. Many Haitains are also successfuly subsistence farmers, so I think this also supports some kind of heavy abstract/concrete distinction.
I don't think we're really disagreeing, just agreeing on something like the correlations that make up IQ being less valid outside the normal range.
Maybe one way to look at it is to go back to the claim from the justice system document above, saying that people with IQ in the 60s are the mental equivalent of third-graders. The third-graders I know are very into Pokemon, and have all sorts of opinions on how if you add X bonus to a Y strength fire-type Pokemon and then play Z combo, it will [commence six weeks of droning on about different Pokemon cards]. Is this the sort of math/reasoning/strategizing that we don’t expect someone with IQ 60 to be able to do? Does the fact that third-graders can do it mean that we’re miscalibrated? I’m not sure.
The part of Lyman’s comment that gives me the most pause is his observation that, if the mean IQ is 60, a decent fraction of people must be 45, and a non-negligible portion 30. At this point, even third-grader comparisons don’t save us. I guess this is where I bring in the claim that IQ breaks down as a guide to practical living skills below some point.
You can see several more layers of response between me and Lyman here, but I was especially grateful for him teaching me two things I didn’t already know:
First, he corrected my misconception about Reich on ancient European cognitive evolution. Reich had said that pre-agriculture Europeans were “2-3 standard deviations” below moderns. I had interpreted that as IQ deviations of 15 points, making them genetic IQ 55-70, which would have been pretty crazy. Stone tells me he actually meant PGS deviations, each of which was about 3-4 IQ points, so he’s claiming that pre-agriculture Europeans had genetic IQ of 90 (they probably also had lower IQ for environmental reasons).,
Second, he linked a post of his where he found that, although IQ accurately predicts GDP at each time point, changes in IQ don’t predict changes in GDP, suggesting something weird is happening. I think the weird thing is the improvement in the abstract/symbolic/”test-taking” aspect of IQ separate from the practical aspect, mentioned above.
Final Comments/Conclusion:
I should have been more careful to highlight the possibility that the subcomponents of IQ come apart at the tails and that this explains some of the effect. I didn’t do that partly because I thought it was closely related to the developmental syndromes point (which is a specific example of how things can come apart at the tails), partly because I expected people to misinterpret it as “IQ is meaningless and nothing is correlated with anything”, and partly because I don’t understand the correlations of different subcomponents with each other and g and practical skills to have a good opinion here. Still, I could have at least said “Maybe academic and practical skills come apart”.
Maybe I should have had a stronger opinion on whether Lynn’s exact studies were correct? Certainly lots of commenters had strong opinions that they weren’t. I had hoped that linking the Aporia article would be a sufficient pointer to my opinion that, while Lynn’s work was a first effort and far from perfect, the general thrust (including surprisingly low IQs in sub-Saharan African countries) has been confirmed by later research which is harder to bias. Maybe it would have been less controversial if I had just asked why sub-Saharan Africa has surprisingly low school test scores, but then nobody would have a good sense of what school test scores were normal vs. surprising and nobody would have cared.
I’m interested in learning more about Lyman Stone’s claim that changes in IQ don’t predict changes in development level. This could mean that development causes IQ (through education), but that sometimes you can get IQ through other means (like an extreme pro-education push by a poor country), and this doesn’t itself cause development. Or it could mean that countries have a real (genetic?) IQ and a fake (educational?) IQ, and the GDP only depends on the real IQ (but then how come Africa, which everyone agrees is underperforming its genetics, is in the expected place on the IQ/GDP correlation?) Or it could mean that the changes are fake, and changes in fake data don’t affect real GDP (but really? No country has ever really changed its IQ? Doesn’t that contradict our suspicion from the African data that something about education must be involved?)
Share this post