The Unz Review: An Alternative Media Selection
A Collection of Interesting, Important, and Controversial Perspectives Largely Excluded from the American Mainstream Media
 BlogviewJames Thompson Archive
Swanning About: Fooled by Algebra?
🔊 Listen RSS
Email This Page to Someone

 Remember My Information



=>

Bookmark Toggle AllToCAdd to LibraryRemove from Library • BShow CommentNext New CommentNext New ReplyRead More
ReplyAgree/Disagree/Etc. More... This Commenter This Thread Hide Thread Display All Comments
AgreeDisagreeLOLTroll
These buttons register your public Agreement, Disagreement, Troll, or LOL with the selected comment. They are ONLY available to recent, frequent commenters who have saved their Name+Email using the 'Remember My Information' checkbox, and may also ONLY be used three times during any eight hour period.
Ignore Commenter Follow Commenter
Search Text Case Sensitive  Exact Words  Include Comments
List of Bookmarks

black swan google

Nassim Nicholas Taleb has tweeted a set of remarks about intelligence research.

https://threadreaderapp.com/thread/1076845397795065856.html

He has now gathered those together into one format, with links and explanations.

https://medium.com/incerto/iq-is-largely-a-pseudoscientific-swindle-f131c101ba39

There is no lack of confidence in his essay. There is much to discuss here, and what follows covers what I see as the main points. I have added some links to relevant publications, but you can put any of the concepts and author names in my search bar to get further details.

1 IQ is largely a pseudoscientific swindle

Given that Taleb criticizes the poor statistics used by intelligence researchers, a mild comment is that it would have been better to be more precise. I have assumed he means that more than half of intelligence research findings are wrong, and for malicious reasons. If this is his point, he is factually wrong.

2 IQ is stale, mostly measures very low intelligence, or a lesser form of intelligence for paper shufflers or those ill-suited to real life. “it explains at best between 13% and 50% of the performance in some tasks”. It is based on poor maths, and promoted by racists and swindlers.

It seems that Taleb has a poor opinion of many people. “Paper shufflers” probably include all the backroom workers who keep the books and process the transactions of star traders. The reason for his doubts about the maths behind IQ, Taleb explains, is that he can computer-generate correlations based on particular assumptions which then look like some of the reported findings on intelligence and scholastic attainment. He implies that if he can do that on a simple basis (create a mythical test which only measures below IQ 100 performance and progressively just add noise above that) then that invalidates the actual observations reported by Frey and Detterman (2004). This is not a compelling argument. A far simpler explanation is that a population wide measure (general IQ) is being compared with a scholastic test taken only by a selection of brighter students (SAT) yet still does a pretty good job of showing the link between the two. This is a real-life finding, of the sort that Taleb supposedly favours.

3 If you want to detect how someone fares at a task, say loan sharking, tennis playing, or random matrix theory, make him/her do that task; we don’t need theoretical exams for a real-world function by probability-challenged psychologists.

In fact, psychologists have understood this point. Hunter and Schmidt and Kunzel point out that the best test of whether a person can do a job is to let them try it. However, this is expensive in time and money, since you have to supervise them to prevent disasters, give them detailed instructions and monitor their performance carefully, all of which takes at least two weeks to get a reasonable estimate of the applicant’s capabilities. You cannot do this for all applicants, or it would take up all the staff time required for doing the actual work of the business. The above researchers show that an intelligence test is a close second-best in terms of outcome, and far quicker and cheaper. Add a test of honesty and you have an efficient selection system.

4 Different populations have different variances, even different skewness and these comparisons require richer models.

Again, most psychometricians agree with that and it has been known for decades. At the very least, they like seeing the data plotted out properly, so the actual findings are visible, and so that they can be analyzed by different statistical approaches. Nothing new or insightful here.

5 A measure that works in left tail not right tail (IQ decorrelates as it goes higher) is problematic.

Lubinski and Benbow have shown in prospective studies with a large sample that IQ is still predictive at the very highest levels, and keeps working at each higher band. Taleb’s point is demonstrably wrong.

6 It (IQ) can measure some arbitrarily selected mental abilities (in a testing environment) believed to be useful. However, if you take a Popperian-Hayekian view on intelligence, you would realize that to measure it you would need to know the mental skills needed in a future ecology, which requires predictability of said future ecology. It also requires the skills to make it to the future (hence the need for mental biases for survival).

Intelligence test items are not arbitrary. They are selected to represent a wide range of abilities drawn from actual tasks and real-life problems. They correlate highly with tests which specifically base themselves on real life tasks in American society, such as the Wonderlic Personnel Test. Linda Gottfredson has shown all this, many times, for decades. As to “mental skills needed in a future ecology”, that is an excellent example of intelligent behaviour, as is survival. In a Scottish population study, Ian Deary has shown that intelligence tested at age 11 predicted lifespan into old age. Brighter people were capable of surviving longer than the less bright. Taleb is wrong again.

7 Real life never never offers crisp questions with crisp answers (most questions don’t have answers; perhaps the worst problem with IQ is that it seems to selects for people who don’t like to say “there is no answer, don’t waste time, find something else”.)

If this were a relevant objection, then the crisp answers required in the Scottish 11+ would not have shown any relation to lifespan and decades of achievement. Equally, the crisp answers required of SMPY participants would not have predicted their mid-life achievements (and will probably predict decades of achievement as the follow-ups continue). Digits backwards is a crisp-answer task. It wastes little time, yet is a good predictor of general ability. Crisp test answers also correlate to many brain structure and function measures assessed by neuroimaging (Haier, 2017). Also, given that all puzzles require brain power, these selected items may tap a general ability to solve puzzles of a far more general and urgent nature.

8 It takes a certain type of person to waste intelligent concentration on classroom/academic problems. These are lifeless bureaucrats who can muster sterile motivation. Some people can only focus on problems that are real, not fictional textbook ones.

Taleb is very free with his insults. It might play to those already taking an anti-IQ stance. A rough measure of ability can be obtained in two minutes, which does not tax concentration. Sure, many people favour the practical over the academic, and might concentrate best on real-life problems. This is testable, and once again, on a broad range of people and a broad range of real-life problems, intelligence tests maintain predictive utility. Detterman shows many of the correlations.

9 IQ doesn’t detect convexity (by an argument similar to bias-variance you need to make a lot of small inconsequential mistakes in order to avoid a large consequential one. See Antifragile and how any measure of “intelligence” w/o convexity is sterile edge.org/conversation/n…)

Taleb makes interesting points about what he has described as “convexity”. In his Edge essay he points out that “chance” is not a good explanation for long term gains. Now we have something we can agree upon. By “convexity” Taleb means that in his view research progresses by “a significant asymmetry between the gains (as they need to be large) and the errors (small or harmless), and it is from such asymmetry that luck and trial and error can produce results”. This is a convex function, hence the name he gives it. Fine. This may or may not be the case, and it is not clear how this hypothesis could be tested directly (and therefore not scientific), nor is it clear why this proposal means that a measure of intelligence would be “sterile”. One test might be to see what the correlation is between IQ and options trading/financial investments. The latter are real world tests of considerable significance, and a null result would strengthen his argument.

As luck would have it, here is a relevant publication.

https://sci-hub.shop/10.1016/j.jfineco.2011.05.016

We analyze whether IQ influences trading behavior, performance, and transaction costs. The analysis combines equity return, trade, and limit order book data with two decades of scores from an intelligence (IQ) test administered to nearly every Finnish male of draft age. Controlling for a variety of factors, we find that high-IQ investors are less subject to the disposition effect, more aggressive about tax-loss trading, and more likely to supply liquidity when stocks experience a one-month high. High-IQ investors also exhibit superior market timing, stock-picking skill, and trade execution.

The authors find that by making better stock selections and achieving lower transaction costs high IQ subjects do 4.9% per year better than low IQ subjects. Given that real returns average 7%, this is a massive difference which will accumulate over time and result in far high net personal worth for brighter investors. By the way, intelligence is measured at conscription age, long before there is much investment history, so is more likely to be causal.

https://www.sciencedirect.com/science/article/pii/S0304405X1100211X

10 Seeing shallow patterns is not a virtue — leads to naive interventionism. Some psychologist wrote back to me: “IQ selects for pattern recognition, essential for functioning in modern society”. No. Not seeing patterns except when they are significant is a virtue in real life. To do well in life you need depth and ability to select your own problems and to think independently.

The ability to see patterns where others cannot has traditionally been seen as a sign of intelligence. Interestingly, Taleb accepts that some problems are shallow. How does he know? Presumably he can see through them, and finds them easy. Good. Item difficulties vary. Depth, ability to select problems and to think independently are signs of intelligence. This is a point of agreement with psychometrics although he may not realize it.

11 Functionary Quotient: If you renamed IQ, from “Intelligent Quotient” to FQ “Functionary Quotient” or SQ “Salaryperson Quotient”, then some of the stuff will be true. It measures best the ability to be a good slave.

Taleb’s argument seems to be that IQ tests only work for common folk in humdrum jobs. Not exactly oblesse oblige. This is a version of the old familiar argument that intelligence tests do not measure creativity. Easy to assert, but the evidence seems to be against it, so long as you measure creativity by quality and quantity. Rex Jung has studied this matter, creatively and carefully.

Kunzel shows that college entrance tests can predict success even in jobs that are far from humdrum.

12 “IQ” is most predictive of performance in military training, w/correlation~.5, (which is circular since hiring isn’t random).

Two criticisms. First, the military training data is interesting in itself, but the key issue in terms of the generality of the findings is that quite a few tasks were identical to non-military tasks. For example, vehicle maintenance is the same task, so what we have is far more detail on the IQ/training link than was usually collected in commercial car repair garages.

Second, far from being “circular” the observed correlation of .5 is in fact attenuated by range restriction. Can Taleb have made a simple statistical error? Impossible. Since the US military are allowed to screen out low ability candidates, they provide a precise test of what Taleb had asserted earlier, namely that the tests only worked for low ability people. The correlation of .5 is achieved on higher ability people. If we assume that at least IQ 100 is required, then the true correlation might be .7

13 I have here no psychological references for backup: simply, the field is bust. So far ~ 50% of the research DOES NOT replicate, & papers that do have weaker effect. Not counting poor transfer to reality.

Why does he offer no references? Probably because, in fact, the main findings in psychometrics have replicated as well as or better than other areas in psychology. It is just that many people hate the results.

14 The Flynn effect should warn us not just that IQ is somewhat environment dependent, but that it is circular.

Well, the Flynn effect did not raise Digit Span or Maths scores, so there is an ignored story here about measurement problems. Not clear that the FE is a g effect. Few psychometricians doubt that the environment affects ability. IQ is not circular, but can be determined by very simple mental tasks which are found in all cultures, and have long term predictive power on far more complex tasks.

Summary

Taleb has made sweeping assertions with great confidence and surrounded by insulting language. Those assertions may well influence people who feel unsure about intelligence, and who assume that someone who is sure of themselves must know what they are talking about. That is understandable: an unsure person is aware they need to do more reading and thinking before feeling confident, and charitably assume that only a knowledgeable person who had done the necessary reading would dare speak with confidence.

Yet, far from giving scientific references at the end of his essay, Taleb confidently asserts that he does not need to do so, because the field is broken because…. Convexity. This is presented as if it were an essential ingredient of statistical analysis, rather than one of his interesting ideas about research strategies. This is amusing, because even in the area which Taleb calls his own, as a financial instruments trader, it is easy to find a careful, long term, large sample study that shows the beneficial effects of intelligence on investment behaviour. On his own home ground he is down 1-0.

The other lapse is to ignore the decades of debate carried out by intelligence researchers, notably Jensen, to improve measures of intelligence so that they conform to the requirements set out by SS Stevens. Digit Span is such a measure. So is Digit Symbol and, if measured extensively, Vocabulary. Simple and complex reaction times are other examples. Overall, Taleb is not providing new or original insights that advance the field. But his aim does not appear to be constructive or even informative.

I don’t know why an able man is so ill-disposed to measures of ability, but can only assume he is well aware of his abilities, and regards himself as above such mundanities. He does not give references, but mentions a book he is about to publish. Better to stick to the facts.

Does Taleb’s boastful dismissal of a field he palpably does not master mean that we should dismiss his contributions to other fields? Probably not. Public figures sometimes stray out of their field of competence. It is an occupational hazard brought on by public adulation, known since Roman times. However, if he can be so bombastic when out of his depth, then it would be prudent to go back to his other writings with a slightly more critical eye. When I read his thoughts on probability I made positive assumptions about some of his pronouncements on risk on the very prudent grounds that I could not contest his mathematical excursions. Perhaps I was Fooled by Algebra. Perhaps I was not the only one.

Taleb describes himself as a flaneur, which is a stroller, the sort of person who swans about. No problem with that. Swans are beguiling, but beautiful shapes can lead us astray.

 
• Category: Science • Tags: Intelligence, IQ, Psychometrics, Race and Iq 
Hide 443 CommentsLeave a Comment
Commenters to Ignore...to FollowEndorsed Only
Trim Comments?
    []
  1. dearieme says:

    I suppose someone who’s been told that his todger is rather small might spend a lot of time shouting that Size Isn’t Everything.

    I wonder which IQ test insulted Mr Taleb.

  2. mijj says:

    speaking from ignorance and total disinclination to look anything up on the subject .. isn’t a core problem that the ‘intelligence’ that IQ tests measure a frothy astrologist-quality faith based notion?

    • Replies: @Anon
    , @Svigor
  3. Prominent Swiss writer Rolf Dobelli thought for a year or two, he’d be befriended with Nicholas Taleb. This led to incredibly unsound and unfair and unreasonable ad hominem attacks by Taleb on Dobelli. When I think of Taleb, I think of Marx (extremely short ignition span) and Bakunin: “If I had but one wish, he yelled, it would be that my feeling of embarrassment, which is holy for me, would stay in perfectly good shape until my last day! – Quietness, that’s what enrages me.” – A Mohammed without Quran.***

    *** Quotes from Mausoleum – ‘Thirty-seven Ballads from the History of Progress’ by Hans Magnus Enzensberger – : – Our age on 126 quite interesting pages (starting with the watchmaker Giovanni de Dondi from Padua***, 1318 – 1389 and ending with Ernesto Guevara de la Serna 1928 – 1967 (“Holes in the People’s War”) – in between: Turing (the best of all ballads) and American inventor Oliver Evans – the poem about him contains the lines in this book I think of most often: About the fact, that mills survived the millers.

    *** G. de D. 1318 – 1389

    Giovanni de Dondi from Padua
    Spent all his life
    Building a watch

    A watch like no one before, unsurpassed
    For centuries. (…)

    • Replies: @Jett Rucker
  4. Polynices says:

    Thank you for summarizing that all so succinctly. I’ve been dismayed by Taleb’s cluelessness but couldn’t quite articulate why. Many of his critics actually are just too dumb to understand what he’s talking about so it’s good to read intelligent criticism that engages with his “arguments” and explains his errors.

    • Replies: @Anon
    , @Alfred1860
  5. “Intelligence test items are not arbitrary.”

    This is 100% false.

    • Replies: @DFH
    , @Anon
    , @Anon
  6. Thank you, Prof. Thompson.

    Taleb is often fun to read and full of insights. His rants are often entertaining and often contain nuggets of truth. Sometimes they are spot-on. They need to be held up to careful inspection on a case by case basis.

    He has a nice section in _The black swan_ in which he distinguishes between “experts who are experts” and “experts who are…not experts,” culminating in the conclusion that “Some fields don’t have experts.” It’s a nice part of the book, and worth contemplating.

    “Some fields don’t have experts” is a provocative hypothesis and is quite possibly true–especially in cases such as stock-picking. His latest polemical output, which you address above, is probably not his best work. Thanks for listening.

    Thank you also for the time and diligence you exercise in addressing comments from the “peanut gallery.”

    • Agree: Wizard of Oz
    • Replies: @m___
    , @Bill
    , @HallParvey
  7. DFH says:
    @RaceRealist88

    1. If there are no laws connecting psychological and physical states, we cannot have evidence that physical states exist
    2. There are no laws connecting psychological and physical states [your premise]
    3. We cannot have evidence that physical states exist [reductio]

    • Replies: @RaceRealist88
  8. @DFH

    How are you defining “physical states”?

    There are no psychophysical or psychological laws. Since there are no psychophysical laws, the mental is irreducible to the physical. Since the mental is irreducible to the physical, then psychological/mental states are not reducible to anything physical – “genes”, physiology, brain states/structure.

    Therefore, psychological traits – including intelligence – are irreducible to physical states

    Provide justification for (3).

    • Replies: @DFH
    , @Loretta
  9. Anonymous[270] • Disclaimer says:

    Jeez, so many people paying attention to the intellectually insecure Taleb! He must be in havens.

    • Replies: @Wizard of Oz
  10. DFH says:
    @RaceRealist88

    “physical states”

    non-mental. I will just use that from now on to be clearer.

    Provide justification for (3).

    Whatever you think evidence is, I do not think it is controversial to say that for A to be evidence for B requires a connection between A and B. I.e. if there is no connection between two things, one cannot be evidence for the other.

    To put it more directly, if there is no connection between our mental states and any potential non-mental states, then our mental states cannot provide any evidence for non-mental states. So there’s not reason to think that non-mental states exist, or that, if they do, they are anything like our (apparent) perceptions.

    • Replies: @RaceRealist88
  11. Anonymous [AKA "The Red Quest"] says:

    If you read his books, you will find Taleb equally bombastic there.

    He is right a fair amount of the time, but it is very annoying wading through his tedious ad hominem attacks to get to the useful material.

    • Replies: @Anon
  12. @DFH

    “non-mental”

    Kinda vague… can you elaborate?

    “Physical” refers to the body; a “state” is a condition.

    “there’s not reason to think that non-mental states exist”

    Take “brain states.” A “snapshot” of the CNS. We know that that exists.

    What do you think of Davidson’s argument?

    • Replies: @DFH
  13. JLK says:

    I’ve gone into more detail on this in other threads, and won’t repeat myself here.

    Taleb makes a decent point on the differing variances between groups (something that is hard to find data on other than the male/female difference). Otherwise it seems to be an attempt to muddle the debate, and a rather clumsy one at that, especially for a person with his supposed intellect.

    There’s no doubt some politics behind this, with the Asian lawsuit against Harvard, the ever more impressive performance of Asian students on the SAT and NMSQT and the Unz Meritocracy article, which probably got some attention in certain influential circles despite the fact that he is probably on his way to being a nonperson. Consider that Taleb is kind of a house Arab for the Wall Street crowd. Here is a list of the 2018 NM Semifinalists in New Jersey:

    https://worldscholarshipforum.com/wp-content/uploads/2017/09/358901119-Semifinalists-in-the-2018-National-Merit-Scholarship-Program.pdf

    New Jersey is an interesting case because the Jewish and Asian populations are both around 6% of the population. I haven’t tried to tabulate the numbers, and some of the names are ambiguous, but a quick eyeball test tells me that Asians are punching well above their weight. This is amazing considering that the double-weighted verbal portion and the lack of a spatial component both work against the Asian strengths.

    Tying this in with Unz’s Meritocracy at Harvard article, there seem to be substantially fewer Jewish names on this list than there were in the 80s when I took the test. Much of that may be due to the truncation of the upper end (80 points were added to the SAT-V) of the verbal component that took place when they re-centered the test in 1995. There was a 1999 article on the Princeton controversy that I cited in another thread that pointed to a drop-off on Jewish enrollment at other institutions. This probably coincided with that 1995 re-centering. Through some mechanism, certain schools like Harvard, Yale, Penn and Columbia seem to have been convinced to keep Jewish enrollment at the previous level anyway. I think they missed a point in deciding to do this, which is that Jewish scores were artificially high in the first place because the tests were gamed to elevate women, who like Jews tend to be more proficient at verbal processing. By truncating the top end instead of re-balancing the test to give more weight to math and spatial, ETS and the NMSQT kept the number of women on the right tail as high as possible while reducing the Jewish ride on their coattails.

  14. Nobody says:

    I keep reading these rebuttals of Taleb on IQ, hoping someone might finally tackle him mathematically. Disappointed yet again. Instead we see more circular references back to the field whose statistical methods are under dispute in the first place, plus a lot of wasteful bloviating about his tone. The history of science will not be any more kind to you charlatans than he is.

    • Agree: onebornfree
  15. Hi Mr. Thompson,

    Thanks very much for providing some point-by-point rebuttals to Taleb, including some names that allow readers to look up relevant research for themselves. This is much better than what’s being offered by some other pro-IQ bloggers.

    Taleb’s criticisms are, for me, a mix of things I don’t care about or can’t follow well enough. However, I am interested in his claim that IQ’s predictive ability works for lower IQ scores, but not for higher ones. Based on your article, I’ve done some reading in SMPY. My understanding of what Taleb means when he says “predictive” and what Lubinski and Benbow mean are different. For example:

    Can IQ predict whether or not you can be successful at achieving a PhD?

    From Lubinski and Benbow:

    In our 20-year follow-up studies of adolescents identified at age 12, for example, 30% of participants with SAT-M or SAT-V scores of 500 or above secured doctorates, compared with 50% of those scoring 700 or above. The base rate for earning a doctorate in the United States is 1%…
    By age 33, approximately 25% of Cohort 1 had secured doctorates, in comparison with more than 30% for Cohort 2 and more than 50% for Cohort 3.

    So we can see that high-IQ achieve PhDs at a much higher rate and that the rate correlates positively with IQ. However, (1) even in top cohort, it is basically a coin toss whether or not the person has a PhD, and (2) this data doesn’t tell us anything about who tried or wanted to get a PhD, particularly in the base rate of 1%. So what good is it?

    (Obviously, Lubinski and Benbow are writing about a lot more than PhD acquisition, but this is an example.)

    The military’s aptitude tests are often cited as evidence for the predictive ability of IQ. As you note, since the military only takes in those with above-lower-than-average IQs, it is test of the right hand side. But again, only .5? I don’t understand why pro-IQ are impressed with this?

    I also wonder if the military uses IQ/aptitude tests in officer promotions, particularly at the highest levels. It seems to me that if IQ tests were predictive on the right hand side of the curve, the best way to select admirals and generals would just be to give them an IQ test. What’s wrong with this line of thinking?

  16. On a slightly different topic, I note that Sailer has re-Tweeted Pinker on IQ research non-replicability, but Pinker’s Tweet links to a page that just has bios of conference speakers. He doesn’t link to evidence:

    This Google link goes to:

    http://www.isironline.org/2015-albuquerque-new-mexico-september-18-20/

    Similarly, you assert replicability, but you don’t drop names of any researchers or provide links.

    in fact, the main findings in psychometrics have replicated as well as or better than other areas in psychology. It is just that many people hate the results.

    Perhaps you could edit this portion of your essay?

  17. utu says:
    @Chrisnonymous

    If “Replicability crisis in psych DOESN’T apply to IQ” Mr. Pinker why do you think Dr. Thompson in this article did not cite a single number for correlation, predictability of success, heritability’s and whatever else the IQ score is supposed to perform so wonderfully? He did not because he does not remember any of those numbers because they are given in very wide brackets and if he cited any of them they could be contested.

    • Replies: @Wizard of Oz
    , @CanSpeccy
  18. Hail says: • Website

    a population wide measure (general IQ) is being compared with a scholastic test taken only by a selection of brighter students (SAT) yet still does a pretty good job of showing the link between the two. This is a real-life finding, of the sort that Taleb supposedly favours.

    This raises the question of how university admissions would be handled in a world where Taleb were omnipotent dictator:

    If SAT and IQ correlate,
    and if IQ must be fought at all costs because it is useless anyway,
    then SAT is also a useless, pseudoscientific swindle,
    and therefore university admissions ought to be judged as follows:

    Holistic interview with the applicant, in which a Taleb-approved team evaluates the applicant, face to face, to determine whether the applicant is a mere paper shuffler (or worse), or well-suited to real life (in which case, the admission letter is on its way).

    • Replies: @Wizard of Oz
  19. Anonymous [AKA "RFN"] says:

    Speaking as an engineering manager

    Taleb is ass backward

    You need to fill your department with people whose IQs are 120 to 140*.

    Almost nobody with an IQ below 115 is going to succeed.

    * – I don’t object to people with higher IQs, but it’s hard to find them.

    (I can’t cite research papers. The above is based on personal experience.)

    • Replies: @res
  20. Anonymous [AKA "Amazo"] says:
    @Chrisnonymous

    I’m watching this all intently, while knowing very little about the specifics of the field. My uneducated impression is despite the coherent point by point response in the article (thank you), the author and Taleb still aren’t exactly talking to each other.

    I want to highlight the previous commenter’s observation that the pro-IQ crowd are defending IQ by insisting that psychometrics haven’t been subject to the replication crises. Yes IQ is a subset of psychometrics, but Taleb’s criticism isn’t about the big 5 etc, although I can see why that could be left in his bombastic language.

    Anyway thanks to the author, I will be banking this with any other future reports for a time that I can muddle through all the references and necessary math to figure it out.

    • Replies: @Che Guava
  21. J says: • Website

    IQ tests are predictive. Why should I let Taleb’s casuistry confuse me?

    • Replies: @Johnny Horton
  22. One aspect of his discussion is most objectionable: he tosses out technical statistical and/or mathematical terms such as convexity, transitive, monotonic, risk measure, in an extremely haphazard way. It is meant to appear dazzling to the reader, but if you are trained to understand the terms it just appears fraudulent. There is no firm connection in his pot-shot style of discussion to the rigorous statistical/mathematical definitions of these terms. Popular writing about scientific topics can be done well or it can be done fraudulently. His article smells like fraud to me. A bit like Stephen Jay Gould’s stuff, but more transparently fraudulent.

    • Agree: res, utu, Philip Owen
    • Replies: @Wizard of Oz
    , @Bill
    , @utu
  23. Regarding Taleb’s point about IQ being circular:

    It would seem that having a high IQ would certainly allow one to be more successful in a social environment that is stable enough to allow for generally predictable outcomes. However, outside of such an environment, IQ is not nearly as useful and in fact the whole concept becomes increasingly meaningless.

    That is how we should interpret the “environmental” aspects of IQ. It isn’t that factors in the environment directly raise or lower the individual’s score; it’s that the nature of the environment determines the degree to which the phenomenon appears in the first place. IQ score reflects the extent to which social predictability is embodied in the individual case. The lower the score, the more socially inert the individual is. He doesn’t perceive the richness of meaning in the world around him and remains something of a vagabond, or a lump of raw material. The higher the score, the more the realm of socially sanctioned meanings becomes identical with the individual’s own thoughts.

    I believe Wittgenstein would have agreed with this were he here to weigh in.

    • Replies: @DFH
    , @Anonymous
  24. J says: • Website

    Instead of another Jewish conspiracy, I’d say young American Jews in 2018 are not of the same quality as the 1980 generation. There has been over 70% intermarriage.

    • Replies: @Franklin Ryckaert
  25. DFH says:
    @RaceRealist88

    Kinda vague

    It’s broad but it is very precise.

    Take “brain states.” A “snapshot” of the CNS. We know that that exists.

    You might have a perception of that, but, if you believe there is no connection between mental and physical, that cannot give you any reason to think that it actually exists, as is the conclusion of my previous argument. That is the absurdity that your claims lead to.

    • Replies: @RaceRealist88
    , @Gssere
  26. J says: • Website

    P.S.: In addition to the general decline of American population.

  27. Loretta says:
    @RaceRealist88

    Then how does brain damage make people stupiderer? Asking for a friend.

    • Replies: @Anonymous
    , @RaceRealist88
  28. Completely unable to understand the fuss made by reasonable persons about IQ tests.
    IQ measures IQ, that is all.
    Nobody, as far as I know, ever was able to define intelligence in such a way that it can be measured, thus, IQ does not measure intelligence.
    IQ tests are relevant, as any employer knows who asks those applying for a job to be tested.
    Thus IQ tests also are relevant for judging in how far anyone, also migrants, fit into a European job culture.
    These things the authorities try to keep secret, but both the Dutch railways and police schools know that those with low IQ have a high chance of failing, either in simple railway work, or in completing police training.
    Police schools lowered minimum IQ in order to let more students with migration background finish the education.
    The background of the discussions of course is clear, if there are significant IQ differences a lot of complaining about racism and discrimination becomes nonsense.

  29. Anonymous[364] • Disclaimer says:

    Excellent article. Taleb is very careless with his claims and seemingly confident. That’s a bad combination.

    A honest person will carefully examine core issues and make their case. A dishonest one will throw every red herring and half-baked (and already disproved) “challenge” at the wall. They’re not even hoping that “something will stick” – they’re hoping that the sheer volume, and confidence, will trick those who don’t know the details.

    • Replies: @Anonymous
    , @Bill
  30. Anonymous[364] • Disclaimer says:
    @Chrisnonymous

    So we can see that high-IQ achieve PhDs at a much higher rate and that the rate correlates positively with IQ. However, (1) even in top cohort, it is basically a coin toss whether or not the person has a PhD, and (2) this data doesn’t tell us anything about who tried or wanted to get a PhD, particularly in the base rate of 1%. So what good is it?

    Are you joking?

    • Replies: @Chrisnonymous
  31. m___ says:
    @charles w abbott

    “peanut gallery.”

    To be coined.

    • Replies: @Sparkon
    , @m___
  32. @dearieme

    I wonder which IQ test insulted Mr Taleb.

    Jesus, good take. That’s the first thing I wondered. Second thing is, he’s like the guys of Men Going Their Own Way. MGTOW, or whatever format. They do not do well with today’s women so they’re going off, women are shit, they don’t need em. Or ugly feminists who hate beauty contests and winner’s circles at the races because beautiful girls are there. Teleb is a guy that failed IQ tests and hates the ones that show him to be less than his own high opinion. He’s hostile, but I know one test set that works when only merit is concerned.

    Me, I’m clearly a moron, barely passed Algebra 1, truthfully, I barely got through High School because girls, motorcycles, golf and pot were far more fun. I got through with C’s and D’s, period. I read a lot, we had three newspapers hitting the door every day, but I was NOT a student, or scholar. Two of my siblings were, but not me, heh

    But the Navy tested me with some sets that measured my utility, the armed forces entrance exam. AFEE? Anyway, before I signed a contract, in ONE sitdown with pen and paper, maybe an hour or two, they decided I would be fine and dandy as a flight deck rat servicing ejection seats, liquid oxygen and aircraft air conditioning. AME they called it. I excelled, I was mostly a good boy, made E5 in three years, got out honorably in nearly 6 before they could kill me.

    Throughout my service I met a million sailors and pilots, and I gotta say, 999 of 1000 of them were perfectly suited to obviously complex technical and aeronautical feats and equipped with the crazy or bravery to carry them out. And sure as heck, Naval Air doesn’t have a lot of time for OJT once you get past the basics. Their tests were accurate as hell, although I’ve read there was skullduggery there for awhile and right in that 1975ish time frame when I signed up, admitting idiots to the service. Maybe I was one of them, I dunno.

    So how did they know, in a purely Merit Based System (caps deliberate because merit is dead today) that I would rise to their standards in training and study, how did they know I was possessed of sufficient crazy to dance on flight decks, to deliver working ejection seats and breathable oxygen at age 18? Because just by looking at my High School records alone, you’d think (and I don’t deny it), ‘loser’. It would be a fair assumption. How did the military’s testing and recruiting figure out I could do it? Because even now, I’m at a loss. Anyone? How did they know in so many millions of recruits, just from these tests? They were very, very good. And so I’m just considering Teleb fulla shit.

    • Replies: @Jim Christian
    , @Anon
  33. @Jim Christian

    Oh, and BTW, how did they know I wouldn’t go batshit crazy being closed up at sea for two years at a time without going nuts? Or any of us?

  34. Let’s cut to the chase on this guy Taleb: He does not want to accept that there are racial differences in intelligence, that blacks are – in general but always – at the bottom of the heap, and he suspects his own people, whoever they might be, don’t score so well either. All this dodging the question and not naming it is tedious.

  35. APilgrim says:

    Nassim Nicholas Taleb is Lebanese, and presumably Muslim. The average Lebanon IQ is 82. https://brainstats.com/average-iq-by-country.html The Muslim country mean IQ of 81 is half a standard deviation below the mean IQ of non-Muslim nations. The mean IQ in Arab countries is 84. http://www.mankindquarterly.org/archive/issue/50-3/2

  36. Anonymous[364] • Disclaimer says:
    @Loretta

    Haha. Good point.

  37. Anonymous[364] • Disclaimer says:
    @Anonymous

    “An honest”, dammit. English is not my first language but it’s still a bad mistake in an IQ discussion.

    • Replies: @Anon
  38. DFH says:
    @Intelligent Dasein

    However, outside of such an environment, IQ is not nearly as useful and in fact the whole concept becomes increasingly meaningless.

    Do you have any evidence for this at all?

    • Replies: @Intelligent Dasein
  39. Anon[379] • Disclaimer says:
    @mijj

    Your opening words were well judged if you couldn’t resist the urge to comment at all.

  40. Anonymous[364] • Disclaimer says:
    @Intelligent Dasein

    It would seem that having a high IQ would certainly allow one to be more successful in a social environment that is stable enough to allow for generally predictable outcomes. However, outside of such an environment, IQ is not nearly as useful and in fact the whole concept becomes increasingly meaningless.

    That makes no sense. All other things being equal, a high-IQ person would have a much better chance of survival if parachuted on a deserted island. They’d figure it out better and faster.

  41. Anon[379] • Disclaimer says:
    @Polynices

    “Many of his critics”???? Surely more true of many of his followers.

  42. Anon[379] • Disclaimer says:
    @RaceRealist88

    Welcome back to the mental anarchist.

  43. @Anonymous

    I think you may have got him in one . Despite some stimulating passages in the first of his books that I tried to read, The Black Swan, the overwhelming impression I formed was of an arrogant but insecure quite clever bullshit artist.

    • Agree: Bill
  44. @JLK

    Sorry if I’m being dim but can you explain how the truncation of the top end continued to allow women to be favoured but not Jews? Maybe I don’t understand fully what the truncation involved.

    • Replies: @JLK
  45. Sparkon says:
    @m___

    To be coined.

    Don’t be a Flub-a-Dub. The term “peanut gallery” has been in common usage since about 1870 according to Google’s N-gram viewer. The term refers to the cheap seats in a (vaudeville) theater where rowdy patrons sometimes hurled their peanuts at the stage in a display of disapproval.

    Most of the boob tube generation was introduced to the term by the Howdy Doody show.

    The Howdy Doody show first appeared on radio in 1943 and made the transition to television in 1947.

    The popularity of Howdy Doody and its Peanut Gallery led executives at United Features Syndicate to use the name Peanuts for syndication of Charles M. Schulz’s Li’l Folks comic strip, reportedly to the lifelong chagrin of Schulz

    Flub-a-Dub was another of the characters on the show.

  46. @Nobody

    Having already asked Ron to arrange the mathematical disembowelling of Taleb by one of his mathematical mates if not by himself I have some sympathy with your comment. But I think you are being a bit hard on the author who has caught out Taleb not knowing the literature.

  47. Anonymous [AKA "Amstin"] says:
    @APilgrim

    Nassim is Greek Orthodox Lebanese minority and ethnic Greek. He is 75% Greek, 25% Armenian, .

    • Replies: @Anonymous
    , @Hail
  48. Bill says:
    @charles w abbott

    Taleb is excellent when he is talking about things he is expert in. His expertise doesn’t seem to extend much beyond the pricing of out-of-the-money options, however.

  49. Anonymous[863] • Disclaimer says:
    @Anonymous

    Are “Nassim” and “Taleb” Greek names?

    • Replies: @APilgrim
  50. AaronB says:

    One can construct a far more information dense portrait of a country’s or a persons potential performance by looking at their pattern of concrete accomplishment.

    IQ is an abstract figure, and thus necessarily conveys far less information than examining concrete accomplishment. It abstracts certain key features – i.e it loses information density.

    The assumption is that say, Japan and some Chinese cities have similar IQs, then we can expect the same pattern of performance from them. We know by now this is misleading.

    IQ “flattens out” contours and bumps on the map to produce a uniform figure that involves a loss of information density and detail.

    Privileging the abstract over the concrete always involves a loss of information density and richness of detail – whenever possible, the concrete is always to be preferred to the abstract (which by design selects).

    So when would IQ be useful and when not? In instances where we have no record of a persons concrete accomplishment – when we need to identify potential talent and have no history of that person’s real world performance, one may use IQ as a tentative tool.

    And that is in fact what it was designed for – never as a substitute for concrete records of accomplishment. Reality, with its rich concrete detail, was always supposed to trump an abstract figure, which is selective. IQ was meant as a stopgap in our knowledge, a heuristic, a temporary tool before a person could demonstrate competence in real life. IQ was designed to help identify potential talent – never authoritatively assert its existence, or nonexistence.

    Somehow over the years IQ became reified – and people now confidently assert that nations that have failed to demonstrate specific kinds of competence are sure to do so in the future because an abstract figure designed to help identify potential talent in the absence of a record of accomplishment is more important than the actual historical record of accomplishment.

    Or that other peoples will definitively fail to demonstrate talent in the future because this abstract figure is more authoritative than the well known historical record that demonstrates nations experience steep rises and falls in terms of concrete accomplishment over time.

    That is what it means to reify a heuristic – dramatic loss of information density and concrete detail, dramatic loss of configurability to patterns observable in reality (and eventual denial of those patterns), a loss of a sense of concrete time in favor of Platonic ideas, and an impoverished sense of reality.

    We can see also that IQ is fairly meaningless when it comes to country comparisons – the historical record of concrete accomplishment through a large variety of conditions in time conveys a far more realistic picture – and a far more granular, specific picture.

    Conversely, the well known pattern of “rise and fall” makes anything but short term predictions extremely problematic. The attempt to set up Platonic ideas of nations is ahistorical.

    Likewise, an individual’s IQ score is far less useful than his record of accomplishment – or even a long conversation – in assessing his intelligence. And a high IQ person who is consistently dull in conversation (I have personally experienced this many times) or who has failed to demonstrate competence repeatedly can be reliably said to not be very intelligent.

    Then of course there is the question of ability being measurable only in relation to motive and goal, and unless all individuals and groups can be said to be identical in these respects (which would curiously contradict the HBD premise that groups differ in all traits), measuring ability is problematic…the elephant in the room no one wishes to address…

    Anyways this is just the tip of the iceberg…

    The move towards abstract thinking and vast generalizations was historically useful for a period in time, but today, we are in a position where constructing a more intelligent approach to reality would mean a move away from the abstract and general and towards the concrete, and the loss in precision would be compensated for by a gain in richness of detail and better conformability to actual patterns in reality, and the inability to establish eternal laws and find Platonic ideas would be offset by a realistic flexibility in responding to reality.

    • Agree: wayfarer
    • Troll: DFH
    • Replies: @utu
    , @Sylsau
  51. @utu

    I suspect that there are virtual or logical replicattions rather than the pedantically identical replicattions which one might get from say slicing up 25 fresh male Scottish cadavers deceased all between 25 and 28 and all BMI 30 to 35.

    • Replies: @utu
  52. @Hail

    Hey, he and his daddy compiled an algorithmic trading system that made them a squillion just from the investors in it convinced by its simulations. Now that was real enough for admission when they put up $10 million for the Nassim Nicholas Taleb Chair of Financial Eloquence (colloquially known as the Money Talks Chair).

  53. Polymath says:

    The sound fundamental point from which Taleb polemically embarks is the real-world mismeasurement and riskiness associated with applying thin-tailed models to domains with fat-tailed phenomena. From information theory alone one can already see that such a model must fail to capture a lot. The argument that measurements can be rescaled from ordinal data to reflect this is spurious, because the construction of the tests themselves has been pushed in the direction of making raw scores follow the bell curve, so rescaling will still give a relatively uninformative measure.

    • Replies: @James Thompson
  54. Anon[379] • Disclaimer says:
    @Anonymous

    But it is anal or ignorant, whatever your first language not to recognise such an obvious typo/literal or whatever a proof reader might call it.

    • Replies: @Anonymous
  55. wayfarer says:

    Most efficient/effective, arithmetic and algebra tutorial, in the Milky Way Galaxy!
    source: https://themathpage.com/

  56. Bill says:
    @Peter Johnson

    Yup. This seems to happen when he is far away from options pricing, topically.

  57. Bill says:
    @Anonymous

    Argues like a lawyer, in other words. Argues like one who wants a particular answer, not the truth.

    • Replies: @Philip Owen
    , @Dieter Kief
  58. IQ in fact measures a somewhat limited thinking capacity, when you use a Piaget informed lens.

    This paper explains the resulting problem of a ceiling effect: https://dareassociation.org/bdev/bdb_archive/BDB_21.1-50-62.pdf

    So IQ assessment needs to be used alongside assessments that don’t suffer from this ceiling effect – unless you’re happy with only a limited picture of a person’s thinking capabilities.

    The paper is titled: ‘An Analysis of the Verbal Comprehension Index of the Wechsler Adult Intelligence Scale—Fourth Edition (WAIS–IV) Using the Model of Hierarchical Complexity (MHC): Why Might Stage Be a Better Measure of “Smarts” Than Verbal IQ?’

    This could be what Taleb is unknowingly stumbling towards trying to say? (And largely failing to).

    If I knew him, I’d ask him myself.

    • Replies: @Anonymous
    , @Wizard of Oz
    , @utu
  59. Anonymous[364] • Disclaimer says:
    @Anon

    I was correcting myself.

    • Replies: @Anon
  60. Anon[436] • Disclaimer says:
    @Jim Christian

    I too have pretty solid observational and hearsay evidence over many years of the fact that IQ tests have a basis in reality that makes them potentially useful and, on average, predictors.

    My wife’s foibles I am of course conscious of. I think my memory of many things is more accurate but she reads much faster than me and seems to pick up the plot or point of a play much more readily. She is a very good no notes speaker who can entertain and inform and has written several non fiction books. She respects my brain it seems. I recall her mother, who was a very experienced teacher, telling me that my wife had appeared dyslexic at an early age (and in fact only became a reliable speller when using scientific names in mature years) so she had her IQ tested (twice I think). Suffice it to say that it was two standard deviations above mine! That is, assuming it was Stanford Binet or Wechsler but it could well have been Cattell which would have made us roughly the same. My experienced teacher mother-in-law had no doubtbof the validity of IQ tests based on her being able to predict the results for most children she was teaching.

    As it happened I was jumped into classes of kids who had been ahead of me three times, at 7 (because I hadn’t been at school), at 13 into a class with average age 16 months greater than mine because the principal had been experimenting with a small select class the year before from which two were accelerated, and at 15 because I was propelled into top set physics and Latin with kids 2 years older. In each case I had the experience of taking one term to catch up and be at or very close to top of the class (before immaturity or the desire to be seen as a sportsman took over). How does this bear upon IQ validity?

    I recalled taking IQ tests at 9 and 10 and remembered that they had been sufficiently similar that I could in effect game the second one. However that may be, when I was in my last year at high school I had access one day to all the school’s academic (and other) records including IQ scores. There was one, a lifelong friend, whose highest was a couple of SDs, or more, above mine and which could be said to be reflected in the fact that he became a distinguished particle physicist at a top university before returning to his home country to pursue another very demanding profession at the highest level. (As I said when proposing his health and recalling, inter alia, that he wrote poetry, remembered other people’s’parts in plays and won science prizes for original projects during the summer holidays I was much better at football). The others included a number of professors and leading professionals but I was left in no doubt that the 9 or 10 point IQ advantage I had over all but one of my brightest contemporaries (who was very different in temperament and interests: he actually did the chemistry experiments) was real as I just did things quicker when motivated. I have known some very foolish people who have tested high on IQ tests, people can have bad days and then there is the Flynn Effect with its presumably multiple causes. Perhaps it takes a brain with a few IQ points above average to be able to make sound use of IQ scores.

    • Replies: @Prof Watson
  61. @dearieme

    “I wonder which IQ test insulted Mr Taleb”

    Appeal to motive.

    • Replies: @dearieme
  62. @DFH

    “You might have a perception of that, but, if you believe there is no connection between mental and physical, that cannot give you any reason to think that it actually exists”

    It’s true. We can see brain states. Thus, physical states exist.

    • Replies: @DFH
  63. @Loretta

    The brain is needed for the mental, without the brain there is no mental. This does not mean, however, that the mental is reducible to the physical.

    However, Sophocles, Shakespeare, Tom Stoppard or Elfriede Jelinek can teach us more about ourselves than neuroscience. Neuroscience attends to our brain or central nervous system and its mode of functioning. Without the brain there can be no mind. The brain is a necessary condition for human mindedness. Without it we would not lead a conscious life but simply be dead and gone. But it is not identical with our conscious life. A necessary condition for human mindedness is nowhere near a sufficient condition. Having legs is a necessary condition for riding my bicycle. But it is not a sufficient one, since I have to master the art of riding a bicycle and must be present in the same place as my bicycle, and so forth. To believe that we completely understand our mind as soon as the brain is understood would be as though we believed that we would completely understand bicycle riding as soon as our knees are understood. (Markus Gabriel, I am not a brain: Philosophy of Mind for the 21st Century)

    • Replies: @Poco
    , @Poco
    , @Loretta
  64. res says:
    @Anonymous

    What approaches do you take to find people between IQ 120 and 140?

  65. Taleb us a bright, insightful guy, but he has his IYI moments. This seems to be one.

    He is right about convexity to some extent, but in bond math it only matters because simpler math is used to model the curve around the effect of changing interest rates (duration), and therefore convexity grows as an issue as you stray from the point where you anchor the measurement. In that respect, Taleb is correct that IQ testing more accurately nails down those closer to average; there is a dispersion of IQ estimated by test around actual IQ, which is not observable, and that dispersion is far higher as you stray from average. Taleb is wrong to junk IQ testing for this reason, since a range of dispersion of estimates around an unobservable actual value is not wrong, just less precise.

    BTW, it would be helpful to add leading zeros to your decimal values.

  66. @Chrisnonymous

    The PhD rates are 50 times higher than the population base rate. That is not a coin toss. The rate at which people wanted to do a PhD would be relevant if we had a population base rate, but the completion rate is the best measure. The difference between the highest levels of intellect and the general population in doctorates is massive.

    .5 is a fairly high correlation in behavioural research. Corrected for restriction of range it would be higher, probably .6 or even .7

    IQ tests are useful at all levels of work. For higher levels a more specialized test designed for high ability candidates would be better. Demanding occupations set their own tests, specific to their requirements.

  67. Research shows that men think logically and women think emotionally. From personal experience, you can not win an argument with a woman if you try to communicate logically with her. Of course, there are some men who have gotten in touch with their feminine side and some women who have gotten in touch with their masculine side. Women have two XX and do not have a Y Chromosome. Not to mention different sex hormones.

  68. DFH says:
    @RaceRealist88

    We can see brain states.

    I grant that you can have a mental state which purports to be that of ‘seeing’ a brain scan (i.e. your visual experience of ‘looking at’ a brain scan). But if there is no link between the mental and non-mental states, then that mental state cannot provide evidence of any non-mental state (i.e. the existence of a brain scan or brain). Therefore, saying that ‘we can see brain states’ is just begging the question.

  69. @Anon

    I look at a person’s achievements, not IQ test results. There are many factors that enter into achievement besides test results. At the University, even SAT tests are viewed as racist. Community activism is rated higher. This helps to achieve our diversity quotas. On campus today, even grades are suspect due to a Professor’s latent racist biases.

  70. Anonymous[364] • Disclaimer says:
    @Miriam Spence

    IQ in fact measures a somewhat limited thinking capacity, when you use a Piaget informed lens.

    Can you please explain the problem in your own words? What is a Piaget informed lens? A quick, superficial, look at that paper doesn’t inspire confidence:

    The MHC is a nonmentalistic, neo-Piagetian,
    and quantitative behavioral development theory.
    It offers a standard method of examining
    the universal pattern of development. A fundamental
    assumption is that development proceeds
    across a large number of general sequences
    of behavior. These sequences exist in
    every domain, including, but not limited to, the
    mathematical, logical, scientific, moral, social,
    and interpersonal domains
    .

  71. @Miriam Spence

    I note that the linked paper doesn’t seem.to make any explicit reference to Raven’s Matrices. I would be interested in Dr Thompson’s opinion on how that affects the force of the paper.

    • Replies: @utu
  72. @Anonymous

    Animals live on deserted islands, and at the bottom of the sea, in the trackless wilderness, under the ice caps, and everywhere else across the globe. Animals have no reasoning ability at all, still less anything that we would call IQ.

    You should think about that for a while.

    • Replies: @Anonymous
  73. Are you attending Ron. I’m still hoping you, or one of your mathematically competent mates will Dona mathematical evisceration of Taleb.

  74. @DFH

    Do you have any evidence for this at all?

    Yes, I have all the evidence. I have the entire HBD argument.

    Black people have lower IQs. Black people have little talent for forming complex societies. Black people nonetheless survive in the African wilds amidst countless dangers—low IQs, poor organizational skills, and all.

    IQ would not seem to be necessary for success outside complex societies.

  75. Rich says:
    @Anonymous

    I don’t know about that, I’ve known some high IQ people who couldn’t hammer a nail, throw a ball or bait a hook. And I’ve known a few guys who either dropped out of high school or barely got by, who could fish, hunt, fix an engine or lift some heavy weight, that I expect would do much better on that desert island.

    Drop the high IQ guy down on Wall Street or into a lab or an accounting office, and he’d probably do better. I’ll give you that.

    • Replies: @Anonymous
    , @Svigor
  76. nsa says:

    Wiki states that back in the 1990s Unz claimed an IQ of 214. If so, how is a number over 200 measured? Or was this just a typical spurious lugenpresse jab to ridicule the Unz campaign for governor of CA? IQ claims for chess maestros like Fischer and Carlsen are only in the 180 range without any supporting methodology i.e. possibly just estimates? Do the two ends of the Bell curve ever connect forming a kind of mobius continuum? For example, Einstein’s sister / caretaker claimed that one day the great man put two socks on one foot, and then spent the rest of his day looking for the missing sock . When the sister explained the mystery, Al solved the problem thereafter by never wearing socks.

  77. utu says:
    @Peter Johnson

    he tosses out technical statistical and/or mathematical terms such as convexity, transitive, monotonic, risk measure, in an extremely haphazard way.

    Absolutely, he is trying to dazzle and confuse. But the spirit of his attack is sound.

    • Replies: @Dave Bowman
  78. dearieme says:
    @RaceRealist88

    Only a fool would ignore the possibility that an appeal to motive might be germane to understanding Mr Taleb’s emotionalism on the topic. After all it’s perfectly possible to harbour doubts about IQ tests without being hysterical about them.

  79. Mike1 says:

    I’ve studied the company Taleb is involved with and their strategy is high leverage. Not exactly ground breaking.
    I like the guy but the “you must live my life” routine gets old.
    You are right about very long term returns being around 7% but the number for the last twenty years is 3.145% (Annualized S&P 500 Return (Dividends Reinvested and Inflation Adjusted). Even to be that high requires the current nosebleed valuations.

    • Replies: @Anonymous
  80. Anonymous[364] • Disclaimer says:
    @Intelligent Dasein

    Well, even you sound like you’re living under an ice cap.

    The point was that an intelligent person would do better even in a new and unpredictable environment.

  81. RW says:

    This is like listening to someone without a background in physics pontificate on climate change.

  82. Anonymous[370] • Disclaimer says:
    @Rich

    That’s why I said: “all other things being equal”. We’re talking same knowledge, experience, toughness, willpower etc. Intelligence is a huge advantage even on a deserted island but it’s not the only one.

  83. JLK says:
    @Wizard of Oz

    Sorry if I’m being dim but can you explain how the truncation of the top end continued to allow women to be favoured but not Jews? Maybe I don’t understand fully what the truncation involved.

    Verbal is the strong suit of both women and Jews, but Jews are higher verbals on average and start to dominate on the right side tail of the distribution curve. The 1995 re-centering truncated that right side tail. In one fell swoop, everyone who previously would have scored 720-800 on the SAT-V were equal. This hurt Jews more than women as well as non-Jews, because the 720+ range is far enough out on the tail that Jews were present in disproportionate numbers and not too many women were affected. Even non-Jewish men were outperforming women in this range before the re-centering.

  84. Svigor says:
    @mijj

    Is that intended as a sentence?

  85. @Intelligent Dasein

    Black people have little talent for forming complex societies.

    Do they ?
    Jan Vansina, ‘Kingdoms of the savanna, A history of Central African states until the European occupation’, London 1966
    Jan Vansina, ‘The Children of Woot, A History of the Kuba Peoples’, 1978, Madison
    ‘The archaeology of Africa, Food, Metals and Towns’, ed. Shaw, Sinclair, Andah and Okpoko, London and New York 1993

  86. Anonymous[370] • Disclaimer says:
    @Mike1

    I’ve studied the company Taleb is involved with and their strategy is high leverage. Not exactly ground breaking.

    Sounds risky.

  87. Svigor says:
    @Rich

    You mean people who had never hammered a nail, or had no inclination to do so, or who had some relevant disability? Because I doubt you’ve actually known able-bodied, high IQ adults who couldn’t hammer a nail, etc.

    • Replies: @Rich
  88. @James Thompson

    Here’s a question that I’ve been hoping to get an informed answer to. Are their any good studies of what leads to failure in life (or even just careers) by high IQ people?

  89. DFH says:
    @Intelligent Dasein

    Black people have lower IQs. Black people have little talent for forming complex societies. Black people nonetheless survive in the African wilds amidst countless dangers—low IQs, poor organizational skills, and all.

    IQ would not seem to be necessary for success outside complex societies

    None of this shows that a higher IQ doesn’t/wouldn’t help surviving in such environments (not that Africa is for the most part a very difficult environment to survive in, putting aside the threat of disease).

  90. Anon[436] • Disclaimer says:
    @Anonymous

    Oops!!! 🙂

    • Replies: @Anonymous
  91. Gssere says:
    @DFH

    I like how he’s so dumb he keeps straw-manning that “the mental is not reducible to the physical”, when that’s actually not the claim being made. Much of our mental activity has to be non-physical, nobody has “seen” a consciousness.

    Anyone who’s old enough to have had a few too many beers knows that there is a direct connection between mental state and physical state. It’s just absurd to argue they are completely unconnected from one another.

    It’s great that he accepts brain scan data as real. I don’t know if he’s familiar with the brain scan studies where a physical change is observed in the brain (activation of motor cortex) before the subject is consciously aware of it (i.e. deciding to press a button). I don’t want to believe in physical determinism, but that’s a pretty horrible finding to contend with.

    • Replies: @jilles dykstra
  92. m___ says:
    @m___

    Every generation has it’s blind corners. You are right in correcting, the term is appropriate as to us “commenters”, that at the least, floats.

  93. utu says:
    @Wizard of Oz

    Haven’t you noticed “Verbal IQ” in the title?

  94. @Chrisnonymous

    Since Taleb thought he could dismiss a century of psychometry, there are rather a lot of references to give.
    Try searching for Detterman, as suggested in the blog.
    Here is a start /www.unz.com/jthompson/dettermans-50-years-of-seeking
    Here is a another short cut: https://www.unz.com/jthompson/intelligence-all-that-matters-stuart/
    Another short cut: https://www.unz.com/jthompson/intelligence-in-2000-words
    Then read Jensen: the g factor
    Then read the journal Intelligence.

    That will give you a start.

  95. utu says:
    @Miriam Spence

    Very good point and the paper by Featherston et al. very interesting but it will go right above the heads of simple minded IQists like Wizard of Oz who has already complained that Raven matrices were not included in the paper. Anyway, Taleb could incorporate this paper in his arguments.

  96. Agent76 says:

    May 14, 2013 The Truth about School

    Did you ever wonder how it is that kids spend 13 years from kindergarten to high school supposedly being prepared for life, yet when they get out they don’t have any real skills?

    • Replies: @jilles dykstra
    , @Anon
  97. Anonymous[370] • Disclaimer says:
    @Anon

    Thanks for caring, though. 🙂

  98. utu says:
    @AaronB

    Good post. You won’t reach the IQ-ists because they are victims of the reification. A serious deprogramming would be required. Once you go to the reified lala land it is hard to return.

    • Replies: @AaronB
  99. @Gssere

    Much of our mental activity has to be non-physical, nobody has “seen” a consciousness.

    There is no such thing as mental activity, unless the physical activity in the brain is called mental.
    Roger Penrose, ‘The Emperor’s New Mind, Concerning computers, minds, and the laws of physics’, 1989 Oxford

  100. @Wizard of Oz

    You should begin by defining what you see as failure in life.
    Was or is Faurisson a failure or a success ?

    • Replies: @Wally
    , @Wizard of Oz
  101. utu says:
    @Poco

    Than you for the link to Michael Egnor’s article.

  102. AaronB says:
    @utu

    Thanks. Unfortunately you are right.

  103. Agent76 says:

    MAY 12, 2017 This 12-Year-Old Surpasses Albert Einstein, Stephen Hawking In IQ Test

    Most of us get elated at getting full marks in class and keep boasting of it for years. But here is a girl who hasn’t stepped into her teens yet but has made India and UK proud by achieving a rare score at Mensa’s IQ test. Let’s know more this girl.

    http://dhunt.in/2hXaR

  104. @Agent76

    supposedly being prepared for life

    Prepared for what life ?
    In societies where knowledge of the Torah or the Quran is the highest wisdom this preparation is the best basis for a good life.

    Seeing one’s own society in the right way happened to me in faraway countries.
    On Sri Lanka we saw a beggar monk walking along a highway, to my surprise our guide said ‘it is an easy life’.
    They can enter any house and get food and a bed, etc.

    In medieval times young sons of not too rich small aristocrats were prepared for life by learning to ride a horse, hunt, fight, and so on.

    • Replies: @Agent76
  105. Rich says:
    @Svigor

    Are you American? Here in the USA we have a term for very smart kids, “nerds”. Usually, they were highly intelligent but physically weak. Of course there were rare exceptions. As a kid I tested well enough to be pulled out of regular class and go to classes with the high IQ people. Some of them remain my friends to this day. Only two of us were decent enough athletes to play on any high school teams and I’m pretty much the only one of our group who worked with my hands because my extended family was very involved in the mechanical trades and they made sure I learned.

    It’s just my experience, I’m not trying to hurt anyone’s feelings.

    • Replies: @jilles dykstra
  106. Forbes says:
    @dearieme

    I’ve read all of Taleb’s books from the first, and watched a lot of his various appearances, and while some might say, “he doesn’t suffer fools gladly,” I think the better aphorism is, “I’m not always right, but I’m always sure.”

  107. What keeps striking me is that few measures of anything are perfect; nevertheless, some are pretty damned good.

    You could be drunk as a lord and still drive home alright; nevertheless, statistically it would be a bad idea to let people roar around with a blood alcohol level of 0.3 per whatever.

    Similarly, IQ associates very clearly with intelligence in numerous respects; and I defy you to find many people with IQ’s of seventy who aren’t obviously stupid; or people with IQ’s of 150 who aren’t pretty sharp.

    What people mind, of course, isn’t that IQ may not be a perfect measure; it’s that the scores vary among racial groups, and blacks in particular score very low. It’s the accuracy, not the inaccuracy, that upsets them.

    • Replies: @FvS
    , @Wally
  108. Loretta says:
    @RaceRealist88

    So the mind is an epiphenomenomenomum, entirely dependant on the brain for it’s existence and its unique metaphysical status….but the health and build-quality of that brain can never determine the qualities of its attached mind….and so the relevant mind performances can never be measured fruitfully. ????

  109. @J

    As others have pointed out Richard Feynman had an IQ test done and he scored 126. How many people in Mensa have proven to be as productive and successful as Feynman? And this also begs the question – how do you define “successful in life?”. Would a person who is a member of Mensa who is an employee of say Amazon, more successful than Jeff Bezos who is not a member of Mensa and may test say 130? This is the main thrust of Taleb’s argument. From that simple example IQ is not predictive. The success of a person can wildly outstrip the mean of what is expected for life outcome of a person with equivalent IQ. Because the possible outcome of a person’s life is “fat-tailed”, one has to question the utility of IQ. IQ also doesn’t factor in real-world conditions like a person with a much higher IQ may be better at some specific tasks, but their opinion of themselves is such that a task that you as an employer are willing to pay for is is beneath their opinion of themselves, whereas someone who scores lower on an IQ test, may do a task reasonably well, and is content in doing the task thus providing stable and predictable output over a period of time – which is much more valuable in the long run. Many business owners have seen such behaviour, and that is why the term “overqualified” exists. In the end, for an individual making their way through life, it is best to ignore the whole idea of IQ, because it cannot accurately predict YOUR particular outcome and by not putting pre-conceived notions in your head, and allowing yourself to try and fail, and try and succeed, can give you the opportunity to far outstrip preconceived notions. In this respect, the concept of IQ is a product of naive rationalism in the face of human existence that is too complex to reduce, and the only way to test through that complexity is trial and error – and done with some wisdom (ie convex tinkering, experimenting, investing, exploring). Life is path-dependent.

    • Agree: utu
    • Replies: @utu
    , @Anonymous
    , @DFH
  110. Che Guava says:
    @Anonymous

    So, you are of African descent?

    • Replies: @HallParvey
  111. @JLK

    This is amazing considering that the double-weighted verbal portion and the lack of a spatial component both work against the Asian strengths.

    Those points are taken directly from Ron Unz’ article. Are you trusting his piece as an accurate source of facts? Apart from his errors on the above items he missed the larger point that everything about the PSAT is optimized for increasing the representation of Asian immigrants (which to an extent also suppresses the Jewish numbers, as Asians concentrate in most of the same states as Jews).

    The double-weighting of the g-loaded verbal section ended in 1997.

    The National Merit Scholarship selection test [PSAT] has 3 equally weighted parts: verbal (a quasi IQ test), math, and “writing skills”, a multiple-choice grammar test derived from the TSWE. Writing was added to the test to reduce the gender gap ( http://www.fairtest.org/revised-psat-debuts-october ). It is the least g-loaded part of the exam and is easily preppable. Asians outscore whites on the TSWE and adding a similar section to the PSAT helps them.

    The test does have a spatial component: the math section loads on visuo-spatial ability. The ceiling on the math that can be tested is very low, and the consequent ease (for those with a lot of extra math classes) of racking up high math scores is what lifts a big chunk of the Asian PSAT distribution into National Merit award range. Math is a cumulative academic subject where enormous test score increases can be attained systematically and predictably by years of study and practice. The 1/3 weight of math is nominally higher than, and practically about the same as, on the ACT and high school GPA.

    Having different cutoffs by state rewards Asians since in less-competitive states they dominate easy competition, and the states with high cutoffs have large Asian immigrant population centers with the full test-industrial complex.

    The PSAT is a test that Asians are much more likely to prepare for, while whites have historically treated as it as low-stakes zero risk form of SAT warmup. Roughly, white test preparation and college admission-focused activity escalates around the time of the PSAT while for Asian “tiger children” it is the first showcase for efforts started years earlier. Jews are intermediate but much closer to the preparation patterns of similar whites. There would be high Asian representation no matter what but the combination of preparation and the structure of the PSAT pushes it sky-high.

    • Replies: @JLK
    , @Ron Unz
  112. stretch23 says:
    @APilgrim

    Taleb is a Maronite Christian (or at least that’s his background.

    • Replies: @Antiwar7
  113. utu says:
    @Johnny Horton

    In the end, for an individual making their way through life, it is best to ignore the whole idea of IQ, because it cannot accurately predict YOUR particular outcome and by not putting pre-conceived notions in your head, and allowing yourself to try and fail, and try and succeed, can give you the opportunity to far outstrip preconceived notions. In this respect, the concept of IQ is a product of naive rationalism in the face of human existence that is too complex to reduce

    Exactly. One thing though: the naive rationalists are not necessary that naive. They understand the consequences and they may actually want to put “the pre-conceived notions” in in the heads of groups of peoples.

    • Replies: @Gssere
  114. @Intelligent Dasein

    Actually, the real experiment would be to instantly transport first world individuals with their current levels of education into the African context starting them off at African levels of poverty and infrastructure, and see what the survival/growth rate is then going forward If the survival/growth rate is horrific (which I suspect it would be), then you have a number of choices, either define IQ as not truly predictive of intelligence, or, high IQ can correlate to a low survival rate/growth rate, or for a society to survive and grow, the number of individuals with IQ over 150 must be below a certain percentage to correct for things like lack practical heuristics to survive/succeed in the particular environment – in which case, intelligence may very well be a quality that society should not bias towards without limit, to insure long term survival.

    And as Taleb pointed out one could also look at Northern Europe, before the Roman Empire. Why was there no superior civilization/culture North of the Alps for millenia, and then all of the sudden after 1600, Northern Europe becomes dominant. A Mediterranean person before 1600 could have easily argued that Germanics are simply not intelligent, but just adapted to the cold climate to survive – essentially the argument that you are making now. Yet who NOW has the largest economy of Europe? IQ is not predictive. Life is Path Dependent.

  115. @charles w abbott

    “Some fields don’t have experts” is a provocative hypothesis and is quite possibly true–especially in cases such as stock-picking. His latest polemical output, which you address above, is probably not his best work. Thanks for listening.

    Stock picking is predicting the future. Best left to the experts. People with insider knowledge. Like Martha Stewart.

    You can identify those who are members of the insider group by the success of their picks. Then there are the brokers who make money whichever way the players bet. The real experts. They aren’t gamblers.

  116. Gssere says:
    @utu

    Implying that the radical egalitarians haven’t been spreading their own mind virus for centuries and it has been completely harmless.

    Love this ridiculous meme

  117. Anonymous[370] • Disclaimer says:
    @Johnny Horton

    Would a person who is a member of Mensa who is an employee of say Amazon, more successful than Jeff Bezos who is not a member of Mensa and may test say 130? This is the main thrust of Taleb’s argument. From that simple example IQ is not predictive.

    Wrong. It’s predictive statistically for virtually every intellectual endeavour but it’s not a guarantee. It’s much less predictive in the maximum-money-amassed category but that’s your red herring.

    Besides, Mensa membership requires a, more or less, 132 IQ score so your imaginary Bezos is very close. He could probably get there with some luck and persistence.

    • Replies: @Johnny Horton
  118. @Rich

    Here in the USA we have a term for very smart kids, “nerds”.

    There is indeed a lot of jealousy in the world.
    Few nerds here, in my opinion.
    IQ tests just measure IQ, that’s all.

    • Replies: @Rich
  119. JLK says:
    @academic gossip

    Thanks for the substantive reply.

    Those points are taken directly from Ron Unz’ article. Are you trusting his piece as an accurate source of facts?

    I reached my conclusions independently. If Unz ever mentioned anything having to do with a spatial component, I missed it.

    Apart from his errors on the above items he missed the larger point that everything about the PSAT is optimized for increasing the representation of Asian immigrants (which to an extent also suppresses the Jewish numbers, as Asians concentrate in most of the same states as Jews).

    A close look at the facts dispels this thesis, as I’ll explain further below.

    The National Merit Scholarship selection test [PSAT] has 3 equally weighted parts: verbal (a quasi IQ test), math, and “writing skills”, a multiple-choice grammar test derived from the TSWE. Writing was added to the test to reduce the gender gap ( http://www.fairtest.org/revised-psat-debuts-october ). It is the least g-loaded part of the exam and is easily preppable.

    Things haven’t changed that much with the selection index. Wikipedia explains the three portions as “critical reading + math + writing skills scores.” There’s no getting around the high verbal weighting of two of the three.

    The test does have a spatial component: the math section loads on visuo-spatial ability.

    Do you have a citation to support this? “Load” is generally thought to mean a maximized correlation. If the PSAT-M still has even a few the old questions for the ’80s, it isn’t loaded to spatial. It may have a hybrid loading to math and spatial. That still means the East Asian spatial strength is about 1/6th of the Selection Index, while the verbal is about 2/3rds.

    The ceiling on the math that can be tested is very low, and the consequent ease (for those with a lot of extra math classes) of racking up high math scores is what lifts a big chunk of the Asian PSAT distribution into National Merit award range.

    Faulty thinking. If the ceiling is low, the Asians are actually being truncated at the top end, just like the Jews are being truncated on the SAT-V.

    Math is a cumulative academic subject where enormous test score increases can be attained systematically and predictably by years of study and practice.

    So is verbal, although most of it is probably be gleaned from reading outside of the classroom. Spatial intelligence received a lot of early criticism from the verbalists because it was said to be trainable. They based this on subjects getting better after practicing the same 3D exercises that they were testing- like giving someone the same set of WORDSUM-IQ over and over again. Now it is generally accepted by experts like Jensen that the best g-loaded tests can be formulated almost entirely within the spatial-verbal plane, with math being largely superfluous.

    The 1/3 weight of math is nominally higher than, and practically about the same as, on the ACT and high school GPA.

    The point of testing is to identify people who can do well in college and thereafter. Verbal is highly overrated in its utility to society. Considering that todays kids are texting more and reading less, its correlation to g is probably dropping as well.

    Having different cutoffs by state rewards Asians since in less-competitive states they dominate easy competition, and the states with high cutoffs have large Asian immigrant population centers with the full test-industrial complex.

    I can’t debate this, because I don’t know. Unz thinks the exact opposite effect is happening with Jews in the low-cutoff states, and he has done the legwork.

    The PSAT is a test that Asians are much more likely to prepare for, while whites have historically treated as it as low-stakes zero risk form of SAT warmup.

    Do you have hard data to back that statement up, or is it a stereotype? Do they offer Kaplan classes for the PSAT? If they do, I’d imagine that the children of well-off white collar parents would be more likely to take them, across the ethnic spectrum. In any event, there’s only so much that test preparation can do to lift scores.

    By the way, I see that you cited Fairtest. I briefly inspected the site a few months ago, and found a lot of their arguments transparently disingenuous. It reminds me of some of the material from Nation of Islam research on this site, superficially well-written but intellectually hollow. Classic high verbal, low spatial.

    • Agree: Ron Unz
    • Replies: @academic gossip
  120. Gssere says:
    @Johnny Horton

    Except Europeans completely dominated native peoples of NA and Africa for centuries.

    From a “survival” context I’m pretty sure having capability to utterly subjugate another society (say, completely annihilate them) but the converse not being true is adaptive. So you may not want to trot out “adaptiveness” as a unmitigated positive. The last 400 years at least have been a record of IQ > Technology > more IQ > sh**ting on the third world and they can’t do anything about it. That this may end in catastrophe for the West is an ongoing experiment. Maybe living in the bush is Lindy and the most stable survival mode throughout time. Who knows?

    The highest IQ people also tend to have the least amount of children, so I guess there is an argument you could have for having enough IQ to meet some threshold is good, more is marginally better, but too much is bad.

    What’s funny about your comment is we already know what happens when you put relatively- high IQ (read: European admixed) people in African conditions.

    https://en.wikipedia.org/wiki/Americo-Liberians

    Oh wow, dominant majority status, but with blaq-ish skin. It must be exposure to the white man’s constitutional form of government that caused it.

    That said, anytime someone argues that IQ isn’t relevant at an individual level, and you shouldn’t worry too much about it, they are right, but they’ll NEVER concede the actual point the IQ folks want to impress, which is at the group, collective, civilization level, it’s absolutely relevant.

    I see very very few (themselves high IQ) “IQ is irrelevant” folks signing up to compete against New Guinean tribespeople in hunting and fishing competitions with native tools to demonstrate the arbitrariness of what we call “intelligence”. My chief suspected explanation is that they tend to be mid-wits, but are limited by their intellect in climbing the Western civilizational status hierarchy, so they play a side game where they poo-poo the hierarchy for status signalling, despite being wholly incapable of surviving without electricity, medicine, running water, supermarkets, birth control and Netflix.

  121. Ron Unz says:
    @academic gossip

    Those points are taken directly from Ron Unz’ article. Are you trusting his piece as an accurate source of facts? Apart from his errors on the above items he missed the larger point that everything about the PSAT is optimized for increasing the representation of Asian immigrants (which to an extent also suppresses the Jewish numbers, as Asians concentrate in most of the same states as Jews)….The double-weighting of the g-loaded verbal section ended in 1997.

    I see that one of our numerous fanatic Jewish-activist-types has popped up again, as incompetently dishonest as always.

    As you note, the PSAT consists of three equally-weighted parts, one of which is Verbal and a another of which is Writing. Obviously, a Writing test is very closely related to the Verbal, rather than the Math or Spatial sub-components of ability.

    Pretending that a Writing test has no connection to Verbal ability is a stupid ploy even for a propagandist. Indeed, since a considerable fraction of Asian students come from a relatively-recent immigrant background, it could even be suggested that a Writing test disfavors them even more than would a simple Verbal test.

    The remaining test component is Math. Most psychometricians regard the three major sub-components of ability as being Verbal, Math, and Spatial, and the latter two are not at all the same thing. By asserting without evidence the that PSAT Math sub-test is actually a Spatial sub-test, you are pretending that the PSAT disfavors Asians far less than it obviously does.

    So just as I said, the PSAT is two parts Verbal to one part Math and lacks a direct Spatial component, therefore being ideally advantageous for Jews and disadvantageous for Asians.

    Given that the elite Ivies have long had a rigid Asian Quota, it’s hardly surprising that Asian students these days work very hard to compete for that limited number of spots, even heavily focusing on the test-prep system, that was entirely created by previous generations of Jewish activists, apparently as a deliberate means of cheating on the SAT.

    Admittedly, since the elite Ivies now admit huge numbers of totally under-qualified Jews, Jewish students have stopped working or prepping nearly as hard.

    Meanwhile, your obvious sympathy for white Gentile applicants suffering from unfair pro-Asian discrimination shines through with your every word. So it’s therefore odd that you fail to note that the Ivies, almost entirely controlled by top Jewish administrators, generally admit Jewish students at a rate roughly 1,000% greater than white Gentiles of similar ability.

  122. DFH says:
    @Johnny Horton

    If your IQ is below 110 (at least), then it is probably a waste of your time trying to become a lawyer, a doctor or a professor. It would be useful to know that.

  123. Anonymous[370] • Disclaimer says:
    @Ron Unz

    academic gossip

    The handle says it all.

  124. Maus says:
    @Poco

    #76 Poco
    Thank you for sharing the Plough article. As a Thomist, I found it gratifying to see the immateriality of the mind demonstrated by neuroscientific studies. One criticism is the author’s statement that material objects are never “about” something. Semiotics suggests otherwise. The Dominican John Poinsot, in his Tractatus de Signis (1632), explored how the relationship between an object and the intellect that apprehends it has a real existence as a signifier of meaning with a triadic nature. Poinsot was the foremost expositor of Aquinas in his day, and his work on signs is a reminder to our era that the pre-Enlightenment “Dark Ages” weren’t so dark after all.

  125. JLK says:
    @Ron Unz

    Meanwhile, your obvious sympathy for white Gentile applicants suffering from unfair pro-Asian discrimination shines through with your every word. So it’s therefore odd that you fail to note that the Ivies, almost entirely controlled by top Jewish administrators, generally admit Jewish students at a rate roughly 1,000% greater than white Gentiles of similar ability.

    As I mentioned yesterday, it looks like most of the Ivies except Princeton (right down the road from ETS) resolved to keep Jewish representation substantially unchanged after the 1995 SAT-V truncation.

    I suspect the Princeton people understood that the truncation was just correcting a longstanding Jewish overrepresentation caused by the high verbal weighting of the test and the absence of a spatial component, probably both maintained to keep women as competitive as possible. That fact that the media stopped squealing about the Princeton decline leads me to think that this was explained behind closed doors. It’s hard to believe that the Harvard people were unaware of this.

  126. Taleb describes himself as a flaneur

    Is that a typo, because he sure looks like a flamer..

    Even as a young camel herder he had that flaming look..

    • Replies: @MikeatMikedotMike
  127. Taleb’s uninformed rantings are all just variations on a single straw man argument — i.e., that g (IQ) must be a “pseudo-science” unless it can precisely predict outcomes in every specific endeavor for every specific individual.

    No one ever made this claim in the first place. In particular, general cognitive ability has never been held up as a sufficient condition for success in making lots of money (which seems to be Taleb’s main concern). I mean, the most determinist geneticists of Who doesn’t think that work ethic, personality, etc. aren’t also important factors.

    Besides, Taleb can’t remotely make the case for anything with better predictive power. His competing proposal for measuring ability is to just let people try real world tasks and see how they do. I guess Taleb wants everyone to handle a few brain surgeries or corporate IPOs and then evaluate how they did afterword. Very egalitarian, but also massively stupid.

    • Replies: @Hail
    , @utu
    , @James Thompson
  128. utu says:
    @Wizard of Oz

    I give you an example of stability of the Moray House Test scores between age 11 and 80.
    https://www.gwern.net/docs/iq/2004-deary.pdf

    The correlation between the two samples is r=0.63. The correction for the restricted range yields r=0.73. Then this correlation is corrected for attenuation (repeatability errors) which yields r≈0.80. (see p. 136).

    So we have three numbers and only one is purely empirical. The other two are improved with corrections that depend on assumption which one may question. Anyway, in this science you have to remember the three numbers 0.63, 0.73, 0.80 and you must be ready to defend the validity of the latter two.

  129. utu says:
    @Gssere

    despite being wholly incapable of surviving without electricity, medicine, running water, supermarkets, birth control and Netflix

    Try to explain this part of you argument. You can use as much of your IQ you got.

  130. Anonymous[370] • Disclaimer says:
    @Gssere

    Great post.

    The highest IQ people also tend to have the least amount of children…

    That’s entirely artificial, and (((deliberate))), in modern Western countries. Mostly through, indoctrination, wrong incentives and a social safety net turned into a warm incubator for invaders. In a cohesive, decent IQ country, you can have a formidable net without a dysgenic effect because women would always prefer real winners.

  131. Ron Unz says:

    Actually, I really haven’t been much following the underlying Taleb controversy itself, but here’s a simple question…

    Taleb likes to get favorable MSM coverage and have people pay him large consulting fees or invest with him. Doesn’t he benefit himself a great deal by making the statements he did?

    Meanwhile, if he had said the opposite, namely that IQ results are highly accurate and tell us important things about people, wouldn’t the NYT and other MSM outlets ask him “Oh, so you agree with James Watson??!!” Wouldn’t that put him in a difficult spot?

    After all, Taleb’s not a Nobel Laureate generally ranked as one of the most influential scientists of the second half of the twentieth century. He’s just a celebrity-intellectual selling books and financial-services advice. Wouldn’t his entire career quickly be destroyed?

    Regardless of what Taleb believes or doesn’t believe, isn’t he just saying what’s strongly in his personal interest?

  132. @Anonymous

    Your comment:

    “Wrong. It’s predictive statistically for virtually every intellectual endeavour but it’s not a guarantee.”

    So, just as risk models for an investment portfolio cannot predict WHEN your stock will tank, the risk models are effectively useless. This is why hedging exists, this is why insurance exists, this is why barbell investing strategies exist. Risk models cannot be blindly trusted because not only is there no guarantee of success, but there is a guarantee that failure WILL happen at some point of probability. Stock Markets are a form of financial Russian Roulette.

    This is THE central thrust of Taleb’s argument. Just like a stock risk model, IQ cannot predict if a specific person will be successful, or will be a failure – in life, or in a particular endeavor. You need other empirical signals to help with your selection AND start with the assumption of a worst case scenario and build up from there to properly hedge your choice. This goes back to Taleb’s other point – statistics should not be used without an understanding of probability. Many who use statistical methods do not realize when the game being played is a form of Russian roulette.

    IQ cannot be used as a ruler. At the very best, it’s a high pass filter – and that’s how the US military used it.

    • Replies: @Wizard of Oz
    , @Stan d Mute
  133. Hail says: • Website
    @Ron Unz

    Doesn’t he benefit himself a great deal by making the statements he did?

    This raises the question of why he “waited” so long.

    I haven’t followed the origins of this affair (seemingly originating the week before Christmas); does anyone know what sparked his anti-IQ rampage?

  134. onebornfree says: • Website

    Stefan Molyneux “rebutts” Taleb:

    Regards, onebornfree

  135. @Gssere

    So let’s start with the assumption that IQ truly is a measure of intelligence.

    How do you propose that we use this information to our benefit?

    – Do we limit immigration from locations where average IQ is equal or higher than our national average? What about an IQ test just for those applying? What about IQ tests in their native language?

    – Do we limit public education to those whose racial IQ averages are above a certain standard to preserve public funds and maximize financial resources and outcomes? Or do we provide public education to a point in age where we know that IQ won’t change, administer an IQ test, and only provide public funds to that point forward?

    – Do we impose birth controls on those individuals who score below a certain IQ – because it is believed IQ does correlate with genetics? This way we insure the average IQ keeps climbing.

    – Should all professions – doctors, lawyers, accountants take a single, nationally recognized IQ test to get a national ranking, and make that information mandatory on the professional license so I can choose the most intelligent professional I can afford?

    – Should all bankers, engineers, technicians, technologists be nationally tested, rank ordered, and published so that banks, businesses, and customers purchase the labor and services of the most intelligent that they can afford?

    I’m just throwing these out there. I’d like to see what is truly the benefit to us to use IQ as a measure.

    • Replies: @onebornfree
  136. Hail says: • Website
    @Hypnotoad666

    a single straw man argument — i.e., that g (IQ) must be a “pseudo-science” unless it can precisely predict outcomes in every specific endeavor for every specific individual.

    No one ever made this claim in the first place

    A lot of parody possibilities here.

    Steve Sailer’s reposting of an anon’s “Deadlifting Is Largely a Pseudoscientific Swindle” satire:

    Deadlifting is a pseudoscientific measure of strength. […] In competition the deadlifted weight is only raised to waist height and returned straight down. In real life streetwise physical workers may need to place weights on truck beds located above their waists.

  137. @Bill

    On target. That’s what had me blocked.

  138. @utu

    It’s not a simple paper, I must admit.

    The underlying point is that IQ more or less measures the final Piagetian capabilities of an 18 year old.

    But later research eventually found that a lot of more complex and interesting thinking emerges after 18, after what Piaget described.

    Various assessments were developed to assess this.

    IQ assessments are not designed to assess that more complex, nuanced thinking, though there is always some correlation with IQ.

    • Replies: @James Thompson
  139. Anonymous[370] • Disclaimer says:

    Just like a stock risk model, IQ cannot predict if a specific person will be successful, or will be a failure – in life, or in a particular endeavor.

    It really looks like you don’t know what you’re taking about. A basket of thousand 130+ IQ people will always outperform their low-IQ peers in the field of intellectual endeavours. Some of them will chose to spend their life surfing but that’s a given. That’s why we use statistics and never use a one-person, everything in one basket, “risk model”. Lol!

  140. Hail says: • Website
    @Anonymous

    Nassim is Greek Orthodox Lebanese minority and ethnic Greek. He is 75% Greek, 25% Armenian

    Misleading:

    Antiochian Greek Christians, also known as Rûm, are an Arabic-speaking ethnoreligious Eastern Christian group from the Levant region. They are either members of the Greek Orthodox Church of Antioch or the Melkite Greek Catholic Church, and they have ancient roots in the Levant […] The designation “Greek” mostly refers to the use of Koine Greek in liturgy

    Mr. Taleb is more properly described as of Arab Christian origin; he has an impressive personal pedigree of notable ancestors in Francophone Lebanese society.

    Too bad about the Muslims and all (and the Israelis who partly shoved them on Christian Lebanon), or Taleb might today be a bigshot in Christian Lebanon.

  141. @Bill

    The reformulation of your argument in Habermasian terms might even be a tad clearer: It is of no use to apply strategic arguments in the context of scientific discourse. –

    – In other words: It makes no good sense to pick an aspect of a chain of scientific arguments (chains of sound arguments – that’s what scientific discourses are all about) and put it in an everyday context and show thus, that at it makes no sense. This approach is senseless in itself in this context, but very useful especially when a – lawyer – tries to convince a jury, let’s say.

    The right way to prove, that a scientific argument is right or wrong is to follow the rules of the science which is at stake and see, whether this science is correct by is’ own claims and standards or not.

    Unless you want to show – like Nietzsche and Heidegger (and Derrida in their footsteps) did, for example, that a scientific worldview is wrong in itself. But that’s an endeavor which has not been mastered on twitter yet.

  142. APilgrim says:
    @Anonymous

    What does the name Talib mean? Meaning & History. Means “seeker of knowledge, student” in Arabic. Abu Talib was an uncle of Muhammad who raised him after his parents and grandparents died.

    Meaning of name Nassim Etymology : Fresh air (Arabic).

    Nicholas Name: From the Greek name Νικολαος (Nikolaos) which meant “victory of the people” from Greek νικη (nike) “victory” and λαος (laos) “people”.

    So, the author appears to have a mixed ethnic name.

  143. utu says:
    @Hypnotoad666

    Taleb’s uninformed rantings are all just variations on a single straw man argument — i.e., that g (IQ) must be a “pseudo-science” unless it can precisely predict outcomes in every specific endeavor for every specific individual.

    The straw men do exist. You can meet them here at UR pages. They indeed believe that the IQ test score is a definitive measure of humans. They conflate intelligence with this unidimensional scalar measure. They will fight you if you try to argue that perhaps we should consider more dimensions. They will fight you if you point out to them that the so called g-factor is mathematically inevitable and at the same is not unique because it depends on the battery of test used to construct the covariance matrix. They will fight you if point out that correlations claimed about the predictive powers of the IQ test score are inflated by various corrections. They indeed live in their own universe of reified belief.

    I do agree with others here that Taleb is a bullshitter and he tries to dazzle and shut up people with mathematical terms that do not have application in this case. He is a bully and and seems to be an obnoxious person. Or perhaps this is just a persona he created that helps him market and sell books and his services. He is a celebrity. But I do believe he has a heart in the right place in this argument.

    • Agree: AaronB
    • Replies: @JLK
    , @MikeatMikedotMike
  144. If you want to detect how someone fares at a task, say loan sharking, tennis playing, or random matrix theory, make him/her do that task; we don’t need theoretical exams for a real-world function by probability-challenged psychologists.

    Recruiters know this as well, which is why they tend to focus on trying to hire people who already have jobs and have proven they can do the work,

    The problem with today’s graduates is that there are too many ways to fake it, especially if you have minority status.

  145. JLK says:
    @utu

    But I do believe he has a heart in the right place in this argument.

    He’s showing his love the wrong way. Once society accepts that IQ is both profoundly important for high value-added tasks and largely heritable, the argument for income leveling and socialism gets better. A lucky set of genes has nothing to do with merit, and wealth differences have less to do with the sweat of the brow.

    That’s the main reason for elite hostility to hereditarianism.

    • Replies: @utu
    , @Wizard of Oz
  146. Rich says:
    @jilles dykstra

    As a tested high IQ person, I happily admit that we are smarter than everyone else and probably should run the world, but my experience has shown me that high IQ people generally have less aptitude for physical pursuits. Maybe I’m wrong, I’m going by what I’ve seen, that’s all. Now I’ll give you that the Dutch, your people, are highly intelligent. My wife is Dutch ad German and she and her sisters are among the smartest people I’ve known, so maybe things are different in their Motherland. Here in the greatest country that ever existed, the Nation that saved the world and dominates it to this day, there has always been a slight aversion to those who were known as bookworms. I’m not saying it’s right, I’m kind of a bookworm myself, I’m just saying that’s how it is. “Nerd” really isn’t a term based on jealousy, but maybe it’s an American thing and hard to understand.

  147. @Nobody

    Math doesn’t necessarily equate to truth. Theoretical physicists write magic math all of the time and a few years later it gets changed. Issac Newton was a great physicist but had a truly child-like understanding of biology and the natural sciences.

    Taleb has quite a few problems and the first is reality. Africans can’t build shit! If your IQ is 100 you can’t be a scientist for the most part. If IQ didn’t matter why does it work in the real world which he is constantly spouting off about? Africans are told not eat bush meat because of Ebola and what do they do….they eat bush meat. South Africans blacks rape little girls to get rid of AIDS because of their wonderful culture and belief systems and these people are going to build a dynamic civilizations.

    Taleb is a genus but he doesn’t know everything. He believes he can apply mathematical models to everything which he has preached against yet he uses it for everything because his models he assumes are correct. I don’t think IQ is what a lot of people think it is, however, there is something there because if it’s in certain range you have a potential to do certain professions or tasks. If you don’t have it….you can’t do it.

    It’s not perfect but Taleb wants throw the baby out with the bathwater and for that with all his genius he is wrong and the real world is against him.

  148. Rich says:
    @HallParvey

    No. The “African descent” canard became popular when certain groups decided that it would help with their agenda. Polygenesis is the more likely beginning of human origins. Though it is extremely unpopular in our Afrocentric society.

    • Replies: @Anonymous
    , @Wizard of Oz
  149. Anon[424] • Disclaimer says:

    • Replies: @wayfarer
  150. Mr. Thompson

    I am a neutral observer on this topic. I do not have a particular leaning.

    What I do have is a curious mind on matters and observations, and therefore seek to understand and to be understood in the light of information.

    My tone is always suggestive and never affirmative.

    I will only speak on an attempt to reconcile the observed group differences in IQ scores. Here are a couple of my highly hypothetical articles looking at worst-case situations on the topic of PGSs and between-group differences in IQ scores:
    http://4-manifold.blogspot.com/2017/10/an-attempt-to-reconcile-group.html
    http://4-manifold.blogspot.com/2019/01/an-attempt-to-reconcile-group.html

  151. Wally says:
    @jilles dykstra

    Oh, Jilles! Your best nugget ever.

    Recall that scholar Germar Rudolf was a “failure” in his quest for a PhD.

    Because The Usual Enemies of Free Speech has his PhD revoked for his study* which scientifically demonstrated the impossibility of the ‘Nazi gas chambers’ propaganda.

    * The Rudolf Report / Expert Report on Chemical and Technical Aspects of the ‘Gas Chambers’ of Auschwitz
    http://vho.org/GB/Books/trr/index.html

    more on Rudolf: https://inconvenienthistory.com/columnists/1465

    https://www.amazon.com/Hunting-Germar-Rudolf-Essays-Witch-Hunt-ebook/dp/B01N0V1C67

    • Replies: @jilles dykstra
  152. @jilles dykstra

    Why should I define my shorthand “failure in life” for the purposes of the inquiry I posed? But I might broaden the class a bit by adding “disappointed expectations”.

    How about popular high IQ kids at school who go to good universities on scholarships, get OK but disappointing degrees, talk their way into potentially good jobs, don’t really focus or show serious ambition, marry attractive spouses and have children, fail to prevent frauds by colleagues/subordinates, have to restart career in late 30s, divorce, remarriage, another divorce etc etc. ? Any problem with using “failure” as shorthand – but without inclusive and exclusive “definition”?

    • Replies: @jilles dykstra
  153. @Ron Unz

    Shrewd enough for the Hercule Poirot prize for superior little grey cells smoothly stirred.

  154. Anonymous[370] • Disclaimer says:
    @Rich

    Correct me if I’m wrong (I really don’t care that much because it can’t change current IQ distributions) but the African-origin theory rested on the fact that the oldest kinda-human fossils used to originate in Africa.

    But didn’t they recently find even older fossils in Macedonia/Bulgaria and China? I’m sure I’ve seen articles about that in 2018.

  155. @Chimela Caesar

    I noted in a glance at your second link that we are far from understanding the details of genetic causation. Your example of a superficial understanding of the genomes of light skinned Koreans causing one to expect them to be black is telling.

    • Replies: @Chimela Caesar
  156. @Rich

    One of the simpler reasons for believing in the out-of-Africa origin of homo sapiens (without ignoring the Neanderthal and Denisovan bits which actually, by their absence in Africa, tend to confirm it) is that there is vastly more genetic diversity within Africa than in the rest of the world. This most obviously results from small groups of related people leaving Africa and even smaller groups surviving in bottlenecks as they adapted or failed to adapt to changed environment, perhaps principally cold.

    • Replies: @Rich
  157. @Stan d Mute

    He kind of resembles Pablo Escobar.

    • Replies: @Wally
  158. utu says:
    @JLK

    Sorry, but I am unable to follow your logic.

    • Replies: @JLK
  159. @Gssere

    “That said, anytime someone argues that IQ isn’t relevant at an individual level, and you shouldn’t worry too much about it, they are right, but they’ll NEVER concede the actual point the IQ folks want to impress, which is at the group, collective, civilization level, it’s absolutely relevant. ”

    I think this is really getting to the heart of the IQ issue. Well said.

  160. @Chimela Caesar

    Dear Chimela

    Thank you for sending me your blog posts. You have made informed comments on several papers, and I can see that you are trying to understand and explain the puzzle of very low intelligence scores in Sub-Saharan Africa.

    Others are trying to do the same, mostly because the scores seem too low to be relied upon. The first point is the one that Jensen made in 1980 in “Bias in mental testing”, which is that if a whole group of people do better in real life than a test or exam result would predict, then that test or exam has under-estimated their real talents, and is biased against them. That leads on to a set of studies attempting to show that actual African ability is higher than the tests would predict. Chandra Chisala has argued that case. I have no argument against the view that such a finding would be very convincing, but think it has not been shown yet, partly because African populations are very large, and there will always be some very bright individuals in all populations. Another option advanced by Chisala and others is that there could be cognitive elites in Africa. I agree that this is possible, and should be investigated, but once again don’t think that there is evidence of that so far, that is no evidence that such groups are at the highest world levels.

    My current view, subject to new findings, is that the levels are probably correct, and are a mixture of genetic and environmental effects, particularly poor educational systems. Whatever the cause, the scores are low and at the moment that would predict, among other things, poor economic progress.

    Have a search through my blog for relevant comments, but here is one that partly deals with the issue:

    https://www.unz.com/jthompson/africa-and-the-cold-beauty-of-maths

    Keep blogging.

    • Replies: @anonymoys
  161. @utu

    “The straw men do exist. You can meet them here at UR pages. They indeed believe that the IQ test score is a definitive measure of humans. They conflate intelligence with this unidimensional scalar measure. They will fight you if you try to argue that perhaps we should consider more dimensions.”

    They (meaning, the bubble wrapped, super insulated intellectual credentialists) will also tell you, here at UR, not to believe your own eyes, in which you use to make observations about all different groups of people every single day of your lives.

  162. wayfarer says:
    @Anon

    Thanks for posting this video.

    Linda Ronstadt reminds me of my mother, with her beauty, her smile, her sweetness, her flawless Spanish, her connection with Mexico and its enduring vaquero culture. She worked for the CIA as a polyglot, with a photographic memory and an ability to accurately recall a seemingly unlimited volume of detail from well over a half-century in her past, and was capable of competing toe-to-toe with most Jeopardy champions.

    I’m looking forward to being with her again someday, in the magic realm of spirits.

    source: https://en.wikipedia.org/wiki/Linda_Ronstadt

  163. @Johnny Horton

    Statistical prediction is still prediction and often useful. If there were eight science graduates being considered for jobs in startups research teams and company A snapped up those who scored 131, 133, 135 and 137 whereas company B got those who scored 123, 125, 127 and 129 you would be mad to weight your investment in them equally based on likely research based outcomes. (Obviously that is not a precise real world example of utility. It merely points to the kind of reasoning that works).

  164. @JLK

    Your conclusion makes logical sense but is there anything like empirical evidence for it?

  165. JLK says:
    @utu

    Sorry, but I am unable to follow your logic.

    If the lion’s share of the economic spoils today are going to people born with smart genes, is that any fairer than feudal times when the wealth stayed with those born wealthy?

    You’ve been born lucky either way.

    • Agree: Jett Rucker
    • Replies: @Anonymous
    , @Wizard of Oz
    , @utu
  166. Anonymous[370] • Disclaimer says:
    @JLK

    You’ve been born lucky either way.

    Huge difference. One is a meaningless title and the other raw capability. It’s a matter of justice and every social animal (and beyond, probably) on the planet is sensitive to it.

    Unfortunately, in today’s decadent system, the lion’s share of the economic spoils are going to the (((unworthy))) and their lackeys. This is why everyone sane is getting restless. The system is unnatural.

  167. Rich says:
    @Wizard of Oz

    Human fossils found in Greece pre-date those found in Africa. The “Out of Africa” theory has pretty much been debunked by this discovery, but due to the new racial politics of the West, the story is given little attention. Similar to the finding of Caucasoid bones in the Americas that pre-date the American Indian invasion.

    • Replies: @Wally
    , @Wizard of Oz
  168. Anonymous [AKA "AB1"] says:

    He comes across as a complete idiot.

    I suspect he worries deeply that he is not as clever as he thinks he is.

    • Replies: @Dave Bowman
  169. JLK says:

    Huge difference. One is a meaningless title and the other raw capability. It’s a matter of justice and every social animal (and beyond, probably) on the planet is sensitive to it.

    You’ve been inured to think that. Look, I’m not advocating Marxism, just pointing out that the masses (who have been programmed to think the same thing) would be more restless if it really dawned on them that the people living in the wealthy suburbs were practically destined from birth to be there instead of deserving it because they worked harder in school.

    It’s like getting dealt a bad poker hand at birth that never changes and having to play it week after week with your luckier friends without complaining.

    Unfortunately, in today’s decadent system, the lion’s share of the economic spoils are going to the (((unworthy))) and their lackeys. This is why everyone sane is getting restless. The system is unnatural.

    That’s part of the broader issue. Most people have been led to accept the notion of money being siphoned from dull people to clever people on Wall Street. In reality, it is parasitical commerce with little or no relationship to market efficiency or raising capital for productive enterprises. This is the platform from whence Taleb speaks.

    See what I mean now about the real concerns the elite have with hereditarianism?

  170. @Wizard of Oz

    Perhaps your glance was too quick to notice that I was quoting someone else:

    Another new paper finds that the GWAS hits for IQ – largely determined in Europeans – don’t work in people of African descent. That was always a possibility: I’ve talked about it. If you look at the frequencies of height alleles (determined from GWAS in Europeans) you would predict that Pygmies are pretty short – but they’re considerably shorter than that. They have their own private alleles influencing height, which make them even shorter than you would think. Or, if you tried to estimate skin color in Koreans by the frequencies of variants that cause light skin in Europeans, you would conclude that they were black as night – but they’re not. They’re pretty light-skinned, but that’s caused by light-skin alleles common in East Asia, almost completely disjoint from the common light-skin alleles in Europeans.

    So you can’t use those GWAS hits to tell how smart sub-Saharan Africans are, at least not today. All you can use are IQ measurements and achievements. It is as if the only way we could determine your height was by using a ruler, instead of GWAS predictions.

    – Gregory Cochran, on his blog post.

    • Replies: @Wizard of Oz
  171. Anon[424] • Disclaimer says:

    Thank you for your comment . Ah , the deep Hispanic roots of the USA !

    I hope we all meet in heaven ,after our trials on earth .

    ¡ Andale !

  172. @Wizard of Oz

    That’s a great question.

    I would define failure as not achieving one’s goals, and my experience at an “elite” university suggests to me that mental illnesses and personality/social quirks are the 2 biggest contributors to failure in high IQ people.

  173. @Anonymous

    I’m not joking, but I don’t think IQ is insignificant. I would like people to articulate better their terms, claims, and significances. I am disposed to agree with the pro-IQ side, but I don’t know if they’re actually addressing Taleb’s criticisms.

    When Taleb says IQ is not predictive, I don’t think he’s talking about populations. If ypu take 2 people, IQs 120 and 150, and ask which person is more likely to be able to get a PhD, what do PhD acquisition rates tell us about that?

  174. @James Thompson

    When I refer to a coin toss, I mean what I take to be Taleb’s definition of predictibility. If we are considering admitting someone to a PhD program, and we only have his IQ score to make a decision on, how predictive is it about his success?

    In fact, I think IQ is probably very predictive, but the base rates of PhD acquisition tell us just that a random person has a 1% chance compared with a coin toss for the top cohort.

    To get an accurate picture, we should compare rates in people who have aspired to PhDs. I’m pretty sure the top cohort would be close to 100%.

    I still think my question about admirals and generals is germane too. Why doesn’t the military use IQ to pick its leaders if it uses IQ proxy to assign its enlisted? I think a fully-explicated answer to that question would answer Taleb’s non-predictability criticism.

    • Replies: @res
    , @James Thompson
  175. Antiwar7 says:
    @stretch23

    No, he’s an Orthodox Christian. Maronites are Catholics.

  176. Wally says:
    @Colin Wright

    said:
    “What people mind, of course, isn’t that IQ may not be a perfect measure; it’s that the scores vary among racial groups, and blacks in particular score very low. It’s the accuracy, not the inaccuracy, that upsets them.”

    Nicely put, Colin

  177. Wally says:
    @MikeatMikedotMike

    I was thinking Sascha Cohen with hair.

  178. Wally says:
    @Rich

    and:
    New Evidence Suggests Stone Age Hunters From Europe Discovered America
    https://www.lewrockwell.com/2018/01/no_author/new-evidence-suggests-stone-age-hunters-from-europe-discovered-america/
    “New archaeological evidence suggests that America was first discovered by Stone Age people from Europe – 10,000 years before the Siberian-originating ancestors of the American Indians set foot in the New World.”

  179. @Wally

    In the meantime there exist several more scientific investigations into the alleged gas chambers, all have as conclusion that none of them has been used for killing humans.

    • Replies: @Wally
  180. @Rich

    “Nerd” really isn’t a term based on jealousy, but maybe it’s an American thing and hard to understand.

    Being nicknamed professor in primary school, in Dutch a professor is a university lecturer, I experienced as jealousy.
    Being, when I left primary school the only one going to gymnasium, the school was even blamed because of me, in a meeting of parents.

    • Replies: @Philip Owen
  181. @Wizard of Oz

    Obviously your definition of success is the common USA dream, family, house, income, etc.
    Was Solsjenytsyn a success, yes, in my opinion.
    Was Stalin a success ?
    Not completely, in his own judgment, no world rule.
    Was Hitler a success ?
    He resurrected Germany, no doubt about it.
    Was he a failure because he underestimated anti German forces ?
    Maybe.

  182. @Ron Unz

    Admittedly, since the elite Ivies now admit huge numbers of totally under-qualified Jews, Jewish students have stopped working or prepping nearly as hard.

    If this is true, can it be the cause of that USA military progress is far behind Chinese and Russian ?
    Several indications exist that indeed the USA is far behind, of course, the incidents are debated.
    The USA, if it is true, is not going to admit openly that a single Russian plane can disable all electronic systems on board of one of the newest USA navy ships.
    This seems to have happened in the Black Sea.

  183. @Gssere

    I see very very few (themselves high IQ) “IQ is irrelevant” folks signing up to compete against New Guinean tribespeople in hunting and fishing competitions with native tools to demonstrate the arbitrariness of what we call “intelligence”

    Such a test is scientifically impossible.
    You cannot in such a test separate intelligence from experience.
    One of course may wonder if IQ tests are not similarly biased.
    I suppose they are.
    But for me it does not matter, in western societies a IQ score is a good indicator of suitability for a job in western societies.
    Any employer, me among them, knows this.

    • Replies: @Antiwar7
  184. When I read his thoughts on probability I made positive assumptions about some of his pronouncements on risk on the very prudent grounds that I could not contest his mathematical excursions. Perhaps I was Fooled by Algebra. Perhaps I was not the only one.

    Maybe you can’t contest his math-based arguments, but presumably there are thousands of people worldwide who could. Have they done so? If so, I’m sure you would have quoted them. Indeed, what stops you from taking up the subject yourself? I have (at age 64), and have found it a revelation.

    I’ve read Taleb’s first 3 books four times and found something new and stimulating each time. Yes, he is bombastic and insulting, but how is that relevant to the issue here? Is this essay partly a hit piece?

    Perhaps you are offended that NNT has little respect for the social sciences (what, exactly, makes them sciences?), mainly for their lack of Popper-style falsifiability. Cue the hit piece.

    • Replies: @James Thompson
  185. @Johnny Horton

    Why was there no superior civilization/culture North of the Alps for millenia,

    What makes you suppose this ?
    Ceasar’s propaganda ?
    Visit the Celt museum in Halle Austria, or read about Ötzi.
    For me it is proof of superior civilisation that no cruel police states did exist.
    The barbarians east of the Rhine for me were far more civilised than the Roman armies.

  186. Anonymous [AKA "duplex"] says:

    “A measure that works in left tail not right tail (IQ decorrelates as it goes higher) is problematic.”

    I guess Taleb confuses psychometric measurement quality with correlation here; does he?

  187. @JLK

    Perhaps the parable of the talents applies equally to both, but maybe not.

    It makes sense to set up a system in which people are given adequate incentive to use whatever kind of capital they have well. So there are good arguments for securing inheritance by law, at least so as to give enough incentive to living owners to be productive and prudent. Could it be argued equally that seeking out the right genes for one’s issue ought to be incentivised. Of course women never doubted that there were such incentives from nature in particular.

  188. @Chimela Caesar

    Thanks. I didn’t notice that it had the credibility of Greg Cochran behind it.

  189. @JLK

    The following answers some of your points. I will try to come back to the rest later (am pressed for time at the moment, and RKU is piping up again…)

    everything about the PSAT is optimized for increasing the representation of Asian immigrants

    A close look at the facts dispels this thesis,

    The thesis looks fine despite your comments thus far.

    Things haven’t changed that much with the selection index. Wikipedia explains the three portions as “critical reading + math + writing skills scores.” There’s no getting around the high verbal weighting of two of the three.

    Yet Asians outscore whites on SAT Writing (formerly TSWE), the very test whose score the PSAT writing section is designed to predict, despite it being “verbal”.

    The test does have a spatial component: the math section loads on visuo-spatial ability.

    …It may have a hybrid loading to math and spatial.

    There is a loading on verbal and visuospatial. I don’t think a “math” (factor or PCA component) exists for the math test score to load upon in the sense that g, verbal, visuospatial, working memory and processing speed exist. It isn’t a robust thing that exists in tests of large populations, across tests and samples.

    The ceiling on the math that can be tested is very low, and the consequent ease (for those with a lot of extra math classes) of racking up high math scores is what lifts a big chunk of the Asian PSAT distribution into National Merit award range.

    Faulty thinking. If the ceiling is low, the Asians are actually being truncated at the top end,

    No. The number of points (60) potentially available from the math test is fixed.
    If the difficulty is increased, the mean and standard deviation of the math test, and the National Merit Semifinalist selection cutoff for that state, all will drop; and it will become harder to grab extra points toward the selection by prepping for the math part of the test. You can see this easily in the extreme case where the math test is replaced by 60 questions from PhD qualifying exams. The contribution to selection index becomes zero because it’s impossible for students to gain any points on that part of the test and the Asian kids would suddenly face a test that is effectively math-free even though its math component has been strengthened.

    In any event, there’s only so much that test preparation can do to lift scores.

    Long enough prep (in multiple forms) can do quite a lot of lifting.

    Fairtest

    Fairtest is what it is, but its webpage on this matter is authoritative. The changes to the PSAT in 1997 were made by ETS to settle a complaint by Fairtest.

    • Replies: @res
    , @JLK
    , @keuril
  190. anonymoys says:
    @James Thompson

    Prof. Thompson,

    If what you wrote, “…and there will always be some very bright individuals in all populations. Another option advanced by Chisala and others is that there could be cognitive elite in Africa.”, is true, as it seems reasonable to conclude, isn’t all this debate about IQ a “non-debate” ?

    I mean, it isn’t obvious that the problem of Africa is that they failed-for several reasons – to produce a “cognitive elite” ?

    It is unreasonable to admit that if African leaders were to ask China to help them build a cognitive elite and a ruthless meritocratic system -, Africa, in less than two generations can probably have one or two nobel in physics/chemistry ?

    Of course this is wishful thinking and it won’t happen because those who are in power in Africa are unfortunately too dumb to understand the importance of having a cognitive elite…

    • Replies: @Wizard of Oz
  191. Number of STEM PHDs by average IQ people=0.

    Number of PHds by high-IQ people >0

  192. onebornfree says: • Website
    @Rich

    Rich says: “As a tested high IQ person, I happily admit that we are smarter than everyone else and probably should run the world, “

    Sir, if you were really as intelligent as you mistakenly believe you are, you would not be here fantasizing about how you and others like you “probably should run the world”, simply because you would already understand basic unchangeable laws of human action and economics, and will have already concluded that top down, centralized planning of societies [by whomever] always fails “bigly” [ as Scott Adams would say] , and ultimately cause the deaths [via war and starvation] of vast numbers of people.

    Unless, of course, that end result is what you are really after. Otherwise, its simply called “hubris” 🙂

    See: “Man Economy and State” by Murray N. Rothbard:
    https://mises.org/library/man-economy-and-state-power-and-market.

    Regards, onebornfree

    • Replies: @Rich
  193. Antiwar7 says:
    @Ron Unz

    He claims to have made enough money trading to not need any more. He calls it his “f*** y** money”.

  194. Antiwar7 says:
    @jilles dykstra

    That’s why Taleb thinks it should be called something else, like Salaryman Quotient.

  195. Rich says:
    @onebornfree

    If you’re against me, you’re against one of the high IQ people you’re supposed to worship.Believe me, some of my teachers were amazed at how high the Italian-American kid from Queens’ IQ was, too, so I was tested and re-tested, but I kept scoring at the top. I’m sorry if my genius is difficult for you to comprehend, but that’s one of the differences those of us with high IQs have with you lower IQs.

    My joke about how we should run the world wasn’t “hubris”, it was just a joke that went over your head. If you’re a fan of Scott Adams you’ll remember his garbage man character who was the smartest man in the world. If not, go look him up, maybe it’ll help you understand.

    Regards.

    • Replies: @onebornfree
  196. Curle says:

    I’m not sure what you teach but I had a law school professor in the twilight of his career tell me the faculty could and would reliably predict class ranking solely by reference to LSAT scores. This was a school where they used blind testing.

  197. Antiwar7 says:
    @James Thompson

    Taleb no doubt thinks getting a PhD is no great achievement.

    BTW, I’m part of the Lubinski and Benbow longitudinal study.

    • Replies: @res
  198. onebornfree says: • Website
    @Rich

    OK so you were joking, you don’t fantasize about running the entire world. Good to know. Got it. 🙂

    Yes, I’m very familiar with the Scott Adams garbage man character. Huge fan. I read the Dilbert cartoon daily at : dilbert.com , and have about a dozen of the Dilbert series books. My personal favorite Dilbert character is Wally.

    Regards, onebornfree

  199. res says:
    @Chrisnonymous

    Why doesn’t the military use IQ to pick its leaders if it uses IQ proxy to assign its enlisted?

    Because it actually does? This page includes a great deal of information about ASVAB subtest scores required for different jobs in the US Army: https://www.military.com/join-armed-forces/asvab/asvab-and-army-jobs.html

    Among others:
    MOS ARMY JOB TITLE Minimum ASVAB Line Scores
    09S US Army Commissioned Officer Candidate GT:110
    Where GT is the line score (combination of subtests) probably most like a traditional IQ test:
    GT – General Technical: VE+AR

    From this Quora answer (more good information, including that the threshold used to be an IQ of 120) that corresponds to an IQ of about 108: https://www.quora.com/What-is-the-average-IQ-of-US-Army-Officers
    I assumed the lowered threshold happened for the obvious reason. I wonder how the average GT scores of officer candidates from different groups compare.

    That said, part of your point still applies. IQ is used as an initial screen, but AFAIK not as a selection criterion at the general officer level. And for that I assume the reason is the best screen seems to be an initial threshold followed by a requirement to demonstrate performance. Which happens to be how much of the work world operates in the US (with college serving as an IQ screen, well, once upon a time anyway). This is the part where Taleb at least has a point (a small one, far from most of his bombast). Though I suspect the initial screen makes a big difference in training expense (e.g. washouts) and the expense of actual failure in practice.

    It would be interesting to see a study of average GT scores at different ranks in the army.

  200. res says:
    @academic gossip

    Faulty thinking. If the ceiling is low, the Asians are actually being truncated at the top end,

    No. The number of points (60) potentially available from the math test is fixed.
    If the difficulty is increased, the mean and standard deviation of the math test, and the National Merit Semifinalist selection cutoff for that state, all will drop; and it will become harder to grab extra points toward the selection by prepping for the math part of the test.

    You are smart enough to understand how the ceiling matters here. Why are you trying to talk around that?

    • Replies: @academic gossip
  201. res says:
    @Antiwar7

    Taleb no doubt thinks getting a PhD is no great achievement.

    I think part of Taleb’s problem is that he is high IQ and thinks the things he find easy are easy for everyone. While at the same time being well aware of his non-IQ deficiencies and advantages relative to lower and higher IQ competitors.

    BTW, I’m part of the Lubinski and Benbow longitudinal study.

    Cool. Can you add any comments about how it appears from the inside? For instance, do they keep you informed about the research being published? What kind of long term followup is there?

    • Replies: @Antiwar7
  202. Agent76 says:
    @jilles dykstra

    This helps explain where I come from in life.

    Nov 19, 2015 You Don’t Have To Hide From God

  203. utu says:
    @JLK

    So that’s how you envision your wealth redistribution program. Take money from the smart ones first because being smart is undeserving. Should we have progressive taxations based on IQ? Would you start with Ron Unz? But then what about taking money form the stupid ones? How would you justify it?

    • Replies: @JLK
  204. JLK says:
    @academic gossip

    Yet Asians outscore whites on SAT Writing (formerly TSWE), the very test whose score the PSAT writing section is designed to predict, despite it being “verbal”.

    What are you trying to disprove here? The literature doesn’t say that Asians are low verbals; it says that their strongest suit is on the visual-spatial axis. I would expect second or third generation Asians to slightly outperform whites on the verbal component because of their socioeconomic status, the education level of the parents, and their tendency to live in the information-rich urban/suburban parts of the country.

    You’re knocking down a strawman that you set up yourself.

    There is a loading on verbal and visuospatial.

    Highly doubtful on spatial. I’m still waiting for a citation.

    I don’t think a “math” (factor or PCA component) exists for the math test score to load upon in the sense that g, verbal, visuospatial, working memory and processing speed exist. It isn’t a robust thing that exists in tests of large populations, across tests and samples.

    I don’t think that argument actually works towards your thesis, but in any event it is wrong because it is well established in the literature that mathematical intelligence exists along with verbal intelligence and spatial intelligence. Jensen thinks the most efficacious tests of general intelligence or g can be constructed by omitting the mathematical component and testing strictly for verbal and spatial. That would mean eliminating the SAT-M and replacing it with an SAT-Spatial. I’ve come to agree with him, and your comment essentially concedes the point on this issue.

    Of course a new SAT constructed along these lines would have tremendous societal and political implications. It has hard to find a data breakdown for spatial intelligence between blacks, whites, Asians and Jews, other than Backman’s paper from decades ago that assigned Jews a spatial IQ of 91. If that’s correct, Jews would probably drop from ~25% at schools like Harvard back to roughly 1-3%, much closer to their proportion of the population, and Asians would be the ones with the ~30% representation. Personally I doubt that the mean Jewish spatial IQ is that low, and suspect that the drop-off would be more moderate, but it would be substantial. We should all keep this in mind when considering the motivations of those who keep these types of debates out of the public eye and besmirch the reputations of researchers like Jensen and anyone who delves too deeply into the topic. Privilege tends to be jealously guarded. In an normal information environment, high-end publications like The Economist would dare to run thought-provoking articles on the topic and disingenuous rebuttals would quickly be stifled by audiences that are capable of seeing through them.

    No. The number of points (60) potentially available from the math test is fixed. If the difficulty is increased, the mean and standard deviation of the math test, and the National Merit Semifinalist selection cutoff for that state, all will drop; and it will become harder to grab extra points toward the selection by prepping for the math part of the test.

    I don’t know if you really believe this, or whether you are inadvertently providing a very characteristic example of how our educational system has been damaged by promoting high verbals who are deficient in reasoning skills.

    Your are correct that average scores on the PSAT-M will drop if it is made harder. However, the most capable students will have a better chance to strut their stuff and put distance between themselves and the average students. In the case of the PSAT-M, that subset of most capable students is disproportionately Asian. On the two verbal components of the PSAT, the most capable are disproportionately Jewish. Both Jews and Asians are being penalized to some extent by the compression of the fat tail, but Asians are being penalized a lot worse because there is no PSAT spatial (despite your undocumented assertions to the contrary), and because the math component is underweighted compared to the two verbal components. Jews were hurt by the 1995 re-centering of the verbal, but the benefit that they get from the lack of spatial testing and the double-weighting of the verbal still strongly biases the college admission and NMSQT tests strongly in their favor compared not only to Asians, but to non-Jewish whites.

    A useful thinking tool is to test a thesis by taking it to the logical extreme. If we make the tests so easy that everyone has a perfect score, no group has the chance to excel.

    Long enough prep (in multiple forms) can do quite a lot of lifting.

    There is plenty of literature that quantifies the amount of lift, and it isn’t much. I think the prep courses ought to be banned because it’s not fair that the wealthy should be able to but their kids even 10-20 points on the SAT. Some of your comments about dragon ladies and so forth seem to suggest they are some sort of Asian conspiracy. I never took prep course back in the ’80s, but my recollection is that they were commonly called Kaplan courses, not Wang courses.

    • Replies: @Ron Unz
    , @res
    , @keuril
  205. @Johnny Horton

    IQ cannot be used as a ruler. At the very best, it’s a high pass filter

    That’s one of the best analogies I’ve seen.

  206. @Ron Unz

    Regardless of what Taleb believes or doesn’t believe, isn’t he just saying what’s strongly in his personal interest?

    Doesn’t he also have a book coming out soon? So this is free pre-publication publicity.

  207. @Chrisnonymous

    During 33 years of my work life, I dealt with hundreds of MDs and PhDs. While all were experts in their fields, some were better than others in those fields. Additionally, while they may have been experts in their fields, I was surprised at the large number of them who had very little understanding about the real world outside their area of expertise. This was particularly true of PhDs, who were more likely to inflate their self importance. Outside of that, they were pretty much like everyone else. Some likeable, some not.
    I have always seen IQ as correlative not causative. I have known many very intelligent people whose personalities got in the way of their success. Similarly, I have known many”average” people who are almost savant like in a particular skill or area of knowledge. A couple of them have patents that have made them a lot more money than most PhDs will see in their lifetime.

    Yes, IQ is a predictor, but not a guarantor.

    • Replies: @ThreeCranes
  208. @Ron Unz

    Taleb has no original ideas and is not a genius. but he does a reasonably good job pretending to be one, so understandably a lot of people are fooled by his facade of intellect.

  209. JLK says:
    @utu

    So that’s how you envision your wealth redistribution program. Take money from the smart ones first because being smart is undeserving. Should we have progressive taxations based on IQ? Would you start with Ron Unz? But then what about taking money form the stupid ones? How would you justify it?

    My initial point wasn’t advocating any policy. I was just pointing out that arguments for wealth distribution will gain more force if it becomes generally acknowledged that intelligence is largely heritable, and this is probably the reason that the word “merit” tends to be conflated with “aptitude” in the establishment-sanctioned information environment, when they really don’t mean the same thing.

    As far as policy goes, a good start would be to teach the cognitive elite humility, compassion and the social responsibility that comes with the gift that they have been given, because but for the grace of God and a lucky spin of the genetic roulette wheel they could be the ones getting varicose veins standing all day at the check-out counter at Wal-Mart instead of someone else.

    Some of them may claim to have compassion, but I don’t the New York Times harping on predatory credit card rates when the banks are borrowing from the Fed at zero interest, or the morality of Goldman Sachs siphoning money from genuine investors and other traders using high-frequency trading because they have better computers. I don’t see the lowbrow TV sitcoms spreading the message about how stupid it is to buy state lottery tickets or to fall into a debt trap by not paying the credit card off in full at the end of the month.

    It’s a slippery slope between the smart taking advantage of the less-smart, and actually luring the latter into situations where they can be exploited. Our societal ethics have fallen to the lowest common denominator. It reminds me of the dystopian alternative future in Back to the Future, where Biff was the mayor and the idyllic small town was full of casinos and strip joints.

    • Agree: Ron Unz
  210. Ron Unz says:
    @JLK

    An excellent summary of the arguments…

    Over the years, I’ve regularly disputed various quantitative factual issues with both Alt-Right types and Jewish-activists. But although the former have frequently seemed stubborn, ignorant, or incompetent, it’s almost only the latter who who come across as being disingenuous or outright dishonest. This “academic gossip” character has always given me that distinct impression.

    It’s a little like the difference between arguing with a fool as opposed to a swindler. I do sometimes lose my temper at the former, but much more easily at the latter.

    • Replies: @jilles dykstra
  211. keuril says:
    @academic gossip

    Long enough prep (in multiple forms) can do quite a lot of lifting.

    This is something white people by and large don’t understand because they literally can’t imagine the effort Asian students bring to test prep. This effort is mirrored by Asian families’ willingness to make much greater financial sacrifices for education than white families will, with the result that Asians also punch above their socioeconomic weight (e.g., threshold income for sending kids to private secondary schools).

  212. dvorak says:

    The reason for his doubts about the maths behind IQ, Taleb explains, is that he can computer-generate correlations based on particular assumptions which then look like some of the reported findings on intelligence and scholastic attainment.

    This is like saying that if 1 million monkeys pounding on typewriters for 1 million years could write the complete works of Shakespeare, then in fact those monkeys did create the complete works of Shakespeare. Oxfordians and Stratfordians, meet the Chimpists.

  213. @Ron Unz

    There is no point in arguing with fools, arguing with intelligent swindlers is possible, they understand when they’ve been cornered.
    The clever ones then are silent.
    Already around 1870 Brit’n Brai, spelling always wrong, formulated discussion strategy in this way.

  214. @JLK

    Should we have progressive taxations based on IQ?

    One might argue that W European countries have such systems, in as far as IQ determines income from labour.

  215. onebornfree says: • Website
    @Johnny Horton

    I notice you have received no answer to your important question[s].

    I had asked the same overall basic question in another, related thread, and likewise received no real answer [unless I’ve missed it].

    What is really being implied in this discussion?

    Is it: who gets to decide the standards and parameters for IQ tests, an individual , or the government with a “one size fits all” ruling?

    Is that what the discussion is ultimately all about?

    It seems to me that the unstated,implied assumption is that “we” need government[s] to step in a make their usual half-assed “one size fits all” ruling, and then everyone would supposedly be content and stop arguing. .

    Or am I just imagining this unstated assumption?

    For if that [government involvement in IQ testing standards etc.] is not what its about, why would anyone even care if IQ tests were used on individuals, except for the individuals concerned, who could consent or not consent to such tests depending on what they thought was in their own best interest.

    In which case, Talebs “astonishing” claims are nothing more than ” a storm in a teacup” seems to me.

    Good publicity for any forthcoming book however, as someone else has pointed out.

    Regards, onebornfree

  216. @JLK

    if taking advantage of low-IQ people is so profitable, why did these same banks almost fail in 2008. It was subprime borrowing that did them in. It’s the other way around. Low IQ people are the recipients of billions of dollars of wealth transferred from productive high-IQ people.

    • Replies: @JLK
  217. JLK says:
    @grey enlightenment2

    if taking advantage of low-IQ people is so profitable, why did these same banks almost fail in 2008. It was subprime borrowing that did them in

    They were selling them to less intelligent or less informed bankers in small town America and abroad. They only got caught with the ones that were still in the pipeline, and they knew that politicians on both sides of the aisle were so bought and paid for that they would be saved no matter what happened. The ultimate moral hazard.

    Low IQ people are the recipients of billions of dollars of wealth transferred from productive high-IQ people.

    That’s happened throughout human history. The great Cheyenne hunters got more respect than the other men, became chiefs and got the best squaws, but the other men weren’t put out to starve. Of course, it’s easier to accept that when they are cousins and friends instead of strangers.

  218. res says:
    @JLK

    Jensen thinks the most efficacious tests of general intelligence or g can be constructed by omitting the mathematical component and testing strictly for verbal and spatial.

    Do you have a reference for where Jensen says this?

    • Replies: @JLK
  219. @JLK

    So it also has to do with the variance problem that got Larry Summers into trouble when he brought it up in relation to math and physics?

    • Replies: @JLK
  220. JLK says:
    @res

    I’ll look. It was a memory cite from something I read a few weeks ago.

    • Replies: @res
  221. JLK says:
    @Wizard of Oz

    Yes, that’s why women weren’t penalized that much by the 1995 re-centering. Their variance (Standard Deviation) is somewhere in the fourteens instead of the male ~15.

  222. @utu

    Is it simplemindeness – a relatively innocent defect – that explains why you, in your previous comment to me seem to have confined your reading to half the title and very little of the article? Or why you falsely say that I complained that it made no mention of Raven’s Matrices? Or that makes you appear not to notice that it was asking whether certain tests were appropriate for measuring “smarts” thereby inevitably raising the possibility of introducing other measures?

  223. Anon[436] • Disclaimer says:
    @Agent76

    In youth I took tutorial classes in, amongst other legal subjects, contract law, and it struck me that the 19 year olds ought to have done other degrees first and grown up a bit as they didn’t really have much of a grasp of why people made contracts. Funny though, the two smartest young men in the class showed their quality by their questions and contributions and later became respectively, a Court of Appeal judge and $15,000 a day Queens Counsel.

    • Replies: @Hypnotoad666
  224. @Rich

    I have searched before asking for your sources and can only conclude that you are referring to hominids generally and not, specifically, to homo sapiens. Otherwise, please, links.

  225. @anonymoys

    It would be interesting to set up a model which included the potential intellectual elite being educated but in a way which systematically led them to do their intellectually demanding work in advanced countries. OK continue the thought and you start asking where all the African origin Nobel Prize winners are in America or Europe. Still, a comparative modernisation model that showed new problems that we don’t normally allow for would be interesting.

  226. Wally says:
    @jilles dykstra

    Indeed, Jilles, there simply were no homicidal gas chambers used by the Germans. Period.
    The chemical agent allegedly used, Zyklon-B (active ingredients: Hydrogen cyanide – HCN) was / is an insecticide.
    In lieu of the fact the Germans had large amounts of extremely potent, quick acting, efficient, & inexpensive gases like sarin & tabun available, the claimed use of Zyklon-B is truly ridiculous.
    More:
    Zyklon B: https://forum.codoh.com/viewtopic.php?f=2&t=10798
    ex:
    Even more silly was the alleged use of a pesticide, Zyklon-B, in the first place.
    The Germans had huge stocks of very efficient, low amounts required, Sarin & Tabun nerve gases they could have used had they actually wanted to exterminate people.

    and:
    The Rudolf Report: http://vho.org/GB/Books/trr/

    Here’s more about what poison gases the German could have used if they were really in the ‘extermination of Jews’ business:
    German Poison Gas (1914 – 1944): https://codoh.com/library/document/976/

    • Replies: @Wizard of Oz
  227. @res

    The ceiling matters in several ways. What do you have in mind exactly? The issues include:

    1. how much score variance is contributed by the math test. If ceiling is too high or low, variance will be low and not many points available from (math) test prep. It is in Asians’ interest that the ceiling be in an intermediate sweet spot where pushing on the math prep becomes most effective, while at the same time…

    2. structuring the PSAT to disincentivize whites from preparing (by not using it as an admission selector like the SAT, and increasing the length and complexity with addition of a third test section), but also…

    3. providing another preppable component (the Writing section) on which Asians near the margin of NMS selection, who are in any case on track to maximize their math score and probably also doing what they can toward the verbal part of the SAT, can gain more points by preparing for this subtest than by ekeing out a few more points on the math or the more g-loaded verbal.

    All these are true for the PSAT. I’d say it’s pretty well configured for the interests of subgroups that score much higher on math and out-prepare others.

    Fantasies of visuospatial puzzles or a calculus section on the PSAT (or making everyone take it in Korean or Chinese) are amusing but fly in the face of the purposes of the test: predicting SAT scores, university grades/admission, and maybe performance after university (all using elementary questions tht plausibly relate to typical high school or college coursework). Accomplishing any of those would require a substantial weighting of the verbal and both the SAT and PSAT may underweight it relative to optimal predictors. I know that in several regression analyses of University of California grades where SAT M, V and W were included, the Writing section for some reason completely dominated in predictive power, with coefficients larger than M and V combined. The most predictive National Merit selection index for college grades might be Verbal+ Math + 4xWriting or something of that nature that is even more skewed toward the verbal test than what is done now.

    • Replies: @res
    , @JLK
    , @JLK
  228. @JLK

    The 1995 re-centering truncated that right side tail. In one fell swoop, everyone who previously would have scored 720-800 on the SAT-V were equal. This hurt Jews more than women as well as non-Jews, because the 720+ range is far enough out on the tail that Jews were present in disproportionate numbers and not too many women were affected.

    This looks like a second- or third-order effect (on Jews), not something that wipes out their representation at the Ivy League schools. The schools that care about the difference between 730 or 770 have a lot of other information available in their admission decisions that did not suddenly disappear, pertaining to ability in general and verbal-type abilities in particular.

    The representation of Jews is not so different at those two thresholds, and if it were, the overall effect of obscuring the score difference is to do some limited local shuffling of the rank order of Ivy League applicants so that some Harvard-Jews randomly go to Princeton (or the reverse), some Princeton-Jews to Cornell and so on up and down the list, with motion in both directions but overall a reverse brain-drain that moves Jews to lower ranked schools. The typical percentage of Jews at the dozen or so schools that care about these score distinctions doesn’t correlate well with ranking, so some schools would gain and others lose, and it would not look like a sudden mass disappearance of Jews from the elite universities.

    • Replies: @JLK
    , @res
  229. APilgrim says:

    The University of Texas (by law) automatically admits most (75%) incoming freshmen, from the Top-6% of Texas high school GPA scorers. UT Austin is the most highly sought state university, with enrollment capped @ ~50,000 undergraduate students. The other universities in the Lone Star State have automatic admits for the Top-10% of Texas High School Graduate GPAs. In any event, most automatic admits in Texas Public Universities have Straight-A high school transcripts.

    In the 1970s, automatic admits were taken from the Top-10% of SAT/ACT scores, OR the Top-10% of Texas high school GPA scorers.

    The present system allows more low IQ Blacks, Hispanics, & Native Americans than would be automatically admitted with a Top-4% system based on SAT/ACT or GPA.

  230. Antiwar7 says:
    @res

    Mostly now it’s a periodic survey we get by email every 5 years or so, and they follow up to try to get full participation. They’ve sent us some info about their research, too. The basic gist is that the people they study tend to be quite successful in life, and don’t adhere to the stereotype of a pointy-headed intellectual stuck in a corner, resentful and passed over by society. However, I’m sure I’m not doing them justice.

    When I was a child, SMPY was run by Julian Stanley of Johns Hopkins, and at that time my family and I received copies of lots of articles and information that helped guide my schooling. (In particular, I accelerated my studies.) At one point I went there in person, to Baltimore, as part of a group to take a battery of tests (and before that I took the SAT at their behest). I met Dr. Benbow then, and later was lucky enough to have several personal encounters with Dr. Stanley: a brilliant, humble, and genuinely nice man who cared about the children he studied. His research in this area started from an attempt to help a child who was brilliant in math. Later, various academic programs were set up to help such children, including academic summer camps, which I attended as a student and subsequently as an instructor.

    My parents and I were also subjects of some of Robert Plomin’s research.

    • Replies: @Antiwar7
    , @res
  231. Antiwar7 says:
    @Antiwar7

    I forgot to mention one of the most famous and controversial papers published by Stanley and Benbow, of which you may already be aware:

    http://science.sciencemag.org/content/210/4475/1262

    It basically states that at the top end of mathematical ability, men tend to predominate, and this appears to be innate. I believe Dr. Benbow, a woman, in particular caught grief for that conclusion.

    I also believe their research tends to support acceleration versus enrichment as providing better outcomes for gifted students. I myself was lucky to have very cooperative administrations at my elementary and high schools, for which I’m eternally grateful. That was certainly not the case for many other similar students around the country at that time.

    • Replies: @res
  232. Che Guava says:
    @HallParvey

    Nice, brief, sharp.

    I do not believe the ‘recent out of Africa’ dogma. Also do think that the aquatic/littoral stage idea does apply to many east Asians, many northern Europeans, some other groups. Physical evidence is there. Subcutaneous fat distribution, even in the slim, streamling of hair (body and head) for water, more.

    A valid hypothesis, and with much physiological backing, but somebody goes ‘oh, I don’t have any likely aquatic/littoral traits, so I hate this idea’, others also lacking take up the idea, et voila!

    A valid idea is beyond the bounds of rational discussion … at least for many years.

  233. res says:
    @JLK

    Thanks.

    • Replies: @JLK
  234. @Polymath

    I think that there is a simpler explanation, which is that many investors do not look at historical records, and thus are fooled by under-sampling the appropriate time periods. For example, Argentina has defaulted 8 times since 1816, yet some investors act as if that never happened.

    As to the raw scores on intelligence tests being pushed into a bell curve, this is partly true, but not necessarily a problem when compared with some other options. For example, one might specify levels of difficulty in Maths so as to construct an absolute scale, but it would be hard to maintain such a set of questions as a benchmark once they became generally known.

    The SMPY is very close to that sort of absolute standard (full marks by age 13 on a Maths test that most 18 year olds have difficulty with) and this has no bell-curve forcing, and works very well.

    • Replies: @Wizard of Oz
  235. res says:
    @academic gossip

    The ceiling matters in several ways. What do you have in mind exactly?

    If you were being responsive rather than just trying to talk around this (you are quite good at that, too bad I have negative respect for that “skill”) you would start with the excerpt you were originally responding to and which I quoted:

    Faulty thinking. If the ceiling is low, the Asians are actually being truncated at the top end,

    With your response (which I also quoted) being:

    No. The number of points (60) potentially available from the math test is fixed.

    The second sentence is clearly true (and important), but ignores that the distribution of those 60 points matters greatly for resolving power in a given range (i.e. the top end here). Thus by no means justifies the preceding No.

    Your followup response which most closely matches this is:

    1. how much score variance is contributed by the math test. If ceiling is too high or low, variance will be low and not many points available from (math) test prep. It is in Asians’ interest that the ceiling be in an intermediate sweet spot where pushing on the math prep becomes most effective, while at the same time…

    You then go on to assert that the PSAT is in that sweet spot for Asians.

    Based on what? Do you have data on the PSAT subtest results by race? The closest thing I am aware of is Dorans 2004 which I think is the definitive reference for the 1995 SAT recentering: https://www.ets.org/Media/Research/pdf/RR-02-04-Dorans.pdf

    What matters here (identifying NMSF) is not total variance captured. It is resolving power at the top end. NMSF is looking for the top 1% of students so that is the range (say double that to account for chance and subtest ceiling effects) that matters. For anyone interested, here is a 2016 look at NMSF qualification by state: https://www.compassprep.com/national-merit-semifinalists-by-state/

    Looking at Figures 18-21 of Dorans (2004) we see SAT subtest distribution histograms broken down by 50 point buckets for four groups (blacks, Hispanics, Asians, and whites) . A bit over 4% of Asians and 1% of whites fall in the top (750-800) SAT math bucket. Assuming normal distributions it is clear that Asians will be truncated more by lower math ceilings.

    And before anyone says “but verbal…”, Asians actually had a larger percentage than whites in the top verbal bucket as well (but by a smaller margin). Though I suspect the comparison with Jews (has anyone ever seen data for that?) would look different.

    I think this is more than sufficient evidence to show how lowering the math ceiling would adversely affect Asians relative to other races in the NMSF selection process. Do you disagree with that specific assertion?

    P.S. At this point I think Ron is right. You argue like a swindler. I’m still trying to figure out if underlying that you are as smart as you think you are (and I bet your credentials indicate). My current conclusion is that you are quite smart, but think the rest of us are stupid. The latter is actually negative evidence for the former. One of the primary markers I use to judge intelligence is the ability to see it in others.

    P.P.S. Here is some recent discussion of the PSAT ceiling (late 2017) which asserts it is lower now: http://giftedissues.davidsongifted.org/BB/ubbthreads.php/topics/240599/Re_Ceiling_Effects_of_new_PSAT.html
    One good thing about gifted forum discussions is the parents tend to be smart too ; )

    Does anyone have good data on PSAT ceiling by subtest and distributions? Especially broken out by race?

    • Agree: Ron Unz, JLK
  236. @Miriam Spence

    Wrong. Try putting Wechsler in the search bar.

  237. JLK says:
    @academic gossip

    This looks like a second- or third-order effect (on Jews), not something that wipes out their representation at the Ivy League schools.

    I never said that it wiped out their representation at Ivy League schools, only that it hurt their scores somewhat. My main point was that the situation would change considerably if there weren’t two verbal subparts and a spatial part was added, either in lieu of or in addition to the SAT-M.

    The published average SAT score of Harvard admits in 1990 was 1390. It’s reasonable to think at least the top quartile had an SAT-V over 720.

    The representation of Jews is not so different at those two thresholds, and if it were, the overall effect of obscuring the score difference is to do some limited local shuffling of the rank order of Ivy League applicants so that some Harvard-Jews randomly go to Princeton (or the reverse), some Princeton-Jews to Cornell and so on up and down the list, with motion in both directions but overall a reverse brain-drain that moves Jews to lower ranked schools.

    That’s no doubt true, which is why the situation at Harvard is of considerable interest. It is at the top of the heap.

    • Replies: @res
  238. res says:
    @academic gossip

    This looks like a second- or third-order effect (on Jews), not something that wipes out their representation at the Ivy League schools.

    Perhaps. One aspect which I think matters is the way the verbal ceiling change affects reported average SAT statistics. I think this could have hurt Jews by reducing the impact of their high verbal scores (as a group) on the overall averages.

    This is probably the kind of thing you had in mind by “second- or third-order effect”, but I am not sure it is so easily dismissed.

    I do think the lower ceiling made it easier for universities to replace higher ability students with lower ability students with more desirable demographic characteristics without harming the averages too much. Though the change to reporting medians and 25-75% ranges indicate that still wasn’t enough ; )

  239. @Wally

    You set me off Wally. I Googled for Zyklon B from which I learned that it was a very efficient killer of people as well as being able to be explained away as a fumigation or deluding agent. And if the lastmentioned wasn’t enough reason for choosing it there was also the safety consideration for the users. A conversation at SS HQ might have gone something like this.

    “Herr General! It says here that, although Zyklon is so cheap the manufacturers of this new stuff Sarin which will kill an elephant in 30 seconds is available as a loss leader for the same price as Zyklon”

    “Ah, Walther my boy, spoken like the Yid I have always supposed is somewhere behind those too-blue eyes! Save money, by all means. You are to be commended. But I regret to say that the Führer’s splendid Aryans that we send to guard the camps are all Wallys who wouldn’t know how to wash their dicks if we didn’t have some doctors in the camps, and health and the standard safety training simply isn’t enough if we were to let one of those Wallys get his hands on our super gases”.

    I should add that, according to links your comment ultimately sent me too, German witnesses who, and/or whose diaries, said they witnessed substantial parts of the gassing deaths of those not selected for work (or medical experiments) included Pery Broad, Dr Johannes Paul Kremer and Rudolf Hõss amongst others. Diaries of a doctor about special actions and people screaming might be thought quite strong evidence.

    • Replies: @Wally
  240. @Chrisnonymous

    Sorry, cannot understand why you do not see that brighter people are more likely to complete a PhD. 50 times more likely. Does that mean, absent all other knowledge, that a single brighter person is more likely to get a PhD than a less bright person. Yes.

  241. res says:
    @Antiwar7

    Very interesting. Thanks! I have read a fair amount of their research and consider it groundbreaking and important. Have you seen Steve Hsu’s blog posts about the SMPY? As you said, it is very useful for rebutting some of the usual stereotypes of high IQ people with real data.

    If you don’t mind my asking (i.e. no need to reply if you prefer not to), which cohort were you in? IIRC they varied from top ~1% to ~0.01%. I suspect they would be more careful following up with the latter since they are so much rarer.

    Are you part of the genetics side of Robert Plomin’s research? IIUC some researchers have been incorporating that in a case/control manner in their intelligence GWAS.

    • Replies: @Antiwar7
    , @Antiwar7
  242. @Taleb Fanboi

    If you look through my blog, you will find that I am often critical of the social sciences, particularly when papers use small and unrepresentative samples, and do not consider confounding variables. Also, if you read my piece again, you will see that I agree with Taleb on various points. I too am critical of models which seek to show that risk has been designed out of particular financial products. Long data sets (at least 50 years) are enough to make the point.

    As to a “hit” piece, tell me which points I have got wrong, or where I have been unfair to Mr Taleb. My view is that he made unsupported assertions about intelligence testing, and in this and the following post I provided the readings to back up that claim.

  243. res says:
    @Antiwar7

    Thanks! I think I managed to miss that paper. For anyone else interested, here is a free text copy: http://beck2.med.harvard.edu/week6/Benbow%20and%20Stanley.pdf

    Acceleration is an interesting question. I was precocious intellectually, but much less so physically and emotionally which made that decision more difficult. Did they look into acceleration considering things like that? One concern I would have is acceleration making participation in high school sports much more difficult. Though age based youth leagues might be a good substitute.

    It sounds like being part of their research was very helpful to you. Glad to hear that. Sounds like exactly what Dr. Stanley was trying to accomplish.

    • Replies: @Antiwar7
  244. @James Thompson

    You might like to check your thoughts about investors’ failings by reference to Jim Collins’s two heavily researched books “Built to Last” and “Good to Great” which both impressed me very much when I read them years ago. But I see now an article titled “Was Built to Last built to last?” which reminds me of Tom Peters and Bob Waterman’s 1980s book “In Search of Excellence” which got those two ex-McKinsey men on the worldwide lecture circuit. At dinner with one of them four years ago he genially admitted, when I tentatively mentioned that Delta airlines had gone bankrupt, that quite a lot of their carefully chosen well researched companies hadn’t done too well.

    • Replies: @James Thompson
    , @res
  245. @utu

    If the “spirit of his attack” (whatever that means) is “sound” – which the endless sullen, spoiled, vicious, back-biting ad-hominems from a supposed highly-intelligent, balanced, detached “professional” indicate is most certainly NOT – why does he feel the need to “dazzle and confuse” ?

    You, being an (ethnic minority) cretin, are an apologist for an (ethnic minority) cretin. Ergo you are a cretin.

  246. res says:
    @JLK

    The published average SAT score of Harvard admits in 1990 was 1390. It’s reasonable to think at least the top quartile had an SAT-V over 720.

    Maybe (I assume you mean 720 or over?), but I’m not sure about that given the imbalance in M/V difficulty at the time. My guess would be more like 700, but I would be very interested in real data. I might be biased and/or wrong given I have a STEM background and am more used to greater M/V tilts than would be typical at Harvard.

    Looking at recent Harvard data (after I wrote the above) we see a top quartile V of 790: http://www.prepscholar.com/sat/s/colleges/Harvard-SAT-scores-GPA
    Which according to this equivalency table corresponds to a pre-recentering 720: http://www.greenes.com/html/convert.htm
    So maybe you are right. This assumes SAT scores have been stable since 1996 though, which I am skeptical about.

    Here is some historical information from Swarthmore (1990 average 1342) which includes quartiles, mean, and median by subtest for 1990: https://web.archive.org/web/20080325012348/http://www.swarthmore.edu/Documents/administration/ir/SAT.pdf

    Their top verbal quartile is 700 so I would say you are probably right after all.

  247. Sylsau says: • Website
    @AaronB

    This is a great comment.

  248. keuril says:
    @JLK

    There is plenty of literature that quantifies the amount of lift, and it isn’t much.

    Those studies probably focus on conventional SAT courses, which are a waste of money. I doubt they consider courses that kids start in fifth grade and stick with for five or six years (including summer and winter breaks). Or courses where students memorize the answers to thousands of actual questions from previously administered tests (which questions are recycled on new tests).

    I think the prep courses ought to be banned because it’s not fair that the wealthy should be able to but their kids even 10-20 points on the SAT.

    Well, these courses are now also available for free from Khan Academy, and they are probably just as good. I don’t think there’s much of a socioeconomic argument here. In any case it’s a red herring, considering that the more effective test hack used by the rich is to pay a doctor or psychologist fees to write up a learning disability diagnosis, and then establish an IEP so as to get special accommodations. Double time is quite helpful on the ACT. It would be interesting to learn what percentage of top 1% scorers have a “learning disability,” but this information is not publicly disclosed AFAIK. I’ve read the odd article claiming that high percentages of private school students receive accommodations.

    • Replies: @res
  249. @Anonymous

    Nailed in two lines.

    I have found many of the entries on this particular Comments thread utterly unnecessary, irrelevant and pointless. I have always noticed in my life that there are very many people in any area of life or experience who seem to enjoy talking about highly-technical, complex, specialist subjects about which they actually know little or nothing.

    And I’m perfectly sure he’s nowhere near as clever as he thinks he is – and never will be.

    Bravo.

    • Replies: @JLK
  250. res says:
    @keuril

    It would be interesting to learn what percentage of top 1% scorers have a “learning disability,” but this information is not publicly disclosed AFAIK.

    That is an interesting question. Learning disability diagnosis does seem like an area that is ripe for gaming the system.

    Here is a 2005 ETS report on The Impact of Extended Time on SAT® Test Performance: https://files.eric.ed.gov/fulltext/ED563027.pdf

    Their three major findings were:

    • Lower-ability test-takers gain little or no benefit from extra time. If students do not have the knowledge or skills, no amount of extra time will improve performance.
    • Section breaks appear to help test-takers at different ability levels, regardless of their disability status.
    • Extra time helps medium- and high-ability test-takers with and without disabilities. Extra time, however, does not help and actually may hinder low-ability students with disabilities.

    Figures 15 and 16 clearly show a benefit of extra time for performance on the later items in the test.

    Here is an article discussing extended time: https://edsurgeindependent.com/a-case-against-extended-test-time-71aa4a82148d
    They advocate for untimed tests.

    It includes this quote which matches your final sentence:

    According to the College Board, up to 46% of students at elite high schools receive special testing accommodations, including extra time. The natural proportion of learning disabilities should be somewhere around 2%.

    Referencing the study I linked above they also state:

    students with average academic abilities and no learning disability scored approximately 43 points higher on their SAT when given extended time.

    • Replies: @keuril
    , @Wizard of Oz
  251. keuril says:
    @res

    Interesting info, thanks for posting. Of course the 2018-19 SAT is very different from the 2005 version. SAT verbal had a reputation for “trick questions” (i.e., g-loaded) and was losing market share to the more straightforward ACT. So ETS dumped the IQ test-style questions for something more akin to the ACT a couple yrs ago. It would seem that the ACT (and, perhaps, the new SAT), which are more performance-based (answering a lot of less-tricky questions, but more quickly), would be more amenable to time-hacking, for Moms who are hooked into the right kaffeeklatsches when Junior is still young enough to get an IEP that passes muster (reportedly you have to get the diagnosis *early*, like preferably elementary school).

  252. JLK says:
    @res

    This isn’t the Jensen citation that I remember reading a few weeks ago, but it refers to verbal and spatial as primary group factors throughout, sometimes but not always in conjunction with memory, but never with mathematical.

    http://arthurjensen.net/wp-content/uploads/2014/06/Psychometric-g-and-Achievement-1993-by-Arthur-Robert-Jensen.pdf

    This particular article is interesting because it suggests that the spatial group factor increases the separation between blacks and whites, but verbal, contrary to popular belief, does not after correcting for I. That’s probably another reason we see two verbal components on the PSAT/SAT and no spatial component.

    Incidentally the idea that verbal and spatial are orthogonal enough to make a good intelligence test by themselves goes back to at least Swineford (1948), with many appearances in the literature since then. Johnson and Bouchard’s “VPR” model (mid ’00s) includes it, and has been tested with some really interesting functional MRI studies that show where verbal and spatial thinking are performed within the brain.

    • Replies: @res
  253. JLK says:
    @Dave Bowman

    I have found many of the entries on this particular Comments thread utterly unnecessary, irrelevant and pointless. I have always noticed in my life that there are very many people in any area of life or experience who seem to enjoy talking about highly-technical, complex, specialist subjects about which they actually know little or nothing.

    I absolutely agree. It’s a travesty that the college admission testing system has to be put under the microscope by a group of dabblers on a free speech site, piecing together their research and analysis using Google and duct tape. Harvard ought to be doing the research and publishing itself instead of covering its tracks, and NYT, WaPo, WSJ and The Economist should reward their subscribers with some quality reporting and analysis so that the graduates of these schools can weigh in on the debate instead of being kept completely in the dark.

    • Replies: @Anonymous
  254. res says:
    @JLK

    Interesting. Thanks!

    For anyone else interested, here is a quote (which you paraphrased nicely) from page 23/139 of that link:

    A spatial visualization factor is the only non-g factor that rather consistently rivals g in its loadings on the black/white variable (see also Naglieri and Jensen, 1987). Hence, the largest black/white mean difference is seen on those tests that are the most highly loaded on both g and a spatial factor. The smallest black/white mean differences occur on tests that are the least loaded on g and the most highly loaded on a short-term memory factor. Contrary to popular belief, the mean black/white difference on the verbal factor (independent of I) is nil.
    Examination of 121 psychometric tests that were factor-analyzed in eleven studies also showed that the g loadings of various tests are distributed as a continuous variable extending over a wide range of values – from about .30 up to nearly .90. On the same set of tests, the black/white mean
    differences (expressed in standard deviation units) are also distributed as a continuous variable, ranging from close to zero up to about 1.3 standard deviations (SDs). From the linear regression of the mean black/white differences on tests’ g loadings, the estimated mean difference on a hypothetically pure measure of g would be approximately 1.2 SDs.

    Below the more is an extended quote from pp. 156-158 which I think speaks to some of what Taleb wrote.

    [MORE]

    It has often been noted in reviews of test validities in personnel selection that validity coefficients for predicting job performance within specific occupations are rather disconcertingly low, for the most part in the range of .20 to .30. The reason usually given for this is restriction of range of ability within occupations. But this is only one factor in a quite complicated picture and probably a minor one at that. It so happens that predictive validities of g-loaded tests are actually somewhat higher in occupations that have a more restricted range of ability than in occupations with a very wide range of ability. (The reason for this seeming paradox is explained in Job Complexity and g Validity below.) When we analyze g-loaded test scores from persons in a very large number and extremely wide variety of occupations, we find that approximately one-half of the
    total variance in scores exists between the means of the various occupations and approximately one-half of the total variance exists within occupations (that is, individual differences among persons within any given occupation). From this empirical observation, it follows statistically that if
    we rank-order occupations so as to maximize the correlation between their ranks and their mean scores on g-loaded tests, the correlation between individuals’ test scores and their occupational ranks would be the square root of one-half, or approximately +.70. In other words, g predicts occupational status with a validity coefficient of +.70. And this degree of correlation is just what is actually found in studies in which many different occupations have been ranked, not on any psychometric criteria, but in terms of their prestige (in the eyes of the subjects who do the ranking), their desirability, and people’s subjective judgments of the amount of intelligence they think is required for successful performance in the occupation. (These three criteria, when based on the pooled ranks by a large number of persons, are amazingly consistent with one another and are highly stable throughout the industrialized world and from one decade to another.) It seems impossible to avoid the conclusion that what people ordinarily mean by “occupational status” is quite highly related to psychometric g.
    On the other hand, the observed correlations between g and measures of proficiency within given occupations are usually very far below +.70, even though one-half the total g variance exists within occupations. This means that, in general, g is much less able to predict occupational performance than occupational level. The main reason (aside from the forms of attenuation of the validity coefficient previously mentioned) is that, once employees are up to the minimum level of qualification for performing in a given occupation, a host of other factors independent of g becomes at least as important as g for successful job performance (or the perception of effectiveness by supervisors and co-workers). Most nominal occupational categories accommodate a surprisingly wide range of g above the minimum level, or threshold, for a given occupation. This threshold level can be estimated from the mean g-loaded test scores of persons in a given occupation whose scores are at the first percentile of the distribution of scores in that occupation. The g “threshold levels” across a wide variety of occupations vary considerably more than do the mean levels of g across occupations, and the very top levels of g across occupations show surprisingly little variation-only about one-seventh as much (in IQ units) as we see at the threshold level. Some very high-g persons are found in some very low-g occupations, but no very low-g persons are found in high-g occupations. (The evidence is reviewed by Jensen, 1980, 343-45.) This widely recognized threshold property of g, with respect to both education and occupations, is probably responsible in large part for people’s anxiety and antipathy concerning tests of mental ability.

    • Replies: @James Thompson
  255. Sparkon says:

    Is Chinese writing even harder than Chinese arithmetic?

    My pet theory is that learning the Han characters confers on the young Asian student several significant advantages. Chinese and Japanese kids spend a lot of time in their childhoods learning the Han characters, a task that requires both self-discipline and also a lot of long, hard study. Who would doubt that those are excellent work habits to develop at an early age that can only pay benefits for any student?

    中文写作比中国算术更难吗?

    The Google translation is accurate, but the pun doesn’t work because making puns in Chinese is definitely harder than Chinese arithmetic.

    • Replies: @keuril
  256. JLK says:
    @academic gossip

    I know that in several regression analyses of University of California grades where SAT M, V and W were included, the Writing section for some reason completely dominated in predictive power, with coefficients larger than M and V combined.

    I ran across a study a few days ago that had the correlation of the writing test to first year GPA around the low 0.6+ range, and the math and verbal in the high 0.5+. I won’t take the time to go back and find it, because you didn’t provide a citation for yours, and still owe me another one proving the PSAT-M has any appreciable spatial weighting.

    I wouldn’t put much faith in these correlations anyway without drilling deeply into the studies. Many of these strong writers aren’t STEM students who have to cope with freshman calculus.

    I don’t know much about the SAT writing section, but I took a brief look at the prep suggestions for it on the Kaplan website and noticed they stressed proper punctuation like putting periods and comments in the right places. If they really grade for that, it is low-hanging fruit that likely compresses the left side tail, almost certainly with the purpose of lifting Hispanics and blacks, just like the compression of the upper ends of the verbal and math suppress Jewish and Asian scores, respectively.

    The most predictive National Merit selection index for college grades might be Verbal+ Math + 4xWriting or something of that nature that is even more skewed toward the verbal test than what is done now.

    Your shamelessness is truly breathtaking.

  257. Anonymous[171] • Disclaimer says:
    @JLK

    It’s a travesty that the college admission testing system has to be put under the microscope by a group of dabblers on a free speech site

    Why? Why is it a travesty? Free speech is messy but the dabblers are exchanging information and even learning. Please try not to be full of shit while you’re waiting for the warm, familiar, sleepy hand of the NYT to serve it in a “proper” form.

    • Replies: @APilgrim
  258. @Polynices

    To me, NNT’s cluelessness on this topic is simply based on his insistence that high IQ should or must correlate with success. I don’t think it is unreasonable to assert that (maybe especially amongst high-IQ types) income\net worth isn’t an absolute measure of success. I’ve tested in the 99.5 percentile and personally don’t give two shits about making millions or advancing in my “career”. By not focusing all my time and effort on making money, climbing corporate ladders, and impressing people, I’m able to pursue my many varied interests and hobbies. I suspect there are many more important factors in explaining financial and career success than IQ.

    • Replies: @Anonymous
    , @Curle
  259. Anonymous[171] • Disclaimer says:
    @Alfred1860

    In my experience, money-making and IQ only correlates for middle classes where academic achievement can lead to a good career. Nowadays, most of the real money is inherited and/or reserved for well-connected individuals. Some bright people will create successful businesses on their own but in many cases the “real” money people will simply buy them off and absorb the innovation/success.

    I’ve also read somewhere that the brightest of the bright are notorious for not caring about finances. Not sure if it’s true but it makes sense that they’ll have “better things to think about”. Tesla would be a good example.

    • Replies: @mikesmith
  260. mikesmith says:
    @APilgrim

    Nassim *Nicholas* (hint, hint!) Taleb is a Christian, not Muslim. Even if I had not learned this while reading two books by him, the name is a dead giveaway. Muslims do not name their children after St. Nicholas. I note, however, that many people, regardless of their faith, become very testy about IQ scores if their particular group tends not to score so well, even if they themselves score quite high. See, e.g., Thomas Sowell, an African-American with a 160 IQ who just can’t accept that Blacks score lower than Whites for genetic reasons.

  261. mikesmith says:
    @Anonymous

    I think many more people than we realize are not strongly motivated by money, but we tend only to notice this fact when such people are also very smart. Thus the eternal question, “If you’re so smart, why aren’t you rich?” Our hunter-gatherer ancestors were nomadic and could not accumulate lots of possessions. A strong acquisitiveness must only have evolved later, but it still is not universal in the species.

  262. @Wizard of Oz

    Thanks. Yes, In Search of Excellence was a good signal for shorting stocks.

  263. @res

    Interesting points. I suppose the best advice to those wishing to understand intelligence is: read Spearman, then Jensen, then Deary.

    • Replies: @res
  264. res says:
    @Wizard of Oz

    Regarding “In Search of Excellence”, here is a laudatory 2002 Forbes article which includes a complete list of stocks and their performance to date: https://web.archive.org/web/20040217221744/https://www.forbes.com/2002/10/04/1004excellent_print.html

    And a Tom Peters 2006 blog post: https://tompeters.com/2006/11/in-search-of-excellence-at-almost-25-and-standing-tall/

    But this 2017 followup speaks the most directly to your point: https://www.marketwatch.com/story/great-companies-are-more-likely-to-do-really-badly-over-time-than-really-well-2017-07-12

    Their conclusions:

    Great companies were more likely to do really badly than really well.
    A few great companies is all it takes for a portfolio to outperform.
    Greatness is no guarantee of survival.

    Link to the 2004 article you mentioned: https://www.fastcompany.com/50992/was-built-last-built-last

  265. APilgrim says:
    @Anonymous

    Taxpayers FUND, or DEFUND education, as we have recently DEFUNDED the MSM & local papers.

    EFFING Deal with it.

    Or don’t.

  266. @res

    Thanks. A fascinating read. Berkshire Hathaway seems a better bet.

    • Replies: @res
  267. keuril says:
    @Sparkon

    My pet theory is that learning the Han characters confers on the young Asian student several significant advantages. Chinese and Japanese kids spend a lot of time in their childhoods learning the Han character

    But Korean kids don’t spend a lot of time on this (Korean nowadays is written 99%+ in the Korean phonetic script, with Chinese characters being about as critical as Roman numerals in English), and they do OK. Koreans might even argue that they do better than, e.g., the Japanese, because they have more time to study STEM subjects instead of memorizing Chinese characters. And the American-born children of Chinese and Korean immigrants do pretty well.

    • Agree: JLK
    • Replies: @res
    , @Sparkon
  268. res says:
    @James Thompson

    Interesting. Based on your followup post to this I would have guessed a good approach would be to read Ritchie’s all that matters short book, then your post on Deary’s review paper along with that paper. Then at some point (perhaps with intervening levels of detail?) Jensen’s g Factor.

    Spearman clearly did revolutionary work laying much of the groundwork, but would you recommend him as a starting point? Which pieces in particular? (I tried a quick search of your blog and did not see an obvious answer, maybe I missed it?)

    • Replies: @James Thompson
  269. res says:
    @James Thompson

    One of my investing regrets is not buying shares of Berkshire Hathaway when it was out of favor during the dot com boom. But this chart makes me wonder how good a bet it is going forward: https://www.theatlas.com/charts/S1mAAANEz

    Let’s try embedding that to save a click through.

    A notable thing about that chart is its behavior near corrections (roughly 1987, 1991, 2000, 2008). I wonder if there is a lesson to be learned there about buying it during tumultuous periods in the market.

    I wonder how long Warren Buffett and Charlie Munger can keep going. It is interesting that Buffett’s age seems less discussed now than it was in 2001

    P.S. I think it would be informative to recast that graph with a log y axis.

    • Replies: @James Thompson
  270. res says:
    @keuril

    Perhaps more important for selection during the evolution of IQ than as a cultural effect?

    • Replies: @Wizard of Oz
    , @keuril
  271. klokman says:
    @Anonymous

    Actually, no they don’t. As a military contractor employee in the 80s I attended the Air Force survival training program. Our job required us to visit remote radar sites on the Dugway proving grounds, where there was always a high risk of being left behind at 7-9K feet because of weather or military exigencies. It was initiated specifically because of the high fallout rate of downed pilots not surviving isolation. Their concern was the loss of the significant investment in pilots, as well as the scarcity of persons smart enough to get that far. The first thing these pilots who either died or came near to dying did was panic, making very stupid and irrational decisions. They had survival kits on their persons. They had training on how to use the gear. But the sudden isolation outside of the support environment that required social and technical intelligence did not aid them.

    Knowing how to save one’s ass in the cockpit did not necessarily carry over into knowing how to do so outside of it in isolation. The qualities that make for an accomplished fighter pilot does not necessarily carry over into ingenuity, creativity, and sensibility in the wild.

    • Replies: @Wizard of Oz
  272. @res

    Thank you for the links. I can’t achieve the genius status as an investor of one of my school friends whose father had bought a farm on the outskirts of a city of a million in 1931 which he sold three quarters of recently for $500 million when the city hit 5 million. So, I try to turn into “wise saws and modern instances” my noting that there are some almost natural features in modern times to a 2 or 3 per cent real return on investment for which I may claim just a little originality for basing it on noticing that, from the foundation of the Bank of England in 1694 to 1914 Consols returned that sort of interest, backed up by 4 per cent being the decent rate of return a big landowner could expect of his tenants and also a pretty standard residential first mortgage rate. After that your knowledge of and adjustment to tax law and rates is probably more likely to give you the modest 1 per cent outperformance compared with market timing or stock picking for equivalent time and effort.

  273. @klokman

    Very interesting but I think you should give him credit for his saying “other things being equal”. Admittedly it’s not quite the same as insisting on the key importance of probabilities and averages because, of course, the cet. par. condition may simply be irrelevant and never apply.

  274. @res

    “of IQ”?? Perhaps evolution of other productive characteristics would be more important?

    • Replies: @res
  275. keuril says:
    @res

    My remarks were specifically direct at the OP’s “pet theory” (“learning the Han characters confers on the young Asian student several significant advantages”). I’m not sure what the effects may have been over a longer (evolutionary) period of time. Perhaps the Chinese imperial examination system had some eugenic effects, but I’m not sure how you would tease out the effects of learning ideographic characters in particular.

    • Replies: @res
  276. @res

    I think in there there must be something about a test for Tiger Mother status 🙂

  277. res says:
    @Wizard of Oz

    Perhaps. Which ones did you have in mind? I mentioned IQ because it is a usual topic here and fairly measurable. Also because the connection of difficult literacy to IQ selection pressure seems plausible to me. Less sure about other things (e.g. conscientiousness?).

    My primary point was keuril’s comment seems like evidence that the contribution of Chinese characters to current good performance was more evolutionary (genetic, for whichever traits) than cultural. I assumed this was the implicit point of his comment.

  278. res says:
    @keuril

    Thanks for clarifying. I wrote comment 285 before seeing this one. Agreed about the Chinese imperial examination system. And I am also not sure about how to separate the effects. Though my guess would be that both were in play. Do you have any idea of the exam success/failure rate relative to the il/literacy rates? My guess would be the exams would have a stronger but less broadly distributed selection effect.

    • Replies: @keuril
  279. Sparkon says:
    @keuril

    Fair point, but per Wiki, the Koreans didn’t discontinue teaching Hanja until 1971, and even now, it is taught to some degree in both North and South. Even with pretty good machine translation, wouldn’t some Korean scholars and historians want to know Hanja to be able to read older documents, especially those written by hand, where the machines stumble? I don’t know, but my wild guess would be that officially Hanja has been deprecated if not eliminated, but in reality, probably not.

    In fact, my suspicions are confirmed, and Hanja are still used for scientific and legal writing in S. Korea according to this comment I found on Quora:

    There are two areas of Korean society, however, that resisted the trend in dumbing down the written language and continue to use Hanja characters heavily and unabated to this day. Those two fields are in scientific research and in the judiciary, both of which rely heavily on Hanja characters to communicate complex ideas whether it be research published in Korean-language scientific journals, or in judicial research where drafting the nation’s legal codes and laws is written almost exclusively using a lot of Hanja.

    https://www.quora.com/How-often-is-Hanja-still-used-today-in-South-Korea

    In any event, no young children first learning the Han characters or Roman characters either one are studying STEM subjects. Rather, they are laying down the groundwork and developing both framework and tools for how they think and learn that will apply to the specific fields they study when older.

    • Replies: @keuril
  280. @res

    You are right, I was making too obscure a point. James Flynn said that all psychometric publications in the late 20th century were footnotes to Jensen. I reported this comment with approval, and Deary remarked to me that all psychometric publications were footnotes to Spearman. Jensen would have agreed with that.

    So, to be clear, if someone now asks me where to start when finding out about psychometry, I would suggest Stuart Ritchie’s introduction.

  281. @res

    Agree that it doesn’t look so good on this graph. I was basing my remarks on the front page of his letter to shareholders, but that is based on book value.

    • Replies: @Wizard of Oz
  282. JLK says:
    @Ron Unz

    So just as I said, the PSAT is two parts Verbal to one part Math and lacks a direct Spatial component, therefore being ideally advantageous for Jews and disadvantageous for Asians.

    It’s disadvantageous for Native Americans as well because they share the East Asian intelligence profile to a large extent.

    https://velesova-sloboda.info/antrop/lynn-race-differences-in-intelligence.html#_12

    The median of the 21 studies is an IQ of 86. The Native Americans obtained higher visualization than verbal abilities in all of the six studies in which tests of the two abilities were given. The median visualization IQ in these studies is 89.5 and the median verbal IQ for the studies is 81. The same strong visualization-weak verbal profile of abilities is present among East Asians (see Chapter 10, Section 1) to whom the Native Americans are genetically closely related.

    A study by Lombardi (1970) compared 80 Native American with 80 white 6-8-year-olds tested with the Illinois Test of Psycholinguistic Abilities and found that the Native Americans obtained a verbal IQ of 73 and a visualization IQ of 93. The visualization IQ is constructed as the sum of six subtests of visualization abilities of which one is visual memory, and on this the Native Americans obtained an IQ of 104. This was the only subtest on which the Native Americans scored higher than the whites.

    Closer to your heart, it follows that the skew of the PSAT/SAT is also disadvantageous to Mexican-Americans, who are largely an admixture of European and Native American stock:

    Row 9 gives results from the standardization of the Stanford Binet 4 showing Hispanics with an IQ of 99 on non-verbal reasoning; this sample obtained an IQ of 93 on verbal reasoning.

  283. JLK says:
    @academic gossip

    Fantasies of visuospatial puzzles or a calculus section on the PSAT (or making everyone take it in Korean or Chinese) are amusing but fly in the face of the purposes of the test: predicting SAT scores, university grades/admission, and maybe performance after university (all using elementary questions tht plausibly relate to typical high school or college coursework). Accomplishing any of those would require a substantial weighting of the verbal and both the SAT and PSAT may underweight it relative to optimal predictors.

    There is near unanimity is the literature that the general intelligence factor, most often referred to as g, is the best single predictor of practically all aspects of academic and professional success. It follows that the PSAT/SAT should be testing for the skills that most closely correlate to g. It is very, very hard to find those correlations on the Internet, which is surprising because they have to be of immense interest to psychometricians and and anyone having an interest in the public policy of college admissions. I finally found something on point from Dr. Richard Haier’s book The Neuroscience of Intelligence:

    When you consider that the PSAT/SAT verbal portions are a hybrid of vocabulary (a relatively low 0.74 correlation to g) and reasoning (items like analogies; 0.96 correlation to g), the spatial correlation to g of 0.92, your suggestion that the already verbal-skewed PSAT/SAT underweights verbal as an optimal predictor is patently absurd.

    This diagram shows that the omission of spatial testing from the PSAT/SAT and other intelligence tests is unjustifiable. It is also inherently discriminatory against Asians, Native Americans, Mexican-Americans and probably a large number of white non-Jewish males, because it is clear that these groups do better on spatially-oriented tests than on verbal tests, even verbal reasoning tests.

    • Replies: @utu
    , @Hypnotoad666
  284. Wally says:
    @Wizard of Oz

    said:
    “You set me off Wally. I Googled for Zyklon B from which I learned that it was a very efficient killer of people as well as being able to be explained away as a fumigation or deluding agent.”

    [MORE]

    LOL. Try reading real science for a change, not Zionist propaganda.
    – BTW, there were already large store of Tabun & Sarin, there would have been no need to use Zyklon-B which was not cheap, nor safe. But they this is the ‘holocau$t’ scam where facts don’t matter.
    – Zyklon-B takes hours to outgas it’s cyanide load, not exactly efficient or safe. Not to mention it’s highly explosive characteristic when used in the large amounts claimed in the fake & impossible story. Note that intense heat emitting crematorium, allegedly just above the ‘gas chambers’, would find volatile cyanide very attractive.

    – Now have a laugh at the alleged gassing process Zyklon-B.
    The alleged Auschwitz homicidal gassing process reviewd and demolished here.
    https://forum.codoh.com/viewtopic.php?f=2&t=11143&p=83723&hilit=model+asmarques#p83723

    and: The Rudolf Report / Expert Report on Chemical and Technical Aspects of the ‘Gas Chambers’ of Auschwitz
    http://vho.org/GB/Books/trr/index.html
    – Your cited “witnesses” mention events which never occurred, were scientifically impossible, and of course the use of Allied torture can be persuasive. I note you are afraid to mention about what they actually claim. But here’s help for you:
    https://codoh.com/search/?sorting=relevance&q=pery+broad
    https://codoh.com/search/?sorting=relevance&q=Johannes+Kremer
    https://codoh.com/search/?sorting=relevance&q=Rudolf+hoess

    And the there is the complete lack of millions of human remains that Jews claim still exist in known locations, but are not there.
    See more in my comments here: Pravda: Holocaust Denial, by Ron Unz: http://www.unz.com/runz/american-pravda-holocaust-denial/

    You’re in way over your head, desperately so.
    http://www.codoh.com

  285. keuril says:
    @Sparkon

    The thing about Korean is that about half of its vocabulary comes from Chinese, especially vocabulary for abstract things. The problem is, Chinese is a tonal language whereas Korean isn’t, so when all these words were imported to Korean, the tones got lost, resulting in a huge number of homonyms. In Chinese, you might be able to distinguish “ma” four or five ways depending on the tone, and it’d be obvious based on the tone whether you meant mother, hemp, scold, horse, or a question. But in Korean you just have a bunch of “ma”. Of course, some of these you could change to “ba” or “mu” or something, but those will get mixed in with other words that were originally “ba” or “mu,” so it doesn’t really solve the ambiguity problem. This isn’t much of an issue in conversation, since context tends to be clearer (and topics, perhaps, simpler), but in writing, there are ambiguities. So even today, in some specialized text they might include the Chinese characters for a word as a sort of disambiguation the first time it is used. For example, I just checked and there’s are Korean wiki entries for “Court of Law” and “Sources of Law” with the exact same Korean pronunciation 법원, that are disambiguated in the introduction based on the difference Chinese characters (法院 and 法源). But Chinese characters are no longer routinely used throughout texts as they are in Chinese (of course) and Japanese. And even the most fervent Korean nationalist would, I think, acknowledge that Koreans’ knowledge of the Chinese characters pales in comparison to that of the Chinese and Japanese.

    Written Korean didn’t used to be like this—they used to use a mixture of Chinese characters and native (Korean) phonetic script, as is still done in Japanese (thank God!). But starting in the 60s and into the 80s, the Chinese characters were done away with in common Korean texts like books and newspapers. Reading Korean is a bit like what reading English would be like if everything were in phonics. If they did it overnight, there would be a rebellion. But if you introduced the change gradually over the years, slowly simplifying the texts and slowly reducing the importance of traditional spellings/characters in education, people would just get used to it. That’s not to say that Koreans have any problem reading their own language, and this is an indication to me that phonics would work just fine in English, after the older generations die off.

    Of course they still do teach Chinese characters in school, or after school in cram schools (hagwons), but the exposure a typical Korean gets to Chinese characters is minuscule compared to what Japanese and Chinese both growing up, and as adults. So getting back to your pet theory, I don’t think exposure to Chinese characters is responsible for Koreans’ strong academic performance.

    • Replies: @Sparkon
  286. utu says:
    @JLK

    How was the diagram from Richard Haier’s book generated? Did they have 15 different tests and they applied them to many (large N) subjects and then get the covariance matrix from which they get the g factor via factorial analysis? And did they apply the tests to Whites, Asians, Jews and Blacks and Native Americans? Are the loadings noted in the diagram valid for all races? Or is the g valid for all races? If you had two covariance matrices where one was for whites and one for Jews would you get the same g from both matrices?

    • Replies: @JLK
  287. JLK says:
    @utu

    How was the diagram from Richard Haier’s book generated?

    I don’t know. I don’t have the whole book.

    Are the loadings noted in the diagram valid for all races? Or is the g valid for all races? If you had two covariance matrices where one was for whites and one for Jews would you get the same g from both matrices?

    I was just wondering the same things as I fixed dinner. Maybe Mr. Thompson or another expert could share his thoughts on this.

    • Replies: @res
  288. keuril says:
    @res

    Sorry, I have no clue on those matters.

  289. Antiwar7 says:
    @res

    Thanks for your interest! I believe I was included in SMPY in 1980, which would put me in cohort 3. I’m sorry, I don’t know what percentile my abilities landed in. I seem to recall that my SAT scores were 700-M, 650-V when I was 12, and 750-M, 750-V when I was 14 (in the early 80’s). I took some sort of IQ test through the public schools when I was 11. I believe the results were: a mental age of 20.5 and an IQ > 160. (Does that even make sense?) I believe they’re trying to get as much participation as possible from all of their cohorts; otherwise, one could reason that the ones that readily respond would have disproportionately successful outcomes, skewing the results.

    Yes, I was part of the genetic side of Plomin’s research. First, a nurse was sent to my home to draw blood from me. A few years later (when the technology improved), my parents were asked to provide cheek swabs.

    • Replies: @res
  290. res says:
    @Antiwar7

    Thanks for responding! Cohort 3 was the 0.01% group and your SAT scores align with that (impressive! you made both M and V thresholds). Table 1 of this paper gives a concise description of the cohorts: https://my.vanderbilt.edu/smpy/files/2013/01/DoingPsychScience2006.pdf
    This won’t format well, but might be readable.

    Cohort | N | Years when identified | Students’ age when identified | Identification criteria
    1 2,188 1972–1974 12–13 SAT-M 390 or SAT-V 370 (top 1 in 100)
    2 778 1976–1979 12 SAT-M 500 or SAT-V 430 (top 1 in 200)
    3 501 1980–1983 12 SAT-M 700 or SAT-V 630 (top 1 in 10,000)
    4 1,130 1992–1997 12–14 Top 3% on any subtest of a grade-level achievement test
    5 714 1992 23–25 Enrollment as a 1st- or 2nd-year student in a graduate program
    at a top-ranked engineering, math, or science department in
    the United States

    I know someone who took the SAT for that cohort but did not qualify (probably would have made the first two). The effective sample size (~5e6 = 501 people at 1 in 10,000 level) is impressive.

    IQ > 160 makes sense for your numbers. I’m not sure how the mental age translates.

    Any idea how Plomin’s research is using the genetic information from your parents? I wasn’t aware of that aspect of it. Did they IQ test your parents as well?

    One thing about the SMPY, when looking at their plots remember that grouping all of the cohorts together and looking at them by quartiles gives a very unrepresentative top quartile. Because cohort 3 has a threshold so much higher than the others the top quartile is significantly smarter than in a purely representative sample. More concretely, looking at the first three cohorts the first is by far the largest and will fill the first two quartiles with roughly 1%-0.5% group. The third quartile will be a mix of cohorts 1 and 2 at roughly 0.5%-0.25% while the top quartile will contain the 0.25% and up people from the first two cohorts along with a spiking of 0.01% people (501, where the first two cohorts probably only have ~40 at that level).

    • Replies: @Antiwar7
  291. res says:
    @JLK

    Are the loadings noted in the diagram valid for all races? Or is the g valid for all races? If you had two covariance matrices where one was for whites and one for Jews would you get the same g from both matrices?

    I was just wondering the same things as I fixed dinner. Maybe Mr. Thompson or another expert could share his thoughts on this.

    I am very interested in this as well. It seems obvious to me that groups with significantly different score profiles should have different principal components, but data trumps “obviousness.”

    The most likely place to get good data would be a European/Asian comparison. Has anyone done that?

  292. Antiwar7 says:
    @res

    Forgot to mention: no, I wasn’t aware of Steve Hsu’s blog posts on the subject. Thanks for pointing them out to me!

  293. Antiwar7 says:
    @res

    They basically presented us with lots of information about various educational options, their research so far, and articles about individuals and their experiences. It was up to my family and the local school officials to decide what to do with that. (One of the options was to simply take an advanced class or two at a time, perhaps in one’s current school or visiting at another for part of a day, which perhaps wouldn’t be too disruptive.) My local school district gave me carte blanche, and I was assigned a fantastic (I’m not sure what his position was, but basically an at-large gifted) counselor who could help pull many strings. It so happened I was quite large for my age and appeared older, so going to college early wasn’t too bad. (However, a couple of local newspaper articles about me made me feel quite exposed and shy for the first couple of years.) I wasn’t much into organized sports, though at college I played on a dorm soccer team. My size helped there.

    Yes, overall, they helped me a lot. Also, attending and later teaching at some of their and related summer programs which they introduced me to (I attended TIP and taught at CTY and MTS) brought me into contact with similar people, which was quite enjoyable and led to many good friendships.

    • Replies: @res
  294. Anonymous [AKA "Arne_K"] says:
    @Johnny Horton

    Do Afrikaners or early Australian pioneers not qualify?

  295. @James Thompson

    This thread seems as good a place as any to raise a question that came from watching a bit of a TV program that was on when I came home from a swim (it’s midsummer here). It dealt with human maturing from an early age and about the 6 year old stage it mentioned and gave some detail of synaptic pruning. This was described as part of the body’s efficient way of saving energy on a very energy hungry part of the body. It seemed to suggest that constantly playing musical instruments to the age of 6 or whatever would build one class of synapses (and neurons???) while letting others go.

    This leads me to two questions
    1. Has their been research on the impact of differential synaptic pruning at various ages including full maturity?
    2. To what extent can synaptic pruning be compensated for and reversed, and at what ages? As we have known for at least 30 years that there are stem cells in the brain there seems to be reason for saying that even “use it or lose it” isn’t the last word.

    • Replies: @James Thompson
  296. Sparkon says:
    @keuril

    but the exposure a typical Korean gets to Chinese characters is minuscule compared to what Japanese and Chinese both growing up, and as adults.

    Thanks for your comment. I knew that Hanja had been deprecated, but I wasn’t sure to what extent Hanja is still in regular use in Korea, so I intentionally did not cite Korea in my original comment about my pet theory. However, the comment from Hyun Kim I linked earlier from Quora, when read in its entirety, gives a rather different picture from what you paint, and attests to the fact that Hanja is in current use in the scientific and legal communities, or do you dispute that entirely?

    Whatever the case, the average Korean’s exposure to the Han characters and familiarity with them is much greater, perhaps orders of magnitude greater, than virtually any Occidental’s exposure to and knowledge of Hanzi, Kanji, or Hanja during the formative years, which leads to such hilarities as the “harder than Chinese arithmetic” expression.

    There is another aspect to my pet theory, which has to do with the hemispheres of the brain, where the left side is thought to be more involved in language, and the right side more involved with processing imagery. I know this is a perhaps entirely too simplistic, but I throw it out here as a starting or reference point.

    Although it may be true that, strictly speaking, the Han characters are not purely pictorial or hieroglyphic, I know from my own experience with Japanese Kanji that “reading” the Han characters is a different process than reading English, where even unknown words may be “sounded out” phonetically and a knowledgeable reader may be able to take a stab at its meaning from the stem or root of the word, especially if that reader is multilingual and knows for example Latin, French, and/or Spanish.

    By contrast, with Han characters, either you know the character or you don’t, although again it is true the the Han characters are built up with various usually familiar radicals (themselves based on simpler characters) and the knowledgeable person might be able to take a stab at the meaning, the fact remains that Hanzi are more cryptic to deconstruct than are English words, and some fairly significant measure of memorization is required to master Han characters.

    The great number of homonyms you mention would seem to be a strong argument for retention or reinstitution of Hanja in Korea. I’ll close with another quotation from Hyun Kim at Quora:

    This was the beginning of an extremely ignorant, shortsighted and ridiculous attempt to increase Korea’s literacy rates by making the written language allegedly more egalitarian by removing Hanja altogether, which for centuries was the real distinguisher between the educated and uneducated classes in Korea. By trying to increase Korea’s literacy in this way, such policies inadvertently increased Korea’s functional illiteracy rates. By dumbing down the language and rendering the public education system almost meaningless, Chun Doo-hwan and his henchmen did irreparable harm to Korea’s potential for intellectual and social development in the future.

    • Replies: @keuril
  297. Antiwar7 says:
    @res

    Good point re: the quartiles. And thanks for the info on the cohorts. I had assumed they were using the same criteria all along, and had kept accepting new members for longer. I was lucky to be born at the right time to fit into that cohort.

  298. @Wizard of Oz

    I see that I have made only one relevant mention of pruning, as shown below:

    http://www.unz.com/jthompson/the-well-tempered-clavichord/?highlight=pruning

    I have not got into the developmental issues surrounding this, just what looks like the finished product as shown above.

    It seems likely that skills which are established early and practiced in childhood would have an advantage. Language and playing a musical instrument come to mind.

    Interestingly, these sorts of issues were known in the 1960s, but were usually described as “sparing” in the neuro-psychological literature. Rats over-trained in particular skills such as maze running lost all their skills after the ablation of the relevant areas in their brains, but re-learnt the tasks much faster than those just trained to criterion, and then ablated. In my PhD I speculated that the skills had been stored in various parts of the brain, in what would now be called a distributed network.

    What is fascinating about all this brain remodelling, particularly in early childhood, is that kids do not lose their memories while it is happening, but tend to have lost the first four years of memory once the process has been completed. Clever.

    • Replies: @Wizard of Oz
  299. Anon[306] • Disclaimer says:
    @RaceRealist88

    The fact that intelligence test items are not subjects of will of a single individual, that is, test items are the result of multiple peoples’ involvement, consultation, testing for relevance, etc, means that they are not arbitrary.

    Intelligence test items are not arbitrary as they are not dictates of an individual will, capricious or otherwise.

    And there are psycho physical laws. Some mental states are demonstrably reducible to physical quantifiable parameters. Come to my lab and I will show you.

  300. res says:
    @Antiwar7

    Thanks! I am interested in the question of acceleration (especially since it seems to be becoming less popular?) and your account is helpful. I had a conversation with someone who was thinking about having his son skip a grade a while ago and my thoughts came down to (assuming the intellectual ability is clearly there):
    – How mature is he in other ways?
    – How good a cohort of friends does he have in his current grade? (if the friendships are short duration the comparison to what is likely a year ahead is what matters)
    – Does he want to play competitive sports? Especially at a high level (research shows even birthdate within the year matters for later outcomes).
    If anyone has more practical advice and/or experience regarding this topic I would be interested.

  301. Pat Boyle says:
    @dearieme

    Dear ‘deariame’ – you are wrong this time. I have taken a lot of standardized tests and done well on them. I have, like you, observed that many of the people who refuse to accept the evidence of the utility of IQ and IQ-like tests didn’t do very well on these kind of tests. In fact since I was involved in computer skill testing for many years. I’ve seen a lot of the similar criticism from students who deprecate the values of the Microsoft or Novell certification tests. Almost invariably you would find that they were one of those poor schmucks who took a test over and over and could never manage a passing score. They were often employed in data processing or networking in some capacity and they preached that the “real” required “hands-on” experience. They seem to think with their fingers.

    But this isn’t the case with Taleb. Taleb is ferociously intelligent. He is quite wrong about IQ tests but not because he’s stupid.

    You have to understand his profession. Taleb contrary to the beliefs of many is not a market trader. He did work at that for a while but according to Malcomb Gladwell was a miserable failure. Thais is to say – he was miserable. So he switched to the guru business. He has thrived and prospered in this role.

    A guru must be self confident. Taleb has that well in hand – he’s one of the most arrogant people alive. They also need to make periodic pronouncements that promote their specialness.

    So the silly statements that Taleb has made are just marketing.

    • Agree: Dieter Kief
    • Replies: @keuril
  302. keuril says:
    @Sparkon

    However, the comment from Hyun Kim I linked earlier from Quora, when read in its entirety, gives a rather different picture from what you paint, and attests to the fact that Hanja is in current use in the scientific and legal communities, or do you dispute that entirely?

    What I meant is that the Chinese characters, or Hanja (as they are called in Korean), would be used in such contexts for *disambiguation*—a bit like the way botanists or zoologists might use the Latin name of a species as opposed to its vernacular name, in order to be as precise as possible. But this doesn’t mean zoologists and botanists actually speak or read Latin.

    know from my own experience with Japanese Kanji that “reading” the Han characters is a different process than reading English

    Yes, there have been studies of Japanese aphasics (people who have lost certain language functions, typically due to a stroke damaging part of the brain) showing that, depending on which parts of the brain are damaged, the subjects may retain the ability to, e.g., recognize Kanji (for those trying to follow this discussion, Kanji is the Japanese word for Chinese characters, i.e., the same thing that Koreans call Hanja) but not pronounce them (such as by looking at the character for tree or sun or water and pointing to the corresponding correct photograph) and also not be able to pronounce the phonetic syllabaries (called Hiragana and Katakana); or may be able to pronounce the phonetic syllabaries but not recognize any Kanji. This strongly suggests that ideographs like Kanji are processed by the brain differently from phonetic writing systems.

    By contrast, with Han characters, either you know the character or you don’t, although again it is true the the Han characters are built up with various usually familiar radicals (themselves based on simpler characters) and the knowledgeable person might be able to take a stab at the meaning

    A lot of the characters have a meaning radical and a sound radical in them. Usually it’s a little easier to guess in Chinese than in Japanese, because the sound radical is more straightforward. For example, the Chinese word for ant, ma yi, is composed of two characters 螞蟻. Both characters have the insect radical 虫 on the left, and sound radical on the right (馬 ma, and 義 yi). In Japanese, the word for ant is ari, and the second of these two Chinese characters (蟻) is used to represent it. In this case, the insect radical is just as apparent to the Japanese reader, but the sound radical 義 (pronounced gi in Japanese) is of no use in figuring out how to pronounce the character in Japanese. So reading Chinese characters in Japanese is actually harder than in Chinese—perhaps not surprising considering that the characters were designed for Chinese, and are jury-rigged to the Japanese writing system (but have been a part of it for so long that the writing system would be completely different without the Kanji, just as English writing would be really different if we used phonics).

    The great number of homonyms you mention would seem to be a strong argument for retention or reinstitution of Hanja in Korea.

    You won’t get any objection from me on that point, but I don’t think it’s likely to happen. Widespread use of the Hanja was pretty much eliminated by the 1980s, so more than an entire generation has grown up without using Hanja as an important part of the writing system. Given the performance of the Korean economy, they seem to get along fine without it, even if some scholars and purists (or people who know Japanese or Chinese) might wish things were otherwise. Getting rid of the Hanja was probably a little more feasible for Korean than for Japanese because Korean has a bit richer sound system than Japanese, allowing for a bit more phonetic disambiguation in Chinese loan words (neither Korean nor Japanese is tonal, so they both face the same fundamental problem, but it’s a bit worse in Japanese).

    This was the beginning of an extremely ignorant, shortsighted and ridiculous attempt to increase Korea’s literacy rates by making the written language allegedly more egalitarian by removing Hanja altogether

    This sounds like someone with an ax to grind (and I even say this as somebody who really likes the Hanja and would be happy if Koreans were to use them). I know a lot of Koreans, and I can’t say I’ve ever heard one complain about their writing system this way. While the conversion to a Hanja-free writing system may have been billed at the time as a way to increase literacy, Koreans are in fact very proud of their native script (Hangul), and in fact they have a national holiday for it (Hangul Day). I must say, from a linguistics perspective it is a really well designed script—it just fits together very logically, and it’s pretty amazing that they came up with it back in the 15th century. Korea has spent a lot of time under the control of the Japanese and Chinese, so having their own unique script (which just happens to be designed really, really well) is perhaps an expression of national pride.

  303. @JLK

    Very illuminating graphic. But what is being represented by the boxes numbered 1-15 at the bottom with separate correlations to the sub-categories of g?

    Are those the correlations for categories of question types?

    • Replies: @JLK
  304. keuril says:
    @Pat Boyle

    You have to understand his profession. Taleb contrary to the beliefs of many is not a market trader

    He is underrated as a comedian. He has the best rude schtick this side of the “I fart in your general direction” French guy in Monty Python.

  305. @Anon

    In youth I took tutorial classes in, amongst other legal subjects, contract law, and it struck me that the 19 year olds ought to have done other degrees first and grown up a bit as they didn’t really have much of a grasp of why people made contracts.

    I think you have put your finger on one of the central problems of our whole education paradigm. To really learn something you need some context to cross reference it with in your brain. Until young people have a basic core of knowledge and experience they simply can’t make sense of much of what passes for “education.”

    Adults with life experience would profit more from education – but the opportunity cost is far higher because they are (generally) gainfully employed and would have to forego paying work to put time into school.

    Nevertheless, we spend all our resources educating young people. Not necessarily because they need it or can profitably absorb it, but because they are otherwise worthless to society and have nothing better to do anyway. (The core mission of schools is really to get kids off the streets during daylight hours, if they learn something in the meantime that’s a bonus.)

    The say “youth is wasted on the young,” but the same is probably true of education.

    • Agree: keuril
    • Replies: @Anon
  306. JLK says:
    @Hypnotoad666

    Very illuminating graphic.

    It’s a smoking gun. I remember being puzzled back in the early ’80s, around when I took the PSAT/NMSQT, as to why they doubled-weighted the verbal. There was no Internet back then. The best I could find was one sentence in the fine print of the test brochure stating that verbal was “more important.” I remember reading several print articles in the ’80s that flat out stated that vocabulary is the best single indicator of intelligence. This graphic shows that was a gross misrepresentation. Even on this site, much is made of the Wordsum test, such as the data from the GSS survey that contains very dubious information, at least in some cases. Wordsum’s 0.71 correlation to g is a pretty good indication that Haier’s 0.74 number for the correlation of vocabulary to g isn’t far off. If it has more than ten questions, it would probably approach 0.74.

    Cui bono? Certainly Jewish test takers, but I think it can be traced back to the biases of the monied WASP elite, when gentlemen didn’t do STEM, and public policy was controlled by white-shoe lawyers and other academics who were trained in the classics and eschewed math and hard science.

    I can remember pre-law types in college looking down their noses at the engineering students, with an attitude that they were all destined to be green-eyeshade types to be kept in back rooms out of public view. I don’t think Harvard has an engineering school, but most of the other Ivies and comparable elite private schools do. It is a little known fact, at least for a few of them that I saw numbers on, that the engineers not only had higher combined SAT scores than the liberal arts students, but even higher SAT-V scores.

    If we’re keeping score on who benefits and who is hurt from the verbal-skewed PSAT/SAT, it would look something like this:

    Groups That Benefit From the Verbal Double-Weight/No Spatial on the PSAT/SAT

    1. Pre-law/PoliSci type students
    2. Jews
    3. Women
    4. Blacks

    Groups That Are Hurt From the Verbal Double-Weight/No Spatial on the PSAT/SAT

    1. STEM-type students
    2. Asians
    3. Native Americans
    3. Mexican-Americans
    4. Men

    There was a post from a while back to the effect that various “human rights groups” monitor this site, but I wouldn’t hold my breath waiting for the ADL or SPLC to pick up the torch for the Native Americans or Mexicans that they claim to care so much about on this issue.

    There’s a problem when far-reaching policy decisions that affect so many people are made in smoke-filled rooms, with no scrutiny from the media. Some groups are more likely to have their people in the room than others. Here’s a blog post with some very interesting time-lapse graphics on how SAT scores of various demographic groups have change since the 1980s:

    http://theunsilencedscience.blogspot.com/2012/04/racial-amplitudes-of-scholastic.html

    It’s clear that they were tinkering with the SAT long before the 1995 re-centering and the later addition of the writing test. It’s a bad idea to politicize the tests, because when you try to life one group, others are dragged along one way or another as a result. Doctored test results are also used as evidence that certain social policies are working better than they really are.

    There are national security implications as well. Test scores are used to recruit and evaluate talented analysts, either directly or indirectly. Probably directly, since they have a handy list of all of the 140+ IQ students from high school each year courtesy of the NMSQT. Spatial intelligence is associated with being able to patterns and associations in data, something critical for this type of work.

    But what is being represented by the boxes numbered 1-15 at the bottom with separate correlations to the sub-categories of g? Are those the correlations for categories of question types?

    We would have to read the book to find out, but they are probably different tests of the same category of intelligence. For example, to test spatial intelligence, you might have 3D rotation, 2D rotation and pattern recognition.

    Haier’s diagram shows the primary abilities most closely correlating to g to be reasoning, spatial ability, memory, processing speed and vocabulary. If you used this list to plan the fairest possible PSAT/SAT, you would eliminate memory (long term memory too dependent on environmental factors; short term memory not testable on a paper test) and processing speed (chronometric, not psychometric). Of the remaining three, vocabulary should be substantially underweighted, not overweighted, in comparison to reasoning and spatial, because of its lower correlation to g. When you consider that many of the reasoning questions (such as the word analogies) are themselves vocabulary-intensive, a separate vocabulary test should probably be dropped altogether.

    • Replies: @res
    , @utu
  307. res says:
    @JLK

    Excellent comment. I dug into that graphic a bit more and thought it might be helpful to summarize what I found here.

    First, some discussion of the graphic online: https://psychology.stackexchange.com/questions/19481/how-is-the-g-factor-of-intelligence-calculated

    The graphic appears on page 25 of Haier’s book where it notes: Adapted from Deary et al. (2010)
    I did not see more detail on the subtests in the book (but did not look that hard).

    Following the reference to
    Deary, I. J., Penke, L. & Johnson, W. (2010). The neuroscience of human intelligence
    differences. Nature Reviews Neuroscience, 11, 201–211.
    Which has full text available at http://larspenke.eu/pdfs/Deary_Penke_Johnson_2010_-_Neuroscience_of_intelligence_review.pdf
    You can also view it at https://www.researchgate.net/publication/41420018_The_neuroscience_of_human_intelligence_differences

    We see a similar figure in Figure 1 with slightly different numbers. The big differences I see are a very different g-Memory correlation and an additional vocabulary subtest. Panel b of that figure gives age effects for each of the sub-categories (which is interesting).

    The caption of the figure indicates it is based on data from
    Salthouse, T. A. Localizing age-related individual differences in a hierarchical structure.
    Intelligence 32, 541–561 (2004)
    Which has full text available at http://faculty.virginia.edu/cogage/publications2/Localizing%20age%20related%20individual%20differences.pdf

    I believe Table 1 is a listing of the 16 subtests. I don’t see a mapping to the numbers Deary and Haier used, but assuming they are in order within categories seems like a reasonable first cut (can probably improve by checking correlations). Cut and paste formatting was terrible, but here as an attempt it reproduce with categories and subtest names.

    Table 1
    Description of cognitive variables included in the analyses

    Vocabulary
    WAIS vocabulary – Provide definitions of words Wechsler (1997a)
    Picture vocabulary – Name the pictured object Woodcock and Johnson (1989)
    Antonym vocabulary – Select the best antonym of the target word Salthouse (1993a, 1993b, 1993c)
    Synonym vocabulary – Select the best synonym of the target word Salthouse (1993a, 1993b, 1993c)

    Reasoning
    Ravens – Determine which pattern best completes the missing cell in a matrix Raven (1962)
    Shipley abstraction – Determine the words or numbers that are the best continuation of a sequence Zachary (1986)
    Letter sets – Identify which of five groups of letters is different from the others Ekstrom et al. (1976)

    Spatial visualization
    Spatial relations – Determine the correspondence between a 3-D figure and alternative 2-D figures Bennett, Seashore and Wesman (1997)
    Paper folding – Determine the pattern of holes that would result from a sequence of folds and a punch through folded paper Ekstrom et al. (1976)
    Form boards – Determine which combinations of shapes are needed to fill a larger shape Ekstrom et al. (1976)

    Episodic memory
    Logical memory – Number of idea units recalled across three stories Wechsler (1997b)
    Free recall – Number of words recalled across trials 1 to 4 of a word list Wechsler (1997b)
    Paired associates – Number of response terms recalled when presented with a stimulus term Salthouse, Fristoe and Rhee (1996a), Salthouse, Hambrick, Lukas and Dell (1996b), Salthouse, Hancock, Meinz and Hambrick (1996c)

    Speed
    Digit symbol – Use a code table to write the correct symbol below each digit Wechsler (1997a)
    Letter comparison – Same/different comparison of pairs of letter strings Salthouse and Babcock (1991)
    Pattern comparison – Same/different comparison of pairs of line patterns Salthouse and Babcock (1991)

    Table 3 has means, SDs, and factor correlations for each of the subtests.

    Table 5 has factor-factor correlations. Reasoning-space is by far the highest at 0.9

    • Replies: @JLK
  308. Anon[413] • Disclaimer says:
    @Hypnotoad666

    Thank you for your interesting and I think valid extrapolation of the point made anecdotally.

  309. Curle says:
    @Alfred1860

    Right. The fact that so many smart guys who are financially successful end up sidelining their businesses to pursue hobbies is a testament to the pull of such things. I can’t help but think of the numbers of such folks who make girls (John McAfee), sports and music (Paul Allen) or art and girls (Getty), their chief interest post money pile. Why then not start out pursuing such things?

    From the wiki on Chris Langan, man with highest reported IQ:

    “Langan took a string of labor-intensive jobs for some time, and by his mid-40s had been a construction worker, cowboy, Forest Service Ranger, farmhand, and, for over twenty years, a bouncer on Long Island.”

    “Langan was also approached and contracted by Disney Research[11] and he previously worked for Virtual Logistix, a technology company.[3]”

    “Langan said he developed a “double-life strategy”: on one side a regular guy, doing his job and exchanging pleasantries, and on the other side coming home to perform equations in his head, working in isolation on his Cognitive-Theoretic Model of the Universe (CTMU).[4]”

    In other words, he arranged his life so he could do what he wanted to do.

    The funniest is McAfee who set about creating a ‘socially engineered system’ in Balize that amounted to a harem. http://www.whoismcafee.com/the-girls/

    • Replies: @DFH
  310. @James Thompson

    Your last par leaves it open for ideologues to argue for all those repressed memories that explain the unfortunate adult, or for someone to argue that we have been shorn of the misguided ideas and values that are all a 0-4 year old brain could acquire and therefore one should be suspicious of claiming that bad parents or carers doomed their characters….

  311. utu says:
    @JLK

    The diagram you have posted already twice as the commenter ‘res’ shows is a result of compilation of data from several independent papers. There is no paper that all 15 tests were compared together and one g based on them was extracted. The correlations in the diagram are not with the same g. The g’s are different. The factor g is battery test dependent.

    This is not the first time authors in the field of IQism overstate their case and misinform. The diagram from Deary is there to indoctrinate that the concept of g is valid. You are a perfect example as a victim of this indoctrination. While you have another axe to grind and I even may support your cause I doubt that you will get anywhere with it by adopting arguments from the para-scientific field.

    • Replies: @JLK
  312. JLK says:
    @res

    Thanks for this information. It should be helpful to other researchers and maybe even a journalist in the unlikely event that one would ever dare to write such a story, or have an editor that would approve it.

  313. JLK says:
    @utu

    The diagram you have posted already twice as the commenter ‘res’ shows is a result of compilation of data from several independent papers

    As res pointed out and as Haier acknowledged in his book, the diagram and correlations are from the 2010 Dearing paper, not multiple papers. Dearing analyzed 33 studies from the same researcher, Salthouse, testing over 7000 subjects 18-95 years of age. Haier’s correlations differ slightly from Dearing’s, but not by much. It is unclear why. Perhaps he consulted with Dearing when writing his book and was provided revised correlations. In any event, it is clear that vocabulary tests are a poor proxy for intelligence in comparison to reasoning and visual-spatial tests.

    The correlations in the diagram are not with the same g. The g’s are different. The factor g is battery test dependent.

    The literature is near unanimous that there is only one g. If factor analysis is done on one small study, the calculated g might differ from the actual g, but usually not by much, and the results tend to converge on the actual g as sample size increases. Another comment raised the question of whether gs might vary between ethnic groups. It is a good question, and I don’t know. I suspect not, because it is an obvious question to raise for those who contend that IQ tests are inherently racially discriminatory, and I haven’t seen it raised. Disparate racial performance is probably somewhere in the residual of various tests after controlling for g. That’s another reason that the fairest, most objective tests are the most g-loaded. The g-loading of the SAT has markedly decreased over the past several decades.

    This is not the first time authors in the field of IQism overstate their case and misinform.

    IQism was the very basis of SAT testing from the mid-1940s. You’re accusing Deary and Haier of misinforming, after decades of brainwashing from the media that vocabulary is the best proxy for intelligence, and eighty years of college admission tests that overweight on it?

    While you have another axe to grind and I even may support your cause

    Which axe do you believe that to be?

    I doubt that you will get anywhere with it by adopting arguments from the para-scientific field.

    There’s nothing para-scientific about a field that has been used to screen college applicants and military personnel for over eighty years, but there do appear to be blind spots and taboos that researchers are reluctant to traverse. It may even go beyond that. If I was on the Asian legal team suing Harvard, I’d subpoena Google to see if they are suppressing search results on such topics. I have deep suspicions after a few weeks of delving into the topic.

    • Replies: @res
  314. Antiwar7 says:

    Summary of the Taleb-IQ dispute:

    1) IQ is well-correlated on both the deficient and successful sides of performance at school and at jobs.

    2) Taleb seems to respect only entrepreneurs and gentleman scientists (aka independent scientists).

    Clearly 2) is rarer and perhaps contingent upon a wider variety of circumstances than intelligence alone. And perhaps that’s why Taleb is so contemptuous of the concept of IQ, since he has little respect for the the accomplishments in 1).

  315. DFH says:
    @Curle

    The funniest is McAfee who set about creating a ‘socially engineered system’ in Balize that amounted to a harem

    Harems had attractive women in them

  316. res says:
    @JLK

    Another comment raised the question of whether gs might vary between ethnic groups. It is a good question, and I don’t know. I suspect not, because it is an obvious question to raise for those who contend that IQ tests are inherently racially discriminatory, and I haven’t seen it raised. Disparate racial performance is probably somewhere in the residual of various tests after controlling for g.

    This is an interesting passage. I superficially disagree with part of it, but wonder if we might be in agreement after a more detailed consideration.

    I think the following propositions are true.

    – g is a meaningful concept and probably has a physiological basis. Perhaps based on overall brain functionality where all areas probably depend on some common physical characteristics (likely both genetic and environment influenced) such as myelin formation, synapse branching and pruning, metabolic factors such as intr/extracellular ion concentrations.

    – The first principal component of intelligence tests given in a particular population represents a combination of g and the (average) specific characteristics common to that population. A corollary of that is different populations would appear to have different gs because the first PCs differ. The interesting question here is how the magnitudes of g and the population specific parts of the differing PCs compare.

    Which brings us to.

    – IMO any test based estimate of g is going to have a dependence on the population tested. I think the best way to evaluate this (and get a better estimate of g itself) is to analyze a large and heterogeneous population. Ideally this would result in a PC for g, a number of PCs for population specific profiles, and a number of PCs for specific skill subgroups (i.e. those in the Haier graphic). One possible issue is that imposing orthogonality on these PCs would give non-obvious results (e.g. how would capturing the Asian/white difference in M/V/S skill profile affect the individual factor components?). Perhaps it would make sense to only use PCA for g and the population differences while not imposing orthogonality on the factors.

    Another thought.

    – Is it possible to derive an approximation for an individual’s g from purely physical measurements? In theory I could envision a comprehensive set of physical measurements (e.g. brain size, neural connectivity, myelin thickness, cell ion concentrations, any other ideas?) giving a reasonable approximation of g for an individual.

    What does everyone here think about those propositions?

    P.S. Is it necessary to consider separately whether an IQ test is racially discriminatory because of the test questions or because of the norming? Your points about PSAT M/V/S weighting are very much a norming question. As far as I am aware the discrimination focus has been much more on the questions.

    P.P.S. I am a bit obsessed with cell ion concentrations for a variety of reasons. But I really do think that matters here because they affect the behavior of action potentials. Anyone interested in this idea might want to check out this online course: https://www.coursera.org/learn/bioelectricity

    • Replies: @JLK
  317. JLK says:
    @res

    I superficially disagree with part of it, but wonder if we might be in agreement after a more detailed consideration.

    You didn’t identify which part that you disagree with, but first let me clarify my statement:

    Disparate racial performance is probably somewhere in the residual of various tests after controlling for g.

    I was referring to any unfair discriminatory disparities in racial performance, which I still suspect are contained within the residual of the of various tests after controlling for g. Most of the disparities of course arise out of innate differences in ability and environmental factors.

    The 2010 Deary article that you cited is a good source on this. Deary states:

    as long as the test batteries are reasonably diverse, g factors from different test batteries are almost perfectly correlated

    The effects of “reasonably diverse” in the sense of including visual-spatial batteries along with reasoning and/or verbal batteries can be see in the quote regarding Native American testing in post #290:

    The Native Americans obtained higher visualization than verbal abilities in all of the six studies in which tests of the two abilities were given. The median visualization IQ in these studies is 89.5 and the median verbal IQ for the studies is 81. The same strong visualization-weak verbal profile of abilities is present among East Asians (see Chapter 10, Section 1) to whom the Native Americans are genetically closely related.

    In other words, the g calculated from a diverse battery of tests on Native Americans is going to be more accurate than one based on a test like Wordsum. Tests like Stanford-Binet (which if I recall correctly was the one used in this study) that have a spatial battery have a correlation to g (0.9 or above) that is much higher than Wordsum. This is in agreement with my thesis that any unfair discriminatory disparities in racial performance can be accounted for within the residual of the of various tests after controlling for g. Wordsum has a bigger residual and is more racially discriminatory.

    The first principal component of intelligence tests given in a particular population represents a combination of g and the (average) specific characteristics common to that population. A corollary of that is different populations would appear to have different gs because the first PCs differ.

    I’m not sure what you mean by “principal component,” but Deary says no as long as the batteries of tests are reasonably diverse. The scores may vary between groups, but that’s a different concept.

    • Replies: @res
    , @utu
  318. res says:
    @JLK

    You didn’t identify which part that you disagree with

    I was specifically referring to “I suspect not” regarding whether gs might vary between ethnic groups.

    I think we are basically in agreement regarding the discriminatory issues. Worth noting those can appear in a variety of guises and the sign may differ depending on the metric chosen. For example, as you say Wordsum might misestimate g across ethnicities (e.g. blacks might perform better relative to Asians on it compared to a test which includes a spatial component) but it will still probably show a difference which some would interpret as discriminatory. And may even be discriminatory in reality depending on the particular words chosen.

    I’m not sure what you mean by “principal component,” but Deary says no as long as the batteries of tests are reasonably diverse.

    Ah, that is key because it underpins the point I am trying to make. I am using principal component as in Principal Component Analysis (PCA): https://en.wikipedia.org/wiki/Principal_component_analysis
    And am referring to the first PC of a diverse set of intelligence tests as one definition of g: https://en.wikipedia.org/wiki/G_factor_(psychometrics)#Factor_structure_of_cognitive_abilities

    Could you elaborate on what and where Deary says something counter to this? For a concrete (extreme and contrived though) example say one group always scores M = 2V and another group always scores V = 2M then the respective first PCs in M-V space will be (not normalizing to length 1 for simplicity) (1, 2) and (2, 1). An equal sample size combination of both groups (assuming same range) would have (1, 1) as the first PC. Which of these should be considered g?

    The scores may vary between groups, but that’s a different concept.

    Agreed (I think). I’m referring to differences in how the scores relate to each other within each group. Not the score values themselves.

    P.S. Jensen presents an argument and data that g is stable across populations so I might be wrong.

    From page 46 of http://arthurjensen.net/wp-content/uploads/2014/06/Psychometric-g-Definition-and-Substantiation-2002-by-Arthur-Robert-Jensen.pdf

    Across Populations

    Provided that all the subtests in a test battery are psychometrically suitable for the subjects selected from two or more different populations, however defined, the obtained g factor of the battery is highly similar across the different populations. By psychometrically suitable is meant that the tests have approximately the same psychometric properties such as similar reliability coefficients, absence of floor and ceiling effects, and quite similar correlations between each item and the total score (i.e., the item-total correlation). When such criteria of adequate measurement are met, the average congruence coefficient between the g loadings obtained from representative samples of the American Black and White populations in a wide variety of test batteries is + .99, or virtual identity (Jensen, 1998, pp. 99-100; 374-375). The same congruence coefficient is found between the g loadings of the Japanese on the Japanese version of the Wechsler Intelligence Scale subtests (in Japan) and the g loadings in the American standardization sample. Similar congruence is found in European samples (Jensen, 1998, pp. 85-86).

    Jensen, 1998 is The g Factor.

    I would be interested in seeing the data so I can understand how different M/V/S tilts give rise to this mathematically. FWIW I think an Asian-Black comparison would be the most likely to show the effect since I believe they have the largest difference in M/V/S tilts.

    • Replies: @JLK
  319. utu says:
    @JLK

    This statement:

    as long as the test batteries are reasonably diverse, g factors from different test batteries are almost perfectly correlated

    you took form Deary p. 204

    http://larspenke.eu/pdfs/Deary_Penke_Johnson_2010_-_Neuroscience_of_intelligence_review.pdf

    which is based on Johnson (ref. 110 in Deary paper):

    http://emilkirkegaard.dk/en/wp-content/uploads/Still-just-1-g-Consistent-results-from-five-test-batteries.pdf

    who did take Dutch covariance matrix 46 x 46 consisting of 5 batteries of tests used on 500 subjects. He did not have individual scores only the covariance matrix. He extracted g for each battery and get correlation between different g’s. In the upper part of the Table 2 he lists the correlations. You can see that correlations of tests 3 and 2 is 0.79, 3 and 4 is 0.77, 3 and 1 is 0.88. Clearly the test 3 which is Cattell Culture Fair Test does not correlate “almost perfectly.”

    But indeed correlations among g’s of tests 1,2, 4 and 5 are high. But so what? After all they are all “intelligence tests” designed by people subjected to common “group think” within the same para-scientific paradigm. Each of this test was concocted (designed) against other existing tests within the industry and tweaked until they maximized correlations among them.

    It is really amazing that adults like Johnson or Deary can keep insisting with childlike obstinacy on the singular g-construct and keep producing papers within their confirmation bias w/o ever stepping out and seeing that the construct itself is ontologically and epistemologically flawed.

    • Replies: @JLK
    , @CanSpeccy
  320. JLK says:
    @res

    Could you elaborate on what and where Deary says something counter to this? For a concrete (extreme and contrived though) example say one group always scores M = 2V and another group always scores V = 2M then the respective first PCs in M-V space will be (not normalizing to length 1 for simplicity) (1, 2) and (2, 1). An equal sample size combination of both groups (assuming same range) would have (1, 1) as the first PC. Which of these should be considered g?

    I’d look in Deary’s Box 2. In your hypothetical, a battery of tests equally weighting M and V would be “reasonably diverse” enough not to skew g.

    The Jensen quote that you provided essentially is saying the same thing, although he doesn’t elaborate on what “psychometrically suitable” means with a real world example. There seems to be a bit of a third rail for the professionals in the field that stops them from being too blunt and specific like I was in post #315.

    I would be interested in seeing the data so I can understand how different M/V/S tilts give rise to this mathematically.

    So would I. If you find some, please give me a heads-up.

    FWIW I think an Asian-Black comparison would be the most likely to show the effect since I believe they have the largest difference in M/V/S tilts.

    I’d propose looking at an Asian-Jewish comparison, because there is a pronounced V-S skew with no big secular difference in IQ between the groups, but good luck finding the data. There is no shortage of articles on the Internet touting Jewish intellect, but next to nothing on their spatial abilities since Backman’s 1972 study that found a spatial IQ of 91.5. You would think that it would be easy to find a study listing Jewish scores on the spatial components of Stanford-Binet, as in the Native American study that I cited, but it isn’t.

    • Replies: @res
  321. JLK says:
    @utu

    It is really amazing that adults like Johnson or Deary can keep insisting with childlike obstinacy on the singular g-construct and keep producing papers within their confirmation bias w/o ever stepping out and seeing that the construct itself is ontologically and epistemologically flawed.

    You’re entitled to your opinion, but there was nothing in your post that demonstrated a flaw in the theory of a singular g.

    I haven’t spent my life in the field like many of the experts, and am open-minded on the issue. Keep in mind, though, that it would be the holy grail of the Gould crowd to prove that there is a black g and a white g, and that psychometricians have spent decades holding blacks to an arbitrary white standard, like Dr. Mengele with a set of calipers.

    There would be no shortage of funding for research to show this, but it hasn’t happened yet. g looks like it is here to stay.

    • Replies: @utu
  322. res says:
    @JLK

    There is no shortage of articles on the Internet touting Jewish intellect, but next to nothing on their spatial abilities since Backman’s 1972 study that found a spatial IQ of 91.5.

    Have you seen these papers?
    http://psycnet.apa.org/record/1975-21109-001
    https://journals.sagepub.com/doi/abs/10.2466/pms.1977.45.1.279

    • Replies: @JLK
  323. may mori says:

    I don’t have a strong feeling one way or the other in this argument, but this article comes across more as defeating a straw man. Here is my best summary of Taleb’s points:

    The conclusions that maths provide are determined by fulfilling certain assumptions. If those assumptions are not true, any results you obtain are spurious.

    Statistical regression requires, among other things, that any unexplained variation is normally distributed and has equal variation for all points. To put things more simply, it requires that it’s possible to break up the relationship between the outcome and predictors into a trend, and a completely random error.

    The relationship between IQ and financial success shows unequal variance, and certainly one that doesn’t follow a normal distribution. This observation violates the assumptions needed for the mathematical proofs to work properly, and thus give statistical conclusions their validity. In particular, results about statistical significance and explained variation can be exaggerated.

    Taleb is not saying that IQ is a random value. He proposes plausible alternative distributions that the results could be generated from, and a big part of making statistical claims is ruling out distributions. What can be ruled out is no relation between IQ and success. What can’t be ruled out is that IQ is only capturing the most simplistic parts of intelligence, which are absent in only the mentally retarded, and missing something terribly important.

    Practically, this means that if you’re trying to predict success, IQ is probably not that useful. For an example, let’s say IQ is equally distributed, with half the population low IQ, and half high IQ. A low IQ person might have 0.1% chance of success (depending on how you define it), while a high IQ person might have 0.7%. If this is the case, the overwhelming majority of successful people will have high IQ. You might even see a “correlation” as in the Grinblatt paper. But it’s absolute rubbish to choose people for jobs this way. Your failure rate will be incredible.

    The point is that there are better ways to estimate competency, and that this one is being given a mathematical facade by people who don’t know the mathematics. If you want to argue with Taleb on this, jumping right in to talking about correlations and never understanding the underlying assumptions that make them work makes you look like a fool.

  324. utu says:
    @JLK

    Do you really think that bringing up Mengele is helpful? Are you a graduate of a Jewish school of rhetoric?

    “The Gould crowd” is not interested in technical arguments about g-factor just as they are not interested in discussing properties of unicorns. Postulating unicorn existences and that all unicorns are alike comes from the Jensen crowd.

    Furthermore if some test batteries showed different PCA structures for Blacks, White, Asian and Jews “the Gould crowd” would rather stay away from it because they do not want to deal with racial differences in the first place. It should be a no brainer.

    You seem to be very keen to point out to racial differences and keep suggesting that Jews have different PCA structure from Asians. If you believe that, you cannot believe at the same time that there is a unique singular g. You need at least two factors to explain test results of Jews and Asians. This means that Jensen’s postulate goes out of the window or Jensen’s para science is valid for white people only? Apparently it did not occur to you. I have told you before “I even may support your cause I doubt that you will get anywhere with it by adopting arguments from the para-scientific field.”

    If you unreflectively follow Deary and Johnson who show a diagram (based on a sample of 500 Dutch cadets) where some tests have very high correlation with their g and some tests have lower correlation with the same g you do not realize that this itself points to the existence of the second factor that can’t be ignored. If you had a mathematical imagination you could see that one could construct a battery of tests where its principal factor would be even less correlated with the visual or verbal subtests (your choice) and that this principal factor would differ more and more from the g in the original battery of test. You can manipulate g and make almost anything what you want by redesigning the tests or by even only shuffling individual questions between the tests, i.e., by beefing up the visual or verbal part of the battery. Your insistence expressed by the statement “g looks like it is here to stay” just shows that you haven’t thought this issues through. Ask yourself the question why we do not have a g-test? Why some international body of IQ researchers did not change intelligence into g units? Is there a g test available? Why not administer a test that correlates 100% with g? Why do we need to use horses to define the unicorn?

    • Replies: @JLK
  325. JLK says:
    @res

    Have you seen these papers?

    No. I notice that they are both from the 1970s, like Backman. I’d be interested to see their numbers, but I’m not going to pay for a copy.

    Richard Lynn published an article in 2002 (not referring to these two studies) on Jewish intelligence concluding that more research was needed on the Ashkenazi intelligence profile.

    http://www-personal.umich.edu/~negisama/asdf2.pdf

    • Replies: @res
  326. JLK says:
    @utu

    This short discussion of the pros and cons of both factor analysis and principal component analysis seems to express the heart of your concerns:

    https://www.thoughtco.com/principal-factor-analysis-3026699

    Problems with Principal Components Analysis And Factor Analysis

    One problem with PCA and FA is that there is no criterion variable against which to test the solution. In other statistical techniques such as discriminant function analysis, logistic regression, profile analysis, and multivariate analysis of variance, the solution is judged by how well it predicts group membership. In PCA and FA there is no external criterion such as group membership against which to test the solution.

    A second problem of PCA and FA is that, after extraction, there is an infinite number of rotations available, all accounting for the same amount of variance in the original data, but with the factor defined slightly different. The final choice is left to the researcher based on his or her assessment of it interpretability and scientific utility. Researchers often differ in opinion on which choice is the best. A third problem is that FA is frequently used to “save” poorly conceived research. If no other statistical procedure is appropriate or applicable, the data can at least be factor analyzed. This leaves many to believe that the various forms of FA are associated with sloppy research.

    I have boldfaced the part that I think most important for this discussion. If you think of an intelligence test as a mountain climbing competition, with Jews climbing one side and Asians climbing the opposite side, it is of course discriminatory to focus too much on one of the infinite possible numbers of rotations (looking down from the top) that are closer to the Jewish side of the mountain than the Asian/Native American/Mexican side. Deary solves the problem by suggesting that the factors have a reasonable amount of diversity.

    Two verbals, one math and no spatial on the PSAT/SAT is not reasonably diverse.

    I’m open-minded on the singular g theory, but keep in mind that factor analysis does not inevitably solve for a single solution and that the tests correlating most closely to g are the longest and most diverse. There seems to be a convergence.

    • Replies: @res
    , @utu
  327. res says:
    @JLK

    Thanks for the Lynn link. I am trying to decide whether to buy this book by Storfer since it keeps showing up as a reference: https://www.amazon.com/Intelligence-Giftedness-Contributions-Environment-BEHAVIORAL/dp/1555421857

    Those two papers are available on libgen.io Search by DOI works best and the pages I linked to have DOIs.

  328. res says:
    @JLK

    There is a long history of discussions about Factor Analysis and PCA with utu on this blog if you are interesting in searching through the history.

    keep in mind that factor analysis does not inevitably solve for a single solution

    This is true of factor analysis, but not PCA. That’s the reason I focus on discussing PCA. Especially when looking at the first component. Factor analysis may make more sense when trying to map to common concepts (e.g. g, math, verbal, spatial) though.

    One way of expressing the point I was trying to make above is that I suspect the first PC consists primarily of g along with that population’s specific profile. The question is how relatively large those two quantities are. I suspect g is a good bit larger, but find the 0.99 correlations bandied about implausible for widely divergent populations.

  329. anonymous[417] • Disclaimer says:
    @Nobody

    There is a lot of math in this world.

    I could go on for hours about prime numbers, I could get as technical as anyone could want, but at the end of the day, there is nothing I know about math that Plato, whose works include a very limited number of equations- there is nothing, I suppose, that I know about math that Plato would not quickly understand, if only I were intelligent enough to explain it well: and I suppose that you, at the end of the day, think Taleb would consider Plato merely “verbose” if Plato disagreed with Taleb. You would, I believe, be wrong.

    Please be more careful about calling entire groups of people – among whom at least one, and if at least one, then likely many more than just one – might be much more intelligent than you would guess – charlatans. Taleb’s assessments are correct to a degree, of course, but the challenges to his propositions are not all charlatanesque, and that fact does not make him, reciprocally, a charlatan.

    Be polite! Show respect! People are generally not very intelligent, from the point of view of actual truth, probably not Taleb, probably not you, probably not me, and mutual respect is helpful in our mutual search for honesty and truth.

    And stop waiting, like a child, for someone to discuss Taleb’s math and show to you the mistakes in the math! Grow up, and assess the math yourself. It is quite possible there are no mistakes therein! But the map is not the territory, as Popper would have said if he thought of it first.

    • Replies: @JLK
  330. utu says:
    @JLK

    This short discussion of the pros and cons of both factor analysis and principal component analysis seems to express the heart of your concerns:

    My concerns should be your concerns. I am trying to bring important point to save you from yourself but you pretend you do not get it:

    You seem to be very keen to point out to racial differences and keep suggesting that Jews have different PCA structure from Asians. If you believe that, you cannot believe at the same time that there is a unique singular g. You need at least two factors to explain test results of Jews and Asians. This means that Jensen’s postulate goes out of the window or Jensen’s para science is valid for white people only?

    As far as my concerns it is also related to FA and PCA and how FA in particular is used by psychometrician. That it is not an objective proces but subjective based on different criteria that are used.

    https://pdfs.semanticscholar.org/1691/2e7c392550ab7f18f9b0096f04a0d239bd0e.pdf
    Principal components solutions to factoring are of-ten rotated (the next step) to achieve a simple structure of the factor loadings. However, methods of rotation do not improve the degree of fit of the factor structure to the data: The covariation explained is only redistributed among factors. Rotation is aimed at generating an easily interpretable solution. The choice of approach to rotation is similarly arbitrary, and what constitutes a “simpler” structure is a matter of judgment, even if the solutions are completely mathematical. An important but arbitrary decision is whether the method of rotation should result in orthogonal (uncorrelated) or oblique (correlated) factors.

    The FA allows you to do a lot of arbitrary data massaging until you get a desired result. The methods are like alchemical recipes that PhD apprentices are taught and they pick them up from their masters and replicate in other settings and this is how apparent consistency and agreement between different groups of alchemists is achieved. Their activity is better described by the model of Medieval craft and guild rather than science that’s why I will insist on calling it a para science. There are exceptions when some alchemist decide to follow their own path like Thurston or Cattell. The same Cattell whose g did not correlate well with other g’s in Johnson ‘s paper.

    My critique goes deeper than just quirkiness and arbitrariness of FA:

    http://www.unz.com/jthompson/fear-and-loathing-in-psychology/#comment-1948489
    I was trying to explain, but apparently failed, that batteries of tests usually used are constructed in such a way so they yield one dominant factor just as Spearman and Jensen wished and this fact is used as rhetorical device (for example by Jordan Peterson) to drive the point that IQ or g is the only intelligence that matters because all the other factors are weak and thus negligible. But I argued that one can construct a battery of tests differently, so the two factors will be obtained of approximately equal strength and then one can no longer say that there is only one single intelligence and the rhetorical force of the argument of singularity of g is lost. I did not get to what the factors g and s might be called and what their interpretation might be. This is outside of my argument and interest. In parallel to Spearman there were other researchers like Thurstone (if my memory serves me right) who advocated that intelligence is not uniaxial as Spearman envisioned and thus other factors cannot be neglected. My argument is more general and stronger than Thurston’s. I claim that this is all arbitrary because I can construct tests and subtests and the covariance matrix in different ways that will produce results that will make Thurston happy or Spearman happy. So the argument by Jordan Peterson that “g eats up all the variation” is only local, for particular matrices used, not a general argument. Also Thurston’s argument that say g is x time stronger than s is also local argument not a general one because one can construct a battery of tests where g will y time stronger than s where y≠x. I guess it might be appropriate to quote G.B. Shaw at this point: “They, Spearman and Jensen are barbarians, they think that their customs of their tribe and island are the laws of nature.”

    • Replies: @JLK
    , @JLK
  331. JLK says:
    @utu

    I don’t necessarily have a dog in your PCA versus FA fight, as most of my remarks on this comment-thread have to do with what I contend is a discriminatory skew in college admission tests. I think my points there stand either way. Your comments have intrigued me, and I’ll probably step in later and address them out of pure intellectual curiosity, since I’m mostly bored and have too much spare time on my hands these days. But before I do that, I’d like you to clarify your position on one issue:

    Do you contend that the Vocabulary Factor is just as important as the Spatial and Reasoning Factors?

    • Replies: @utu
  332. utu says:
    @JLK

    ” I’ll probably step in later and address them” -If you do it please share you findings. I myself got interested in the issue only 2-3 years ago when I started to read the UR and for the first time I ran into the true believers of the IQism and began to listen to their language and arguments they used. I have never knew of people like Sailer or Derbyshire. Where do they come from? What makes them tic? Their belief system, their psychology, their epistemology? Before that I knew nothing about it. To understand what was going on I had to deconstruct their belief system to some extent. Now I know that the g-concept is just a rhetorical device that is used by the IQists. Using a phrase that something is g-loaded or not suppose to shut up your opponent – a killer argument. The mechanism is not much different from Soviet apparatchiks who would use the argument of the scientific nature of Marxism that communism is scientifically inevitable or an argument about the hidden hand of market forces by liberal economists.

    I do not have an opinion about which parts of tests (spatial, vocabulary or reasoning) are more important. Important for what? I do believe however that intelligence if we want to quantify and measure it is not a matter of one number like g or IQ but it is rather a vector. We need several measures not one. Are there ethnic or racial differences between the relative strengths of these measures? To what extent they are heritable? If we decide to use the measures in social engineering how should we use them to build a just society? Are the measures purposefully skewed to benefit one ethnic group? If so, how did it happen? These are important questions which you seem to be concerned with. You should pursue it. But you need to navigate around many falsities of IQism and psychometric research. You will find that there is not that much data on, say Jewish IQ and that estimates of its average differ by a lot.

    • Replies: @CanSpeccy
  333. CanSpeccy says: • Website
    @utu

    why do you think Dr. Thompson in this article did not cite a single number for correlation, predictability of success, heritability’s and whatever else the IQ score is supposed to perform so wonderfully? He did not because he does not remember any of those numbers because they are given in very wide brackets and if he cited any of them they could be contested.

    For those who would like a reference to work showing the correlation between IQ and life success, here’s one — a paper by Nobel prize winner, James Heckman, and others, revealing that IQ provides only the vaguest intimation of life success, and is inferior in this respect to both school grades and achievement test results:

    What grades and achievement tests measure

    Lex Borghans, Bart H. H. Golsteyn, James J. Heckman, and John Eric Humphries
    PNAS November 22, 2016 113 (47) 13354-13359; published ahead of print November 8, 2016 https://doi.org/10.1073/pnas.1601135113
    Contributed by James J. Heckman, September 19, 2016 (sent for review January 22, 2016; reviewed by Armin Falk and Patrick Kyllonen)

    The Terman study showed the same thing, with many high IQ individuals working at menial jobs, and earning pretty much minimum incomes, often afflicted by alcoholism, divorce, depression, etc.

    • Replies: @utu
  334. CanSpeccy says: • Website
    @utu

    I do believe however that intelligence if we want to quantify and measure it is not a matter of one number like g or IQ but it is rather a vector.

    Why do you think intelligence is quantifiable? Is it even one thing? Is the ability to solve Rubik’s cube in seconds dependent on the same quality of mind that allows one person to compose a great symphony or another to devise an effective strategy of battle?

    The brain is modular, with many types of neurons and neurotransmitters (over 80 at recent count), which means there has to be independent genetic variation among mental capacities, variation revealed in most the extreme form by savants: for example, Gloria Lenhoff, IQ around 60, who knows the score of operas in 15 languages — languages in which she is able to communicate (She also knows sign language.); but who, of numbers has no clue, probably because the Williams syndrome from which she suffers has robbed her of the number- crunching brain module.

    A general intelligence factor can only exist at a very basic level. For example, cardio-vascular health will affect all brain functions, likewise many other physical and metabolic factors affecting the brain as whole. But just because those who are healthy in all respects will, all other things being equal, have a higher mean psychometrically determined g factor than those with circulatory or metabolic diseases, it doesn’t mean they’re all geniuses. Most will be just normal healthy people.

  335. utu says:
    @CanSpeccy

    Conclusion form Borghans et al.:

    “Cognitive skills predict life outcomes. This paper reinterprets the evidence on the relationship between cognitive skills and a variety of important life outcomes by analyzing the constituent components of widely used proxies for cognitive skills—grades and achievement tests. Measures of personality predict achievement test scores and grades above and beyond IQ scores. Analyses using scores on achievement tests and grades as proxies for IQ conflate the effects of IQ with the effects of personality. Both measures have greater predictive power than IQ and personality alone, because they em- body extra dimensions of personality not captured by our measures.”

    “Despite variation across datasets, consistent patterns emerge. Personality is a powerful predictor for most life outcomes across all datasets. Grades and achievement test scores are more predictive of adult outcomes than IQ.”

    “Note that most of the variance in both measures remains unexplained.”

    • Replies: @CanSpeccy
  336. JLK says:
    @anonymous

    Taleb’s assessments are correct to a degree, of course, but the challenges to his propositions are not all charlatanesque, and that fact does not make him, reciprocally, a charlatan.

    Many of the attacks on mainstream psychometric science these days have everything to do with the US college admissions tests, primarily the PSAT/SAT, because they are the closest thing to an IQ test that most Americans will ever take, and they function as a critical gatekeeper to academic and life success.

    Many parties including senior policymakers have no doubt known of the issues that I pointed out in Post #315, in some cases for decades, and there hasn’t been a scintilla of public debate, except here among the various deplorables, and this article will soon be off the front page and into oblivion forever.

    Sooner or later, the issue will be exposed in a class action lawsuit, and policymakers will face a quandary. They could fix the tests in a manner as I suggested, but then they would have to face the fallout of possible drop-offs in performance among certain demographic groups, who would then have to explain why a large proportion of their Ivy League grads and National Merit Semifinalists over the past 50 years don’t deserve a Roger Maris-type asterisk on their resumes.

    Alternatively, they could do away with the tests altogether, which would be a disaster for American long-term competitiveness as well as probably freeze existing and unjustified levels of privilege in place.

    I’d be very watchful for disingenuous suggestions that IQ testing is suddenly bad science that should be thrown out with the bath water.

    • Replies: @Wizard of Oz
  337. CanSpeccy says: • Website
    @utu

    “Note that most of the variance in both measures remains unexplained.”

    Exactly.

  338. CanSpeccy says: • Website
    @utu

    It is really amazing that adults like Johnson or Deary can keep insisting with childlike obstinacy on the singular g-construct and keep producing papers within their confirmation bias w/o ever stepping out and seeing that the construct itself is ontologically and epistemologically flawed.

    But there has to be an underlying factor, g, i, or whatever you want to call it, otherwise the whole concept of intelligence, at least as psychologists think of it, has to be bunk. For if the facility underlying the composition of Solfegietto was different from the facility underlying the formulation of the equations of electro-magnetism, or the painting of the Mona Lisa, then intelligence cannot be a single thing measured as a single point on a linear scale, but must be a whole bag of tricks.

    The psychologists have, for over 100 years, staked their credibility on the claim that intelligence is not only a single thing, but a single thing that can be reduced to a score on a 20-minute multiple-choice test. And, now, as the evidence to support this claim (the soundest thing there is in psychology, according to Jordan B. Peterson) fails to accumulate, psychologists must either admit to another colossal blunder to go along with Freudianism and behaviorism, or go on BSing. For now they seem committed to continuing with the BS.

    • Replies: @Wizard of Oz
  339. @CanSpeccy

    It is not my understanding that psychologists have staked their credibility on intelligence being a single thing, let alone one measurable by 20 minute tests even if you are not talking about the multiple intelligences crowd but rather the child psychologists or those sifting applicants for employment. Where do you get your contrary view from?

    My understanding is that typically there will be guidance based on aa probabilistic range so that a parent can be assured that their child is (almost certainly) completely ready to be challenged with arithmetic or an employer told that only 2 of the 10 applicants are likely to be able to master the required skills within the 4 week training period.

    • Replies: @CanSpeccy
  340. @Intelligent Dasein

    Timbuktoo and Ancient Mali in general were complex societies, more advanced than some European ones in their day. All this talk on this blog is nothing but generalizations routinely kiboshable with obvious and easy-to-find exceptions.

    • Replies: @Colin Wright
  341. @JLK

    It is a bit frustrating to know that it would take me well beyond your and other commenters’ interest in following this thread to get up enough maths on FA, PCA, orthogonal transformations, rotations etc to discuss the technical questions with confident expertise. But allow me some attempt(s) at simpleminded common sense.

    Is it reasonable to approach IQ as g by supposing that there are a number of similar combinations of physiological features and functions of the brain which, with the exception of some outliers (such as savants), are overwhelmingly important for the speed and correct (in modern terms: survival might have been the palaeo criterion) outcome of cognitive tasks (not excluding the optimum choice of pigments or the choice of player to pass to)? Could g be legitimately be regarded as a measure of the quality and efficiency of these elements of the brain even though the precise combinations of processing speed(?s), working memory etc. might, amongst non-savants, vary within a small range?

    Slightly different topic…. Given the SCOTUS mandated limits on using IQ tests on employment applicants, how might an employer fare who said “we choose 10 research and IT staff each year, 15 sales and 20 general admin. We whittle the fields down by using respectively our psychometric consultants’ Math, Verbal and Visuo-Spatial tests, and we are happy with the results (and btw all the top performers on each tend to show up well when it comes to general management promotion)”?

    Back to the technicalities. Given 0.7 correlations of Verbal and Math tests with g and, say 0.9 correlation for a Visuo-Spatial test (though the difference between 0.7 and 0.9 is not key to my question) what prediction of g (IQ) can be made with what probability given the score on the V,M or V-S tests? What are the likelihoods of big differences in individuals’ scores on the different tests despite equal g?

    • Replies: @JLK
    , @anonymous
  342. How very strange that employers are told how they must select employees.

  343. CanSpeccy says: • Website
    @Wizard of Oz

    It is not my understanding that psychologists have staked their credibility on intelligence being a single thing

    Here’s how Wikipedia defines IQ:

    An intelligence quotient (IQ) is a total score derived from several standardized tests designed to assess human intelligence.

    That sure implies that IQ, a single number, measures intelligence, which by implication must be a single thing, since you cannot measure independent variables with a single number.

    If Wikipedia is wrong, why doesn’t some psychologist correct them?

    But whatever your understanding, psychologists do promote the idea that IQ measures intelligence. Here’s Jordan Peterson, Toronto University Professor and perhaps the world’s most famous living psychologist, certainly among the richest*:

    People have been studying intelligence, IQ intelligence, since the 1920’s and it is a very well established branch of psychology. One of the the things I have to tell you about IQ research is that if you don’t buy IQ research, you might as well throw away all the rest of psychology. … And the IQ people have defined intelligence in a more stringent and accurate way than we have been able to define almost any other psychological construct.

    Thereafter, Peterson goes on to trash efforts to differentiate among forms of intelligence, practical versus analytical, etc.

    Moreover, Peterson argues IQ is the best predictor we have of life success. But as the paper by Nobel prize winner, James Heckman and others show, IQ is actually a poor predictor of life success, as measured for example by income, and is inferior to achievement tests, school grades and personality tests.

    ———
    Peterson, incidentally, has laid claim to an IQ of 150, which, if one accepts his claim about the predictive power of IQ tests, accords with his status as a multi-millionaire academic-entrepreneur and public intellectual. However, according to evidence assembled by Theordore Beale, author of the Peterson debunking book “Jordanetics” Peterson, who failed the Canadian LSAT exam, cannot have an IQ above 124, which happens to be right in the middle of the range for professors, and seems consistent with Peterson’s somewhat dodgy reasoning processes. But if we are to consider hand waving a distinct form of intelligence, (see video link above) I’d rank Peterson a genius.

    • Replies: @CanSpeccy
  344. JLK says:
    @utu

    As far as my concerns it is also related to FA and PCA and how FA in particular is used by psychometrician. That it is not an objective proces but subjective based on different criteria that are used.

    We could devise a so-called intelligence test using a decathlon-like scoring system, where ten (or whatever number) different subtests are weighted equally, individually scored on a percentile basis, and the scores tallied. But without some type of statistical analysis, we would have no way to assess how important each of the subtests are to overall intellectual and professional efficacy, or whether the selected tests overlap and hence overweight some factor qualities out of proportion to their importance.

    Here’s a Factor Analysis of a real decathlon: https://eric.ed.gov/?id=EJ172671

    It solved for four solutions, not for one as well-balanced intelligence tests invariably do for g, suggesting that decathletes are better off with their current scoring system in determining the best track and field athletes than they would be with a FA analysis.

    There are well-established correlations of g to things like achievement of advanced degrees, income, criminality and longevity. We could modify the Deary/Haier graphic to insert them above the little circle with g in it. But we could just as easily use path analysis to eliminate g altogether and calculate the direct correlations between spatial intelligence, for example, and achievement of advanced degrees, income, criminality and longevity. We could even take it a step further and calculate the correlation of just one of the spatial tests to such life achievements.

    So, at least for my purpose of constructing the best possible battery of college admission tests (and demonstrating the relative unimportance of vocabulary), we could do without g and essentially reach the same weighting. It would just be harder, which is why g has utility. Intelligence tests converge a singular FA solution, unlike the decathlon, so there is a lot better justification to use it.

    • Replies: @res
    , @utu
  345. JLK says:
    @Wizard of Oz

    Is it reasonable to approach IQ as g by supposing that there are a number of similar combinations of physiological features and functions of the brain which, with the exception of some outliers (such as savants), are overwhelmingly important for the speed and correct (in modern terms: survival might have been the palaeo criterion) outcome of cognitive tasks (not excluding the optimum choice of pigments or the choice of player to pass to)? Could g be legitimately be regarded as a measure of the quality and efficiency of these elements of the brain even though the precise combinations of processing speed(?s), working memory etc. might, amongst non-savants, vary within a small range?

    g is just a mathematical construct, and all the physical factors you mentioned plus many more like neuronal density and glucose uptake probably have an effect. Just like heart capacity, leg length, body fat percentage and muscle fiber ratios have an effect on marathon times.

    Slightly different topic…. Given the SCOTUS mandated limits on using IQ tests on employment applicants, how might an employer fare who said “we choose 10 research and IT staff each year, 15 sales and 20 general admin. We whittle the fields down by using respectively our psychometric consultants’ Math, Verbal and Visuo-Spatial tests, and we are happy with the results (and btw all the top performers on each tend to show up well when it comes to general management promotion)”?

    Griggs v. Duke Power outlawed the use of IQ tests for hiring purposes, which has led companies to use the much more discriminatory SAT scores (for reasons I pointed out in Post #315) instead! There is tremendous lawsuit potential here for men, Asians, Native Americans and Mexicans! I hope they go for it, it would be amusing to see how the New York Times tap dances around the corollary issues.

    Back to the technicalities. Given 0.7 correlations of Verbal and Math tests with g and, say 0.9 correlation for a Visuo-Spatial test (though the difference between 0.7 and 0.9 is not key to my question) what prediction of g (IQ) can be made with what probability given the score on the V,M or V-S tests? What are the likelihoods of big differences in individuals’ scores on the different tests despite equal g?

    It would depend on the details of the tests. There are plenty of ways to insinuate bias into them other than by stacking the deck by adjusting the weighting of V/M/S.

    • Replies: @Wizard of Oz
  346. res says:
    @JLK

    Thanks for that decathlon FA link! DOI is 10.1080/10671315.1977.10615462 for anyone else who wants full text. That paper looks at four different factor decompositions which is interesting in light of the FA vs. PCA conversation above.

    But what is really striking (and a rebuke to the “g is just a mathematical artifact” camp) is how performance breaks down into 3-4 roughly equally important factors (as you noted). Some excerpts:

    Thus, the following four procedures were applied.
    (A) Principal (truncated) components analysis with oblique rotation according to the direct quartimin criterion (Jennrich & Sampson, 1966)
    (B) Principal (truncated) components analysis with orthogonal rotation according to the varimax criterion (Kaiser, 1958).
    (C) Maximum likelihood (uniqueness rescaling) analysis with oblique rotation according to the Harris-Kaiser criterion (Harris & Kaiser, 1964).
    (D) Maximum likelihood (uniqueness rescaling) analysis with orthogonal rotation according to the varimax criterion (Kaiser, 1958).

    Which factor solution to prefer was, except for interpretability, judged from magnitude of eigenvalues (Guttman, 1954; Kaiser, 1960),total variance proportion extracted (Gorsuch, 1974), and pattern simplicity in terms of salient and hyperplane loadings (Cattell, 1966). Salient and hyperplane loadings were defined outside the interval +/- 0.40 and within the interval +/- 0.20, respectively (Cattell, 1966; Gorsuch, 1974).

    According to the principal components analysis, the first three eigenvalues-3.78, 1.51, 1.10-were larger than unity; the fourth was 0.92 and the fifth, 0.72. The total variance proportions extracted by three, four, and five factors were 64.0%. 73.2%, and 80.4%, respectively. Corresponding variance proportions for the maximum likelihood analysis were 53.1%, 58.5%, and 62.5%. Concerning interpretability, the respective four-factor solutions-either obliquely or
    orthogonally rotated-were considered most meaningful.

    From Table 2 it is obvious that the alternative procedures-A, B, C, D-mutually support the validity of a commonly indicated factor pattern. Thus, the identification of marker variables and pattern simplicity correspond very closely

    The presented factors are consecutively interpreted in terms of running speed (Factor I), explosive arm strength or object projection (Factor II), running endurance (Factor III), and explosive leg strength or body projection (Factor IV).

    Each of the running speed, explosive arm strength, and explosive leg strength factors extracts about 20% of the total variance according to procedures A and B and about 16% according to procedures C and D. This indicates that these three factors should be about equally important for the decathlon championships. The running endurance factor is less important yet extracts as much as 12% and 9%, respectively, of the total variance.

    Something similar (but not as good IMO) for women’s heptathlon: https://www.mdpi.com/2075-4663/4/1/12/pdf

  347. Diversity hire on Wall Street they got rid of pretty fast.

    Demonstration of the downside of even legal immigration.

  348. @JLK

    Thanks for your patience. I have also been trying to get a clear idea of what it would mean for there to be a different g for different races (interbreeding extended families). Would it show up by way of the g factor for race A having markedly different correlations with the V, M and V-S subtests (or some other set of sub-tests) compared with the g factor for race B?

    • Replies: @JLK
  349. anonymous[417] • Disclaimer says:
    @Wizard of Oz

    Professor Taleb is a little too excitable when it comes to measuring economic success.

    A large subset of the people who have economically succeeded, in their capacity as the people who created our technical civilization – a large subset of those people who have succeeded due to various causes ( including but not limited to, in my understanding of the map and the territory genetic adaptation to complicated societal and economic challenges- what we would unhesitatingly call “g” if we could measure it correctly) – a large subset of those people are the people who have a vested interest in making the current world a world where those who remind them of themselves materially succeed (in the future, from their point of view, as the results of high g take a few years to kick in).

    For example, the highest paying blue collar jobs – for example, service on a nuclear submarine or manipulation of the tax codes – are most beneficial to people not all that high along the “bell curve” – which Taleb correctly points out does not exist when the measurement is real world economic success.
    But the reason for that is that people at the extreme (not the fat) tail of the bell curve have designed our current real world that way.

    Such questions as Professor Taleb poses will be more easily understood when a different civilization, designed by a different generation of people at the narrow end of the bell curve, a century (or maybe less) from now, has rendered all the specific results which may or may not transpire under that different civilization, designed in its different way, by a new subset of people, who may or may not seem all that gifted in the quantity of “g” as those among our predecessors whom we now guess were so gifted.

    Numbers, genetics, the current opportunities for success (From low hanging fruit to innovations of genius), and very intelligent people gaming the system —- that is the raw material of this debate. And Professor Taleb has been convincing in pointing out the fault of certain assumptions on the part of many of us, but not convincing at all in substituting better assumptions.

    Like I said earlier in this comment thread, there is a lot of math in this world, but there is also a lot of complexity that is way beyond the large amount of accurate math that we have discovered.

    Imagine reading Eco, and saying, that is amazing, then reading Borges, and saying, well, that is even more amazing, and then reading a few of the authors who are even more inspired and intelligent than Borges. (I will not name the ones I think are that good you would not believe me anyway, unless you had an hour to listen to my insights ….). Now, remember we are discussing the real world, the relation of finite individual human beings, with their rationality, to the real world. We need to be humble, even the Kolmogorovs among us. Taleb’s arguments sometimes, from that point of view, meet the standard of common sense, and sometimes do not.

    • Replies: @anonymous
    , @Wizard of Oz
  350. anonymous[417] • Disclaimer says:
    @anonymous

    wwebd said – for the record, in the fields I am most familiar with, there are no fat tails at the extreme right of the bell curve.

    Eco, who was one in a million, was impressed with Borges in his 40s and 50s, one in ten million, who in turn was clearly to the left on that curve when compared to Joyce, one in a hundred million, in his 40s and 50s, who in his turn was clearly at least a little to the left of Proust in his 40s and 50s, who in turn was clearly to the left of Peguy , who died young (and who was at his best, for example, in the best two or three thousand lines of his epic poem Eve).

    Chandrasekhar, who long ago in a simpler pre-Silicon-Valley world figured out in a very impressive way how big stars live their big lives (imagine that!) devoted his last years to trying to understand the Principia as well as he could. He was under no illusion that Newton was not more intelligent than him, or vice versa.

    These clear differences at the extreme right of the bell curve are a hint that closer to the middle there are real differences too. Just a hint, and for goodness sake I know it is just a hint, but it is what it is.

    Unless of course you think that divine inspiration is more important than genetic ability.
    That is what I believe, after fortunate decades of observing intelligent people, intelligent animals, and intelligent insects, from the humble cockroach to the noble lobster (Not technically an insect, I know) but I have not found many people who agree with me.

  351. @obwandiyag

    ‘Timbuktoo and Ancient Mali in general were complex societies, more advanced than some European ones in their day…’

    …and not at all coincidentally exposed to Arabic influence.

    Left to their own devices, blacks literally would not not be able to write, would not be able to smelt metals, would not be able to build any structure more elaborate than an unmortared stone wall, would have no notion of even the crudest principles of health care, would not have devised a weapon more elaborate than a spear, would have no stable form of political organization beyond the tribe, and in general, would appear to be as reasonably described as unusually advanced primates rather than as human beings.

    …but there. I exaggerate. Still…

    I doubt if you’re full black yourself. It’s possible, but statistically unlikely. You can spell.

    • Replies: @obwandiyag
  352. JLK says:
    @Wizard of Oz

    Thanks for your patience. I have also been trying to get a clear idea of what it would mean for there to be a different g for different races (interbreeding extended families). Would it show up by way of the g factor for race A having markedly different correlations with the V, M and V-S subtests (or some other set of sub-tests) compared with the g factor for race B?

    You’re probably better off thinking in terms of groups having different average scores on the various R/S/V subtests rather than having different gs.

    These are statistical generalizations, of course. There are plenty of women who have high STEM aptitude, and plenty of Jews who are or were high spatials (think Einstein).

    Here’s an interesting summary of spatial thinking traits:

    http://www.gifteddevelopment.com/resources/visual-spatial-resource/spatial-strengths

    Creative, curious, out-of-the-box thinkers, VSLs learn by intuitive leaps. They remember what they see and forget what they hear. They may forget details but remember the BIG PICTURE forever.

    In other words, they learn better by reading out of a lecture environment. We need to test them, because this trait isn’t necessarily conducive to becoming a valedictorian.

    Sometimes seen as having poor organization skills, picture thinkers have their order. It centers around significance, an emotional response. Rather than outline as step-by-step learners do, where main ideas stand out like trees on the plain, spatials respond to feelings about importance. If something strikes them as worthwhile, it becomes part of their web of essentials, a mental map of things worth paying attention to. Instead of outlines—so comfortable to the stepwise —a picture thinker’s scheme of reality is more like a 3D star map. The various stars and constellations stand out in different degrees of brightness, all shining against the dark space surrounding them and all interconnected in some way. Those connections are based on feelings and sensed importance.   At times, picture thinkers not only see but feel their way through concepts. They have kinesthetic input like those cyberspace reality games that evoke muscle response to what players “experience.” These spatials grope through space as if they could touch ideas and possibilities to find what is there. Einstein, who could visualize thought experiments, spoke of using “a kind of imagistic, kinesthetic shorthand” in his thinking process. (He was groping for words, typical of spatials when trying to explain themselves.)

    Out-of-Box Thinking

    An important aspect in understanding picture thinkers is that they need to think in their own way. They are uncomfortable with following some one else’s line of thought, partly because such linear thinking is not the way their minds operate. Sometimes they really just can’t follow along step by step. They can take in each step but without that Big Picture, the steps fade away. They don’t remember details well unless those details vibrate with significance, are tagged with their own feelings, or are part of as sudden gestalt. Picture thinkers blaze their own thought trail. Most need processing time to put together their own Big Picture. There seem to be no real steps in their thinking. Often they have a sudden insight that “things go together like this!” Either slowly or in a flash, a whole concept emerges, which may be brilliant or flawed.

    Probably the closest we mortals can get to divine inspiration.

    I’m not a pure IQist, because it is clear to me from close to six decades of navigating life that things like emotional intelligence play a large part in who becomes and remains successful. In addition, material success is not always the best measure of an individual’s contribution to society. We need contrarians, iconoclasts and people who like to stir the pot, even if society isn’t always forgiving to them.

    Maybe I’ll write a book for special needs children in my considerable spare time these days, where the child who shouts that the Emperor has no clothes is somewhere on the autism spectrum and missed the social cues that he was supposed to distrust his lying eyes like everyone else.

    • Replies: @Wizard of Oz
  353. @anonymous

    Just a quick response after a first glance through your eloquent contributions as my sceptical eye picks up occasions for quibble.

    First is your 4th par reference to “design” by a future generation. What “generation” has ever “designed” the world? To try and understand what you are saying let me ask what previous generation can be said to have designed the world or any epochal aspects of the world?

    Next, you seem to be denying fat right tails on, presumably, the Gaussian IQ curve. But Cyril Burt’s figures for English schoolboys showed just that. And it is not surprising once you notice assortative mating long before the Ivies’ marriage markets. Consider the evidence in Greg Clark’s “A Farewell to Alms”. Another way to look at it is to note that the American population is, for relevant estimation, made up of several fairly distinct populations each with their own averages and standard deviations. You would expect a single Normal curve including the whole American population to have a fat right tail because of the Jewish population’s (still) higher average IQ and an even fatter left tail because of the low average IQ of African-Americans.

    • Replies: @anonymous
  354. @JLK

    Thanks. I look forward to following up on the visuo-spatial.
    Your first par is no doubt correct but I was trying to pick up on someone’s comment that lefty anti racists (or some such) would be making much of any demonstration that there was a different g for blacks. I was hoping to understand what that might mean in practice, and what would count as proof.

  355. @Colin Wright

    All lies. Just simple-minded lies from a stupid bigot. Read some real history paleontology archeology. Oh, I forgot. You can’t. Too illiterate.

  356. utu says:
    @JLK

    Did you come up with the following statement by yourself or you lifted it from somewhere?

    It solved for four solutions, not for one as well-balanced intelligence tests invariably do for g

    You realize it is begging the question fallacy or some variant of the true Scotsman fallacy. It is meaningless. The cleverly placed adverb ‘invariably’ makes it sound like a relevant and profound factual information spoken by a true expert on batteries of intelligent tests. Did you come up with it by yourselves or you lifted it from some Einstein of IQism?

    There is no definition nor test for “well-balanced intelligence tests ” apart from finding one dominant factor. So when a battery of tests yields a second strong factor then the battery is tweaked by adding extra tests or removing some tests until only one dominat factor is found or it is rejected and labeled as being not well balanced. This is the whole mystery behind the single g factor.

    I am not sure why are you bringing decathlon here. Decathlon does exist. Nobody is groping for the definition what decathlon is employing factorial analysis. It is what it is and it has three or four factors that can explain 70% of variance and this factors can be interpreted as speed, strength and so on. But in case of intelligence we do not know what it really is. We are in a process of groping for it. Psychologists are trying to define it by means of batteries of tests. The questions is what tests. There are batterie of test that yield one dominat factor but there are batteries of tests that yield two or more strong factors.

    Let’s go back to decathlon and let suppose we want to find most suitable athletes for decathlon and let suppose we measure the three factors discovered by the factor analysis in candidates. This would work but if for some reason in next Olympics scoring was changed that, say, would emphasize competitions requiring speed over strength after several generations different athletes would be attracted to decathlon and factor analysis based on their results would yield different factors than previously, i.e., there would be different proportion between factors.

    And finally ‘res’ comes up with a straw man:

    But what is really striking (and a rebuke to the “g is just a mathematical artifact” camp) is how performance breaks down into 3-4 roughly equally important factors (as you noted).

    I thought we here were arguing on a level that took for granted the fact that a covariance matrix being always semi-definite positive has positive eigenvalues out of which one must be the largest. Then from the property of Rayleigh quotient it follows that one of the eigenvectors explains more variance in data than any other out of all possible vector in this space. This vector is the best candidate for g. It is not an artifact but a mathematical inevitability necessitated by the property of covariance matrices. So g exists in this sense and thus it is not a big deal because it must exist and in this sense it it can be thought in dismissive way as an artifact. Why in dismissive way? Because many Einsteins of IQism like to use the existence of g as an argument justifying their own existence and activity.

    This strongest eigenvector to be g the way Spearman and Jensen postulated must significantly dominate the other eigenvectors, so the other vectors can be dismissed. And indeed it is often the case for batteries of tests designed by psychometricians and psychologists. And as we said already this so not because the batteries of test are ‘well balanced’, which is a meaningless notion, but because they were designed so. If you beef up the battery of tests with additional tests skewed towards the visual imagination you will increase the second eigenvalue and the question about two factors not just one is then raised.

    • Replies: @JLK
    , @res
    , @JLK
    , @JLK
  357. JLK says:
    @utu

    There is no definition nor test for “well-balanced intelligence tests ” apart from finding one dominant factor.

    I’ll assume that you are a truth-seeker in good faith, but be careful of the motivations of those attacking that “one dominant factor” (g), especially the ones who get traction in the mainstream media, because their real goal appears to be to freeze unmerited existing levels of privilege (like Ivy League representation) in place, and to conceal past evidence of pro-elite bias in the college admission testing system.

    To understand the roots of the bias, you have to go back to the mid-1940s, when the SATs were being constructed for a roll-out out on a national basis. The elite of the time understood that we were in for a long ideological and economic competition with the Soviet Union, which made an egalitarian commitment to the working classes that their brightest children would be on a level playing field with those of the party elite as far as admissions to university were concerned. It was good optics for the USSR, and it helped them economically, because their system was doing a better job of placing competent people in critical positions.

    On the other hand, the mainly northeastern American elite didn’t want their institutions swamped with bourgeoisie from Middle America, who were the “Yellow Peril” of the time. The Ivies were used to drawing most of their student bodies from a pool of about 10-15 million northeastern WASPs and perhaps 3.5 million northeastern Jews. The Midwest alone had over four times as many people of roughly comparable native ability, and there were plenty more in California and in the South.

    None of us are true experts in psychometric testing, but a little common sense applies here. If one is trying to identify people who have inherent aptitude, he or she will make an effort to ensure that the results are contaminated as little as possible by environmental factors such as whether the test subject grew up rich or poor.

    That brings us to vocabulary. There are two possible reasons for why a test subject can fail to recognize a word on a vocabulary test: (1) he or she has a poorer verbal memory than someone who correctly answered the question; or (2) he or she never ran across the word in the first place. Reason (2) has everything to do with environment, and nothing to do with ability. Vocabulary was nevertheless made a pillar of the SATs from the beginning, giving the kids who had access to the Philips Andover school library and the New York Times laying around the house a big leg up in the college admission tests over those in small town Iowa, who were lucky if their parents received the Readers Digest once a month. Of course, most people of color were in an even poorer information environment than the Iowa kids.

    An Iowa farm boy is on a much more level environmental footing with the Phillips Andover kid when they are both given a surprise visual-spatial type question of the type found in the modern long-form Stanford-Binet or Wechsler IQ tests, but there have never been such questions on the PSAT/NMSQT or SAT.

    Mainstream media outlets like the New York Times were speaking out of both sides of their mouths in the ’80s when they were simultaneously assuring us that vocabulary was the best single proxy for intelligence, while rueing that college admission tests were culturally biased against black people. It was crocodile tears, while the elite made sure that their bread stayed well-buttered.

    They can’t credibly claim ignorance, either, because there are plenty of signs in the scientific literature and other full-battery intelligence tests that the relatively weak association between vocabulary and IQ exposed in the Deary/Haier diagram has been known for decades, quite like back to the very inception of the SATs. In fact, I have run across several publications from Jensen and others where the high correlation between spatial and g, and the lower correlation of vocabulary to g was couched in inference, but they tiptoed around actually stating it.

    Science relies on finding sources and it does not exist in a political vacuum. Not even politically incorrect science is immune from politics and elite influence.

    This tells you everything you need to know should the New York Times join Taleb in his assault against g. g may not be perfect, but it is a very effective bullshit detector against post facto justifications for the pro-elite bias of the PSAT/SAT, that are essentially trying to convince numerate people not to trust their lying eyes when they look at the Deary/Haier diagram.

    As another poster suggested, if you don’t like the factor analysis methodology of g, do the math yourself, and submit it for peer review like Spearman and Jensen did. Their works survived that cauldron in the face of considerable political headwinds, so they are the best model that we have for the time being.

    • Replies: @DFH
  358. DFH says:
    @JLK

    Reason (2) has everything to do with environment, and nothing to do with ability. Vocabulary was nevertheless made a pillar of the SATs from the beginning, giving the kids who had access to the Philips Andover school library and the New York Times laying around the house a big leg up in the college admission tests over those in small town Iowa, who were lucky if their parents received the Readers Digest once a month. Of course, most people of color were in an even poorer information environment than the Iowa kids.

    But Vocabulary has a smaller gap between black and white scores than other subtests

    • Replies: @JLK
    , @JLK
  359. res says:
    @utu

    And finally ‘res’ comes up with a straw man:

    But what is really striking (and a rebuke to the “g is just a mathematical artifact” camp) is how performance breaks down into 3-4 roughly equally important factors (as you noted).

    Not a strawman. Yet another example of projection?

    See g, a Statistical Myth

    A sample:

    Factor analysis is handy for summarizing data, but can’t tell us where the correlations came from; it always says that there is a general factor whenever there are only positive correlations. The appearance of g is a trivial reflection of that correlation structure.

    BTW, I seem to recall you arguing just this “strawman” a while back. At least you are capable of learning.

    To be clear, the important point is the amount of variance explained by g compared to what would appear in the first principal component for random data or real data like for the decathlon.

    • Replies: @utu
  360. anonymous[417] • Disclaimer says:
    @Wizard of Oz

    “design” was used by me in the sense of “the attempted arrangement of economic opportunities for the less gifted by the more gifted” – something which varies by generation, I think we can all agree. I was not trying to imply a more ambitious design than that.

    As to “fat tails” – I meant to refer only to those “fat tails” observed in graphs that feature “g” as measured by available IQ tests along one axis and “economic success” as measured by rough indicators such as salary and accumulated wealth along the other axis.

  361. JLK says:
    @DFH

    But Vocabulary has a smaller gap between black and white scores than other subtests

    Vocabulary tests tend to discriminate against poor people, whether they are in the South Bronx, the barrio, the reservation, Chinatown or in rural Appalachia. All the households where you’re not likely to see Harpers Magazine laying around.

    On the other hand, as you point out, the black intelligence profile tends, like the Jewish profile, to be verbal-dominant. It’s just that they are more than a standard deviation apart. Overweighting vocabulary may slightly raise black scores, but it is going to vault Jewish scores a lot more, and it is going to hurt every other group of poor people relative to the wealthy and the middle class.

    • Replies: @James Thompson
  362. JLK says:
    @utu

    There is no definition nor test for “well-balanced intelligence tests ” apart from finding one dominant factor. So when a battery of tests yields a second strong factor then the battery is tweaked by adding extra tests or removing some tests until only one dominat factor is found or it is rejected and labeled as being not well balanced. This is the whole mystery behind the single g factor.

    You previously wrote that you don’t have an opinion on which factors are more important, or for what. So what is the “second strong factor” that you think is being suppressed in favor of a singular g?

    Name it, so we can evaluate your argument.

  363. @JLK

    Not sure Harpers Magazine has quite the effects you imagine.
    http://www.unz.com/jthompson/vocabulary-humanitys-greatest

    • Replies: @JLK
  364. Joe Mack says:

    No King but Jesus.
    Talib may be right about much, insights aplenty, but he ain’t Jesus come to Earth.

  365. CanSpeccy says: • Website
    @CanSpeccy

    Here, according to Wikipedia, is how g or general intelligence is estimated:

    The g factor … is a variable that summarizes positive correlations among different cognitive tasks … composite scores (“IQ scores”) based on many tests are frequently regarded as estimates of individuals’ standing on the g factor.

    That scores on individual tests must inevitably be correlated in some degree is self-evident to anyone with any biological awareness. All forms of mental function depend on the activity of neurons. Therefore, any condition that impairs neuronal function, whether at the cellular or organismal level, will negatively affect all intellectual functions, i.e., the g factor.

    At the cellular level, dozens of genetically determined metabolic abnormalities impair neuronal performance as do environmental toxins and nutritional deficiencies. Likewise, multiple conditions at the organismal level, cardiovascular health and hence the brain’s oxygen supply, for example, affect overall mental function. Such factors could largely account for differences in mean population IQ between sub-Saharan Africans and, say, Europeans: the Africans being more severely afflicted by brain damaging disease, e.g., malaria, and nutritional deficiencies, e.g., of iron, and dietary ingestion of toxic heavy metals.

    But the brain does not work like a PC, where all operations, and hence the g factor, depend on one central processing unit. The brain is modular: with areas specialized for processing different types of sensory input and delivering different types of behavioural output. That means that perfect correlation among different mental functions is most unlikely, each brain module being subject to independent genetic variation and susceptible in varying degrees to environmental toxins, disease, or dietary deficiency.

    Furthermore, estimated values of the g factor seem to be entirely arbitrary since they depend on what tests provide the basis for the analysis. If one goes beyond the tests generally preferred one sees huge individual variability in aptitude. For example, between musical aptitude and IQ test the correlation is trivial. A cursory examination of related literature suggests essentially zero correlation between IQ and artistic ability. Thus, by taking a more realistic view of the scope of the human mind, the g factor will surely diminish to the point of total insignificance.

  366. JLK says:
    @James Thompson

    Not sure Harpers Magazine has quite the effects you imagine.

    As you can probably imagine, I was being a bit flippant. I’ve been focusing on US college admissions tests, which are given around age 16-17. It would be the total childhood informational and nutritional environment that counts. This publication puts the heritability of the vocabulary portion of the WISC-R subscale at 0.72, which allows up to a 0.28 variation based on environmental influences. That’s not insignificant.

    By the way, what is your assessment of the reliability of the Deary/Haier correlations? I notice that you cited Deary in an earlier post.

  367. JLK says:
    @utu

    So when a battery of tests yields a second strong factor then the battery is tweaked by adding extra tests or removing some tests until only one dominat factor is found or it is rejected and labeled as being not well balanced. This is the whole mystery behind the single g factor.

    I’ve actually found a study for you that concludes a two-factor model (spatial separate from g) is a better fit than the singular g theory. Top right, page 2.

    https://www.researchgate.net/publication/313885447_Phenotypic_and_genetic_evidence_for_a_unifactorial_structure_of_spatial_abilities

    CFA was also used to test the distinctiveness of spatial abilityfrom g, as indexed by verbal and nonverbal ability measures(Methods). As expected, spatial ability has considerable overlapwith g(perhaps driven by the substantial overlap with Raven’sMatrices) (24), but a two-factor model (spatial ability and g;SIAppendix, Fig. S2B) fitted the data better than a one-factormodel (g;SI Appendix, Fig. S2A), indicating that spatial ability isdistinct from other cognitive abilities.

    Interesting.

    • Replies: @res
    , @utu
  368. res says:
    @JLK

    Thanks for that link. I am having some trouble interpreting how their spatial factor relates to g. If I understand correctly, Figure S2 is the most relevant for that question.

    Panel S2b indicates that g has a 0.79 correlation with their spatial factor. Their measures of fit (e.g. AIC, BIC) indicate the two factor model is better, but the improvement over the one factor model looks small to me. They seem to studiously avoid quantifying how much difference the second factor makes. Which seems odd to me given that two techniques I would use to look at that (ANOVA and scree plots) are used for other purposes in the paper.

    It seems to me that their spatial factor is too entangled with g to be useful as a separate concept. I think it would be much better to force orthogonality on the two factors and look at the relative explanatory power of the second. Any alternative viewpoints?

    This 2018 review paper looks useful to me. Any thoughts?
    A Heuristic Framework of Spatial Ability: a Review and Synthesis of Spatial Factor Literature to Support its Translation into STEM Education
    https://link.springer.com/article/10.1007/s10648-018-9432-z

    • Replies: @JLK
  369. JLK says:
    @res

    Thanks for that link. I am having some trouble interpreting how their spatial factor relates to g. If I understand correctly, Figure S2 is the most relevant for that question. Panel S2b indicates that g has a 0.79 correlation with their spatial factor. Their measures of fit (e.g. AIC, BIC) indicate the two factor model is better, but the improvement over the one factor model looks small to me. They seem to studiously avoid quantifying how much difference the second factor makes. Which seems odd to me given that two techniques I would use to look at that (ANOVA and scree plots) are used for other purposes in the paper.

    It seems to me that their spatial factor is too entangled with g to be useful as a separate concept. I think it would be much better to force orthogonality on the two factors and look at the relative explanatory power of the second. Any alternative viewpoints?

    Figure S2 appears to be very interesting evidence on the issue of a singular g, which may serve utu’s point. When g is calculated using eight non-verbal, predominately spatial subtests and just two verbal subtests, the correlation of Verbal to g drops from Deary’s 0.7+ to just 0.36. They look like altogether different gs.

    Maybe Deary’s g is a characteristic Dutch g, and it will follow that there are characteristic black, Jewish, Asian, Native American and Mexican gs. That might raise questions about which g one should be loading for when designing a long-form IQ test like the next edition of Stanford-Binet or WAIS, and whether the putative correlations of g to longevity and various forms of life success will vary by ethnic grouping.

    Jensen published this publication several decades ago on testing for the “Spearman Hypothesis.” Spearman had earlier postulated that the black-white intelligence difference would be most pronounced on the subtests having the highest g-loading. Jensen claimed to have confirmed this, but the devil is in the details, and the general conclusion that blacks are better verbals than spatials does not appear to be uniformly supported across the tests. Jensen’s g correlations of various subtests also suggest, just based on my quick eyeballing, that he may have been working with a different g-norm than Deary.

    It would be nice to have a published expert playing the contrarian rather than people like Gould, Taleb and Howard Gardner, who, whether by design or not, can be pushed aside as strawmen.

    • Replies: @JLK
    , @res
    , @utu
  370. JLK says:
    @JLK

    Maybe Deary’s g is a characteristic Dutch g,

    Correction for the record: Deary’s g is based on the Salthouse studies, which tested 7000 subjects, probably American. I was repeating something utu wrote about Dutch cadets, which appears to have been a mistake.

    • Replies: @utu
  371. res says:
    @JLK

    When g is calculated using eight non-verbal, predominately spatial subtests and just two verbal subtests, the correlation of Verbal to g drops from Deary’s 0.7+ to just 0.36. They look like altogether different gs.

    Good point. I had not picked up on that. It also argues against my point about excessive correlation of their spatial factor with g (if calculated in a more conventional fashion).

    Table 2 (page 11/203) of your Jensen link is interesting. It gives the g correlation with Spatial Reasoning as 0.31 with a B/W difference of just 0.19 SD. But what really caught my eye is the note: “The g loading derived from the larger sample (black or white) was used in this analysis.” Why bother with that if the calculated g does not vary between the populations?

    It makes sense to me that the calculated first PC should be sensitive to the composition of both the population and subtests, but I am having trouble thinking through the implications. How can we extract g as a subtest/population invariant quantity in the face of that?

    Perhaps a way of thinking about a general intellectual functioning factor (giff) is as a multiplier of a population specific profile? So we would have:
    – A population specific profile (psp) vector which has both length and direction. The length captures the magnitude of the average difference in group abilities. The direction the M/S/V/etc. balance differences. The directions probably don’t vary that much, but I think enough to matter.
    – A giff as a scalar quantity (say normalized as a Z score) which multiplies the psp.
    – An individual specific profile (isp) vector which has both length and direction which captures the difference between an individual’s full profile and their projection onto the appropriate psp line.

    So an individual’s ability profile would decompose as:
    psp * giff + isp
    With the giff presumably being an analog to what we know as g.

    This does not account for group differences in variation (i.e. the giff SDs will likely differ between populations). I think each axis of the space should be Z scored, but not sure which group to use as a reference.

    For actually working with this construct probably best to use both the full subtest space and a reduction into 2 or 3 PCs for visualization. Using a group of subtests which is a superset of existing popular test batteries would allow an interesting look at test bias (e.g. your PSAT doubled verbal comments) and how those align with group characteristics.

    Any thoughts? Am I off in the weeds?

    • Replies: @JLK
  372. utu says:
    @JLK

    Deary cites only:

    (110) Johnson, W., te Nijenhuis, J. & Bouchard, T. J. Still just one g: consistent results from five test batteries. Intelligence 32, 81–95 (2008).

    (see) http://emilkirkegaard.dk/en/wp-content/uploads/Just-one-g-Consistent-results-from-three-test-batteries.pdf

    for the claim:

    “As long as test batteries are reasonably diverse, g factors from different test batteries are almost perfectly correlated (110). That is, as long as one administers enough tests, the general intelligence factor from one group of tests will agree closely in ranking with the general intelligence factor from any other group of tests.”

    You notice he makes the same fallacious (tautological) false predicate “reasonably diverse” as you did with “well-balanced” reinforced with “invariably.” Keep in mind that g derived form Cattell battery does not correlate well so it must mean (invariably) that it was not “reasonably diverse” and not “well-balanced” .

    Anyway, the data Johnson used:

    “We made use of the data matrix of 46 mental ability tests published by de Wolff and Buiten (1963). This matrix is reprinted here as Appendix A. We did not have access to individual participant data of any kind. The sample on which the matrix was based consisted of 500 professional seamen of the Royal Dutch Navy. “

  373. JLK says:
    @res

    Any thoughts? Am I off in the weeds?

    I’m agnostic on the math, because (1) it’s not really in my present toolset this far out of college without putting more effort into it; and (2) there is probably more than one road to Rome as far as designing the most effective aptitude tests are concerned. Whether one uses a statistical uniaxial standard of g that is calculated by FA/PCA, a bi-factor solution or a decathlon-type (which is kind of what the SAT does, with an overload on verbal) scoring system, one first needs to identify the salient components (such as Deary’s top three of Reasoning, Spatial and Vocabulary) and make a decision on how to respectively weight them.

    The singular g versus decathlon debate has everything to do with the weighting process. Some of the g proponents seem to feel that only reasonable diversity is needed for accuracy when weighting. Personally, I’d try to acquire more data on the relationships of R, S and V to competence in a wide cross-section of professions, and design tests that weighted them more precisely accordingly to importance.

    The real problem is the pernicious influence of politics on science. Too much is at stake in terms of future national competitiveness, in all spheres, to continue to allow these critical gatekeeping mechanisms to be clandestinely gerrymandered and manipulated with trickle-down psychometric patronage schemes. The process needs to be transparent, calculated for the better good of the nation and kept in the hands of scrupulously unbiased people. Affirmative action should be a completely separate policy debate. Don’t mess with the tests.

    I’m not complete fan of the Chinese system, but I was reading a few weeks ago how the most promising and ambitious junior party members are sent to rural villages as part of their training. They don’t come back or get promoted until they demonstrate making the villager’s lives better in a tangible way. We could use a high-level, principled monastic class like that, sworn to public service and integrity. When they’re finished with the psychometric system, we could send them to oversee Wall Street.

    • Replies: @utu
  374. utu says:
    @res

    Yes, I am learning. You can use me as a role model. Three years ago or so I knew nothing about the whole subject and had no opinion or prejudice about it. Only after entering the UR here I encountered people speaking the language of IQism in which they were unreflectively spouting tautologies based on reifications which clearly indicated cult like behavior based on some epistemological errors and indoctrination. I wanted to get to the bottom of it and answer the question how otherwise intelligent and educated people (like for instance yourself) got caught into it. I got interested in it and I have learned quite a lot since. And keep in mind that I was never hostile to research on intelligence whatever it is or to the idea that it has a strong genetic component. I was hostile to the cult like group think which unfortunately many morons and yahoos pick up and keep spreading.

    The “trivial reflection of that correlation structure” or “an artifact” are valid rhetorical objections because the IQists have reified g while it was just a mathematically inevitable necessity and the fact that it was dominat over other eigenvectors which it was just a happenstance of matrices of tests they used or perhaps by tweaking of the tests to make it to suppress the second eigenvalue. When they had a battery of tests that produced the second eigenvalue too strong to their liking they would pronounce the battery as not sufficiently “well balances.” So they would stay forever dancing within the vicious circle of magic spells cast by Spearman and Jensen. A clinical case of cult behavior.

    Already in 1920’s. Godfrey Thomson made a similar objection that the dominat vector called g is not a big idea because an entity with properties of g appears in randomly generated data:

    http://www.aceintelligence.com/detailed_history_of_iq.php
    The British statistician Godfrey Thomson touched off a debate that continues to-date when he demonstrated in the 1920’s that even scores randomly created with tosses of dice could produce a hierarchical pattern of correlation coefficients similar to those Spearman utilized (Fancher, 1985)

    • Replies: @Wizard of Oz
  375. utu says:
    @JLK

    Good catch. Thanks for the link. I found two papers challenging singe g:

    http://www.iapsych.com/articles/castejon2010.pdf
    Confirmatory factor analysis of Project Spectrum activities. A second-order g factor or multiple intelligences?

    Overall, the results indicate that the Spectrum activities are not as separate from g as proposed by the defenders of multiple intelligences theory, nor as unitary as argued by the defenders of g factor models.

    Non-g Factors Predict Educational and Occupational Criteria: More than g
    https://www.mdpi.com/2079-3200/6/3/43

    As discussed below, my research on non-g factors calls into question the primacy of g hypothesis, which assumes that g explains the predictive power of cognitive tests and that non-g factors have negligible predictive power.

    The key result was that the non-g residuals of the SAT and ACT predicted college GPA almost as well as g predicted college GPA (βs ≈ 0.30).2 The results are inconsistent with the primacy of g hypothesis, which assumes that non-g factors have negligible predictive power

  376. utu says:
    @JLK

    “Personally, I’d try to acquire more data on the relationships of R, S and V to competence in a wide cross-section of professions, and design tests that weighted them more precisely accordingly to importance.”

    I believe that three scores are better than one weighted score because the weights would depend on utility. How do you want to utilize the human material? Which cattle will be milk producing and which cattle is better for veal production? A school that wants lawyers will use different weights than a school that wants mechanical engineers.

    And I agree that more research should be done to look at many facets of mental abilities. But the IQists have been caught in Spearman and Jensen dogma. The dogma is an obstacle to progress. The stagnation may also come form the fact that the racists are happy seeing Blacks down and Jews seeing themselves up while it is possible that picture is more complicated that Blacks are much better on verbal intelligence and Jews are much worse on spacial intelligence than it is thought.

    • Replies: @JLK
  377. utu says:
    @JLK

    It would be nice to have a published expert playing the contrarian rather than people like Gould, Taleb and Howard Gardner, who, whether by design or not, can be pushed aside as strawmen.

    Very good point. All three of them are just good talkers. Some point they make are good but they do not go to nitty gritty and they miss some important points. To make a significant dent in the structure built by the IQists we need somebody with high SQ (Sitzfleisch Quotient) and with significant mathematical skills. This would require time and money and courage as whoever would do it would end up making a lot of enemies.

    (1) The Factor Analysis aspect should be attacked. It is an alchemy like recipe based practice hiding behind mathematical procedures which each in itself can be mathematically valid but as a whole it may be (I am sure it is in some cases) quackery and charlatanry.

    (2) Particular studies and data should be reanalyzed from the point of view of skeptics or somebody who have opposite bias.

  378. JLK says:
    @utu

    I believe that three scores are better than one weighted score because the weights would depend on utility. How do you want to utilize the human material? Which cattle will be milk producing and which cattle is better for veal production? A school that wants lawyers will use different weights than a school that wants mechanical engineers.

    I’m inclined to agree with you, especially for undergraduate admission tests. It makes sense to style the LSATs, MCATs, etc. for the profession in question.

    And I agree that more research should be done to look at many facets of mental abilities. But the IQists have been caught in Spearman and Jensen dogma. The dogma is an obstacle to progress. The stagnation may also come form the fact that the racists are happy seeing Blacks down and Jews seeing themselves up while it is possible that picture is more complicated that Blacks are much better on verbal intelligence and Jews are much worse on spacial intelligence than it is thought.

    It’s about individuals, not groups. The only reason that I mentioned ethnic groups on this comment-thread is to point out that some groups (like Native Americans and Mexicans) will, intentionally or not, be harmed if the tests are gamed to improve the performance of other groups.

    The system was pretty good for me, and I’m retired now, so the only dog I have in the fight is the desire for a competitive and fair America.

    • Replies: @utu
  379. @utu

    “Reification” is not really helpful in advancing the dialogue. It can be simply an insult as in suggesting that someone mistakes the map for the territory but even the treatment of reification as a fallacy amongst Wikipedia’s many explanatory pars about its different applications leaves room for ambiguity. Can you walk us through what you see as fallacious or plain wrong or nonsensical in the way Steve Sailer, Peter Frost, Greg Cochran or John Derbyshire, for example, have used “g” or “IQ”?

    How about this for common sense? At least some university courses require what would be generally regarded as intelligence. Law, for example. So you want to find a quick and simple way of determining whether some potential students have the intellectual wherewithal to complete their studies satisfactorily. With a lot of intuition and a lot of trial and error you put together a likely seeming battery of tests and validate them by reference to the performance on them of recent graduates. You find that a g which you extract mathematically serves to rank the graduates as well as or better than any one of the tests. You use the testing for g thereafter to provide at least the threshold that applicants for admission must cross. Then you find a simplified, shorter battery of tests which will give you almost the same g scores. From this point you discover how well g correlates with all sorts of other desirable outcomes where cognitive abilities are important. So you recognise g and IQ as useful, up to a point, without having to apologise for simple minded reification.

    • Replies: @utu
  380. utu says:
    @Wizard of Oz

    “Reification” is not really helpful in advancing the dialogue. It can be simply an insult as in suggesting that someone mistakes the map for the territory but even the treatment of reification as a fallacy amongst Wikipedia’s many explanatory pars about its different applications leaves room for ambiguity. Can you walk us through what you see as fallacious or plain wrong or nonsensical in the way Steve Sailer, Peter Frost, Greg Cochran or John Derbyshire, for example, have used “g” or “IQ”?

    Since you brought up Korzybski let’s stay with him. Any time your heros use a construct IQ is… they are guilty of reification

    ” (From Wiki on Korzybski) Note that this premise uses the phrase “is not”, a form of “to be”; this and many other examples show that he [Korzybski] did not intend to abandon “to be” as such. In fact, he [Korzybski] said explicitly that there were no structural problems with the verb “to be” when used as an auxiliary verb or when used to state existence or location. It was even acceptable at times to use the faulty forms of the verb “to be,” as long as one was aware of their structural limitations.”

    because I doubt they are “aware of structural limitations” of the form “IQ is” they use. They do not use “IQ is…” in the same sense as “his GPA is..” or “his batting average is…” became they know that if a student worked harder his GPA could have been better or if the baseball player had less challenging opponents his batting would be higher. But in case of IQ or g they gave them different ontological attribute that is unwarranted.

    It is OK to ask what was your IQ score the last time you took the test and what was the test you took. Keep also in mind that test and retest correlate at 0.9 level which means ±1 sigma of difference between two tests is ±6.7 IQ points. So on simple empirical level stating his IQ… is at least very inaccurate.

    Clearly you are misinformed:

    You find that a g which you extract mathematically serves to rank the graduates as well as or better than any one of the tests.

    If you give students T_1,..,T_n tests from which a psychometrician will derive g according to Jensen’s secret recipe this g will not be the best predictor of, say, GPA. Another linear combination of T_i’s different than g will predict GPA’s better than g. And to complicate it further the coefficients that maximize the fit with their GPA’s will be different for student at Caltech and at Georgetown.

    Here you are pleading utilitarian value:

    So you recognise g and IQ as useful, up to a point, without having to apologise for simple minded reification.

    The key phrase is “up to a point” because predictive power of IQ is not as high as you were made to believe by your heros. Furthermore I do not think that g is being reported on the tests from which it potentially could be extracted. You can’t extract g from Raven matrices test. Do you know g score of your children, btw?

    The whole purpose of g is ideological one. Part of it was reductionism and physics envy and then the effect of self-fulfilling prophesy as the test which as you say were designed with “a lot of intuition and a lot of trial and error” where in the trial an error part the tests that were too strong on the second eigenvector were weeded out.

    • Replies: @Wizard of Oz
  381. utu says:
    @JLK

    It’s about individuals, not groups.

    I picked it up from you. You were using groups in your arguments.

    • Replies: @JLK
  382. JLK says:
    @utu

    You were using groups in your arguments.

    There’s no getting around the need to reference groups in order to assess the presence or absence of discrimination against groups.

    If you give students T_1,..,T_n tests from which a psychometrician will derive g according to Jensen’s secret recipe this g will not be the best predictor of, say, GPA. Another linear combination of T_i’s different than g will predict GPA’s better than g. And to complicate it further the coefficients that maximize the fit with their GPA’s will be different for student at Caltech and at Georgetown.

    I had an exchange upthread with academic gossip about this. She (I am assuming) pointed out that the SAT writing test predicted first year GPA at Berkeley better than SAT-V or SAT-M. I questioned whether they were controlling for course difficulty. I subsequently read about another study where they were correcting for that, but I suspect there are other important variables still not being controlled for. How do you control for the likelihood that a lot of high school whiz kids who ace the SATs never had to develop study skills until they are plunked down among students of like aptitude at an elite school? The plodders who had to discipline themselves to study all along will no doubt have a relative advantage for a while.

    • Replies: @utu
  383. @utu

    It is hard to take this as a serious contribution to discussion when you start your response with “Since you brought up Korzybski” and I have not mentioned him, whoever he is.

    I note that you criticise g for not being the best “predictor” of GPA. So what, if it is efficiently useful unless you can offer a better substitute of comparable efficiency – even if the purpose was actually predicting GPAs, which is unlikely?

    “The whole purpose of g is ideological one”. Pure bluster. Meaningless. I have a simplified version of how and why g might be developed and used which has nothing to do with ideology and you haven’t even tried to dispute it and prove that what I was describing was really ideological.

    • Replies: @utu
  384. @Curmudgeon

    “I was surprised at the large number of them who had very little understanding about the real world outside their area of expertise.”

    Exactly what Socrates observed when he set about querying those around him who boasted of their knowledge and their ability to impart wisdom.

    What he found was that their mastery of a particular field gave rise to excessive self confidence which misled them into falsely assuming that they were knowledgable in all things. And, that the consequences of this overconfidence were worse than if the person had not acted on the basis of his assumed knowledge to begin with.

    Blinded by overweening confidence in their omniscience, they blundered into crises of their own making, which were worse than if they had proceeded as though they knew nothing. Not too different that Taleb’s Black Swan thesis.

  385. utu says:
    @JLK

    One would expect that colleges would do the evaluation of the predictive powers of SAT and its subparts in almost real time (or at least once a year) and that there would be a lot of data floating around. Most likely the reason we do not see it and no specific number indicating the quality of prediction was imprinted in public space (did academic gossip was specific?) is that the numbers are all over the place. This makes one wonder what is the real utility of SAT and it subparts.

    “The plodders who had to discipline themselves to study ” – Obviously the Sitzfleisch Quotient plays much bigger role than what the IQists would like to admit.

    • Replies: @utu
  386. utu says:
    @utu

    I saw somewhere that GPA in HS predicts GPA in college better than SAT. Which makes sense because of the Sitzfleisch Quotient.

    Here is some research that contradicts Spearman’s Law of Diminishing Returns:

    SAT predicts GPA better for high ability subjects: Implications for Spearman’s Law of Diminishing Returns
    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3090148/

    SAT correlations with GPA were higher for high than low ability subjects. SAT g loadings (i.e., SAT correlations with g) were equivalent for both groups. This is the first study to show that the predictive validity of the SAT varies for ability groups that differ in g. The results contradict a presumption, based on Spearman’s Law of Diminishing Returns, that a test’s predictive validity should be lower for high ability subjects. Further research is needed to identify factors that contribute to the predictive validity of the SAT for groups that differ in g.

  387. res says:

    This might be of interest: Spearman’s law of diminishing returns: A statistical artifact?
    https://www.sciencedirect.com/science/article/pii/S0160289613000809

    I’m trying to decide how much that SAT/GPA finding really is contrary to SLODR: https://en.wikipedia.org/wiki/G_factor_(psychometrics)#Spearman’s_law_of_diminishing_returns

    I think this part is: “SAT g loadings (i.e., SAT correlations with g) were equivalent for both groups.”
    But the other part: “SAT correlations with GPA were higher for high than low ability subjects” seems to be addressing something different.

    The latter makes perfect sense based on the Sitzfleisch Quotient. High ability subjects are much more dependent on intellect in general even if they may tap more in one specific area than another.

    As imperfect as the SAT (or g, IQ tests, etc.) is as a metric for ability, IMO GPA is far, far worse.
    – Subject difficulty.
    – Wildly varying standards between schools and subjects.
    – Change over time.
    – Grade inflation compressing the usable scale.
    – Possibility of non-academic bias in more subjective classes.

    Steve Hsu made some good blog posts discussing GPA/SAT at the University of Oregon. For example: https://infoproc.blogspot.com/2010/04/dating-mining-university.html
    Search for “Schombert” in his blog for more.

    These papers have some good data looking at SAT/GPA/course difficulty.

    An Alternative to Traditional GPA for Evaluating Student Performance
    https://projecteuclid.org/download/pdf_1/euclid.ss/1030037959

    Abstract. In response to the growing problem of grade inflation in American undergraduate institutions, alternatives to GPA and GPA-based student assessment are discussed. One alternative summary, based on a Bayesian latent trait formulation, eliminates many of the inequities associated with GPA-based measures and has been proposed as a replacement for GPA-based class ranks at Duke University

    Adjusting GPA to Reflect Course Difficulty
    https://www.researchgate.net/publication/241813085_Adjusting_GPA_to_Reflect_Course_Difficulty

    Abstract
    The computation of Graduate Point Average (GPA) incorrectly assumes that grades are comparable across courses and instructors. GPA overstates the performance of students who elect an “easier” course of study relative to those who choose a more “difficult” course of study. This paper proposes a method of adjusting GPA and applies it to data from one cohort of undergraduates at Carnegie Mellon University. Adjusted GPAs are more highly correlated with students’ high school Grade Point Average and with SAT scores than are the raw GPAs or GPAs adjusted using a prominent alternative method, Item Response Theory. A survey of students finds that the new methods’ estimates of relative course difficulty are consistent with students’ perceptions of relative course difficulty.

    This 2017 article has a short discussion of that research: https://www.1843magazine.com/ideas/the-daily/the-extraordinary-silliness-of-american-college-grading

    P.S. I think restriction of range is more of an issue than SLODR in most cases.

  388. JLK says:
    @DFH

    But Vocabulary has a smaller gap between black and white scores than other subtests

    I’d like to update my answer on this. Most of what I’ve read in the past has suggested that your statement is correct, including this open letter written by Jensen:

    blacks do better on the culture-loaded subtests of vocabulary, general information, and verbal comprehension than on the nonverbal performance tests such as the block designs. Just the opposite is true for such minorities as Orientals, Mexican-Americans, Indians, and Puerto Ricans.

    Based on statements like that, my post #315 identified blacks as a group that are probably being helped by the double-weighted verbal, one math, no spatial format of the PSAT/SAT.

    However, this actual scientific paper written by Jensen seems to suggest otherwise, especially as shown in Table 2, where the black-white difference is far greater on vocabulary tests, the SAT-Verbal and arithmetic tests than it is in the spatial tests that are listed. Also, looking at the correlations of the g that is used in the Jensen paper to various spatial, arithmetic and verbal tests, it appears that Jensen’s g is quite a bit different than the g that was calculated by Deary based on the Salthouse studies.

    Based on this limited comparison of studies, the Deary g, which seems to give a greater weight to spatial intelligence than the Jensen g, actually looks fairer to blacks than the Jensen g.

    The PSAT/SAT format may as a result have been more discriminatory against blacks all along than generally thought, with the double-weight of the verbal portion hurting rather than helping. Re-weighting them to have equally important verbal, math and spatial sections might actually help blacks along with Asians, Native Americans, Mexicans. males and prospective STEM students.

    The Jensen study may be an outlier, and my observations are no doubt simplistic, because some types of spatial tests are more conceptual than others. But Table 2 of the Jensen study certainly seems to tell a different tale than his statement.

    The differences in correlations between various subtests in the Haier/Deary/Salthouse studies, the Jensen study and the Rimfeld et al. study that found a two factor solution to be superior to a singular g solution all seem to raise legitimate questions about whether college admissions and IQ tests that have been rationalized on the Jensen-style g are fair to various minorities.

  389. CanSpeccy says: • Website

    If you are going to use statistics, use the appropriate methods and apply them properly by all means. But it hardly seems to require difficult mathematical analysis to conclude that IQ tests measure nothing much other than ability to do IQ tests or, which is pretty much the same thing, to achieve academically in subjects that rely rather heavily on the limited range of verbal and numerical operations that feature in IQ tests.

    Certainly the idea that IQ tests measure some innate and unalterable mental capacity in all domains of cognition seems ridiculous in view of the evidence that schooling enhances IQ. Here, for example, it is shown that when the years of compulsory education in Norway were increased during the 1960’s from seven years to nine, IQ’s of students measured at age 19 increased by something like five points per year of additional schooling.

    Assuming that there are diminishing returns to IQ in the effect of schooling, such evidence seems to establish beyond reasonable doubt that the 20-point increase in average IQ of Americans over the last 100 years can be attributed largely if not entirely to an increase in duration of schooling. Likewise, it seems reasonable to infer that if the average number of years of education in sub-Saharan Africa increase from the current duration, said to be around three years, to something like the North America norm of 12 to 24 years, then sub-Saharan Africans will be found to be as innately intelligent as white Americans, or more so. More even, perhaps, than Chinese Americans or American Jews.

    One should not place too much reliance on a single research finding. However, the finding from many studies that there is essentially zero correlation between IQ and musical aptitude, a faculty dependent on many complex mental processes, lends support to the idea that IQ measures nothing much other than innate ability in a limited range of tasks as modified by Western-style education, and is in no way a valid measure of aptitude beyond the limited range of cognitive processes that the test directly measures.

    By obscuring the fact that IQ test results are not much more than a proxy for academic achievement, a conclusion evident to common sense evaluation of readily available data, the psychologists’ claim that difficult analytical methods are needed to properly understand IQ test data seems calculated to conceal from the public the otherwise plain fact that the IQists have no clothes.

    • Replies: @utu
  390. utu says:
    @CanSpeccy

    Instability of IQ scores:

    Test-retest gains in WAIS scores after four retest intervals.
    https://www.ncbi.nlm.nih.gov/pubmed/457898
    Administered the WAIS to 76 male college students on two occasions with a retest interval of either 1 week, 1 month, 2 months, or 4 months. Essentially all Verbal, Performance, and Full Scale IQs increased significantly on the retest. Increases in Verbal IQ for the four time intervals were 4.7, 1.8, 2.3, and .8 IQ points, respectively; the latter was the only nonsignificant gain found. The Performance IQ increased by 11.4, 9.8, 8.7, and 8.0 points for each subsequent time period, and the Full Scale IQ increased by 8.0, 5.7, 5.4, and 4.2 points for each respective time period. Test-retest correlations for the four intervals ranged from .91 to .72 for the Verbal IQ, .87 to .79 for the Performance IQ, and .94 to .74 for the Full Scale IQ. The results were pitted against similar test-retest studies, and clinical and research implications were discussed.

    Long-Term Stability of the Wechsler Intelligence Scale for Children- Third Edition
    http://edpsychassociates.com/Papers/WISC3LongStability%281998%29.pdf
    With an average test-retest interval of 2.87 years, test-retest reliabil· ity coefficients for the Verbal IQ. Performance IQ. and Full Scale IQ were .87, .87. and .91. respectively.

    • Replies: @James Thompson
  391. @utu

    The Wechsler Manual used to recommend at least 6 months between tests to avoid practice effects, and now recommends a year. This is because of stability of memory leads to higher scores as material is remembered. If the measures were truly unstable, they would not all have risen.
    All this well-known. The paper should have used periods longer than 6 months.

    • Replies: @CanSpeccy
    , @AaronB
  392. CanSpeccy says: • Website
    @James Thompson

    James,

    When you say that the positive effect on IQ of successive tests is because “material is remembered” (i.e., from one test to another), does that mean that the second test uses the same questions as the first, or merely questions of the same type? If the latter, as I assume, how does this effect of test preparation differ, really, from the effect of education in general, the duration of which is positively related to IQ, and which fades during periods of greater or lesser academic idleness as manifest by the summer-vacation decline in childrens’ aptitude test scores, a fall down that is modified by social class, reading habits, etc. ?

    What the evidence seems to indicate is that IQ tests measure not innate ability, but innate ability as modified by culture and life-style, the educational component of the traditional Western life-style, with its emphasis on numeracy and literacy, having a strong positive effect on test scores.

    That being the case, it seems certain that as Western civilization proceeds along its present trajectory of decline there will be continued declines in childrens’ academic achievement and IQ, the latter already said to be falling at the rate of seven points per generation. At that rate, we can expect IQ convergence between the West and sub-Saharan Africa by around 2050, as the Africans get more schooling, while our kids enrich the Silicon Valley tech billionaires by devoting ever more of their one and only lives to video games, porn, and social media.

    • Replies: @James Thompson
    , @AaronB
  393. @CanSpeccy

    The same test is used. It was not expected that these time consuming face to face Wechsler tests would be repeated frequently, but given that some would be given repeatedly, the six month gap was recommended. As I described, they now suggest one year is better, and warn that there will still be a learning effect.
    Group tests, on the other hand, often have an A and B version, because it was assumed that pupils might re-sit the tests.
    So, the retest effect on the Wechsler is not an education effect per se. It is a limitation brought about by the repeated use of the same material. An intelligence test is designed to be “school far”. That is, not to be very dependent on the detail that one student might learn at a particularly good school, but using material that would be expected to be known by all school children. “School near” tests like PISA and NAEP and other school examinations are far more influenced by curriculum and school variables.
    In broad terms all human creations are cultural. The few countries in which there is not an education system will have more difficulty with conceptual tasks. However, intelligence testing tries to minimize such effects. Therefore, it is interesting to see how PISA results differ from country IQ results.

  394. AaronB says:
    @James Thompson

    But if you have a group that is consistently practicing “IQ-like” tasks, perhaps as a lifestyle or cultural choice, their test scores will be higher without reflecting differences in innate ability between groups.

    The re-test effect makes clear that lifestyle and culture greatly impact IQ scores, which has been a common sense conclusion for some time, and better consistent with certain observed patterns between groups and countries than the heritability notion.

    Since culture is heritable, heritability measurements may be masking the re-test effect, and cultures change slowly and we only have IQ data for recent times

    • Replies: @AaronB
    , @James Thompson
  395. AaronB says:
    @CanSpeccy

    What the evidence seems to indicate is that IQ tests measure not innate ability, but innate ability as modified by culture and life-style, the educational component of the traditional Western life-style, with its emphasis on numeracy and literacy, having a strong positive effect on test scores.

    This seems most consistent with the mounting evidence. Problem is they cannot accept this. As utu says, they want to “reify” IQ – nothing less. They want it to be a stable “thing” – not some measure in flux, dependent on circumstance both cultural and political.

    I have been trying for some time to get them to grapple with the issue of how motivation (a lifestyle or cultural factor, perhaps) can affect IQ – common sense would suggest one cannot measure innate ability separate from motivation.

    The response has been the well known rhetorical device of dealing with a logical objection one cannot answer by studiously avoiding seeing it.

    One simply “assumes” that motivation is invariable across time and space. It is a primary assumption – because it must serve as a primary pillar of their tottering edifice.

    What makes it particularly ironic is that a key tenet of HBD is that groups differ in all traits. But the HBD movement, of course, seeks to reify races and ethnicities, and cannot accept that the attributes of races are also in flux.

    The HBD movement wants a world of “things” – to banish the Heraclitean flux. This suggests an exhaustion with the long period of Western innovation and a desire for stability. HBD and IQism are instincts towards a healthy stupidity, towards a self-protective blindness to reality, towards seeking refuge in a world of stable things. It is part of the “closing of the Western mind”.

    While this may be a response to the long period of innovation, it may be a needed sleep. The HBD and IQ reification movements may be a negative manifestation of the instinct for stability and sleep, but perhaps the way of meeting it is not refuting it – which is trivial and ineffective – but providing another outlet for the same desire and recognizing the West now need a good, rehabilitative nap.

    • Replies: @utu
  396. AaronB says:
    @AaronB

    Even if the same test is used, it makes no difference.

    Suppose two people take the test for the first time – one person may have been practicing these sorts of problems his whole life, the other not. Even two kids in modern society.

    In reality, no two people take take an IQ test for the first time at the same level of preparedness.

    The studious kid whose parents force him to study all day will have been practicing IQ style problem solving skills his whole life. The athletic kid may have spent little time practicing.

    So, the retest effect on the Wechsler is not an education effect per se. It is a limitation brought about by the repeated use of the same material.

    Obviously, an IQ test is designed to reflect the kinds of problem solving tasks one encounters during a modern Western education and society.

    Obviously someone who studies harder and takes school more seriously will be practicing the problem solving skills an IQ is designed to test.

    Clearly, the re-test effect means IQ style problem solving skills reflect practice – and thus motivation – as much as innate ability.

  397. @AaronB

    Don’t think that practicing intelligence tests boosts intelligence. Evidence against it, other than when the same test is used again, but little generalization to other tests.
    Education, if extended in later years, seems to add 2 points.
    Open to any evidence on this, but best result so far is Ritchie and Tucker Drob, who say 2-5 points, but believe the 2 point advantage is the most likely to be replicated.
    http://www.unz.com/jthompson/boost-your-iq

  398. Art says:

    Sorry but isn’t this getting why to complicated. Isn’t the answer to northern advancement – Darwin’s evolution? Survival of the fittest?

    To survive cold weather requires intelligence – not brawn. To survive cold weather requires planning, responsibility, and cooperation between individuals. In cold weather conditions the weak are culled. This produces a faster evolution.

    Those with intelligence genetics and cooperation genetics survive in cold weather.

  399. utu says:
    @AaronB

    From what I have read about Alfred Binet who having a great tradition of French philosophy of science behind him I doubt he would fall for the vulgar reductionism of Spearman that lead to the sleight of hand of reification. Unfortunately he did not live long. No different from Henri Poincare who if he lived longer would probably imparted a beneficial influence on modern physics and relativity in particular. Poincare was very deep into the philosophy of science (see also Pierre Duhem) and he had much deeper and more nuanced view than most and Einstein in particular.

    Here you may get an idea about Alfred Binet:

    “Binet would argue that ‘intelligence’­ whatever else it was­ could never be isolated from the actual experiences, circumstances, and personal associations of the individual in question”

    About the time he was ending his work at Salpetriere, he began experimenting with tests and puzzles on his young daughters Madeline and Alice and published three articles in 1890. He noted that the difference in reaction time of children compared to adults was contingent on whether the child was paying attention. The importance of the child’s state of mind and the influence of the environment on testing would remain significant to Binet (Fancher, 1985).

    He reported empirical data on his daughters as they grew and included a discussion of the different styles of intelligence between them in a 1902 publication entitled L’Etude Experimentale de l’Intelligence (The Experimental Study of Intelligence) (Zusne, 1975). He noted his daughters’ different styles when they were learning to walk and equated the difference to two equally intelligent adults using different ways to solve a problem:

    Mere numbers cannot bring out . . . the intimate essence of the experiment. This conviction comes naturally when one watches a subject at work . . . The experimenter judges what may be going on in [the subject’s] mind, and certainly feels difficulty in expressing all the oscillations of a thought in a simple, brutal number, which can have only a deceptive precision . . . We feel it necessary to insist that the suggestibility of a person cannot be expressed entirely in a number. (Fancher, 1985, pp. 62-63)

    If France in 19/20 century like Israel in 20 century had a hit squad killing enemy scientists perhaps we would be in a better shape now.

    • Replies: @AaronB
  400. AaronB says:
    @utu

    Interesting. It is not unusual for an intellectual movement to be taken over by people with a different agenda and lesser capacity. I wonder if in 50 years from now IQ will be put to uses quite unforeseen by us.

    Your quotes show a sensitive, nuanced position that is world’s removed from today’s talk on IQ – Binet seems to have a different goal than today’s IQ researchers; to genuinely understand a complex subject. The closing of the Western mind was not so advanced in his day, and he was still operating within the great tradition of European exploration and curiosity.

    Today’s researchers aren’t so keen to explore a fascinating topic and accept interesting facets of it as they appear but rather wish to “establish hard facts” (mostly preconcieved) within a narrow conceptual frame (few concepts acceptable) – a very different agenda, and lacking in genuine curiosity.

    Taleb btw is a huge fan of Poincare, and I’ve been meaning to read about him.

    • Replies: @utu
  401. utu says:
    @AaronB

    More on Binet.

    Binet believed intelligence was a multifaceted psychological faculties which were tied together in a real world and controlled by practical judgment.

    There is in intelligence, it seems to us, a fundamental agency the lack or alteration of which has the greatest importance for practical life; that is judgment, otherwise known as good sense, practical sense, initiative, or the faculty of adapting oneself…Compared to judgment the rest of the psychology of the intellect seems of little importance. (Fancher, 1985, p. 74)

    His 1909 publication Idees Modernes sur les Enfants (Modern Ideas about Children) presented ideas for improving a child’s intelligence, even at the lowest levels where he felt attention could be expanded. He believed individuals did have a mental ceiling, but few people ever approached that limit. Clark noted (1979) that Binet’s belief that intelligence was educable would not again be sounded until the 1960s and that many of his article and speeches could be considered radical even today.

    In 1909 Binet and Simon listed the following three criteria for intelligent thought: la direction, the taking and maintaining of a given mental set; l’adaptation, the adaptation of thought for the purpose of obtaining a given end; and la critique, the taking of a critical attitude toward one’s thought, and correcting it where necessary (Carroll, 1982).

    Galton and Binet both died in the same year. Galton was an old man and had long ceased working; Binet was 54 and at the height of his career. Despite the 277 publications credited to him (Dennis, 1954 cited in Albert, 1983), Binet did not leave behind many followers and his techniques were soon applied with “Galtonian” ways (Fancher, 1985).

    You made me curious about Taleb. What does he find in Poincare? If you check Poincare see also Pierre Duhem. He ha interesting observations on English science and German science.

    • Replies: @AaronB
  402. CanSpeccy says: • Website

    Taleb seems to be something of a jerk, but his third point:

    If you want to detect how someone fares at a task, … make him/her do that task

    seems sound.

    James argues that the point is good in theory but expensive to apply. Yet most employers follow Taleb’s approach, which is not necessarily that difficult. Employers look for qualifications, either school grades or past employment responsibilities and achievements, or they devise their own job-specific aptitude test. That approach seems generally to work. When I was hiring STEM graduates I selected from those in the top ten percent of their graduating class. They were invariable good. They knew their subject, they were conscientious, they were able to handle a wide range of tasks as might be expected of those having scored A’s in just about everything, and they acquired new skills quickly.

    Not everyone can be hired on the basis of academic achievement, but it is generally not difficult to devise a job-related aptitude test, whether it be for bus drivers, clerical workers, or whatever, that is more useful than finding how readily candidates solve anagrams, matches geometric patterns, or deduces the next item in a numerical sequence.

    If IQ score was more useful than track record in selecting good workers, most companies would use them, yet few do. Google may be an exception, but look at the results: the James Damore defenestration, damage from the fallout still to be fully calculated; the fiddling with search results in the interests of the SJW agenda, the trust in Google’s search results thus damaged; their compliance with the demands of the Chinese government for censorship, their readiness to work with tyrants thus demonstrated; all the while user confidence, as reflected in Google’s share of Internet search, diminishing. Hardly the profile of a smart company. And now they say that they never will be able to make autonomous driving work under difficult driving conditions. Oh, great: just watch out for those Google Waymo vehicles when their’s black ice or mist patches on the highway.

    • Replies: @James Thompson
  403. AaronB says:
    @utu

    I read Taleb many years ago, and remember Poincare was one of his heroes. I found some relevant quotes from Taleb on the web –

    Henri Poincaré, in spite of his fame, is regularly considered to be an undervalued scientific thinker, given that it took close to a century for some of his ideas to be appreciated. He was perhaps the last great thinking mathematician (or possibly the reverse, a mathematical thinker). Every time I see a T-shirt bearing the picture of the modern icon Albert Einstein,

    I cannot help thinking of Poincaré—Einstein is worthy of our reverence, but he has displaced many others. There is so little room in our consciousness; it is winner-take-all up there.

    Again, Poincaré is in a class by himself. I recall my father recommending Poincaré’s essays, not just for their scientific content, but for the quality of his French prose. The grand master wrote these wonders as serialized articles and composed them like extemporaneous speeches. As in every masterpiece, you see a mixture of repetitions, digressions, everything a « me too » editor with a prepackaged mind would condemn—but these make his text even more readable owing to an iron consistency of thought.

    Poincaré became a prolific essayist in his thirties.

    I will repeat that Poincaré was the true kind of philosopher of science: his philosophizing came from his witnessing the limits of the subject itself, which is what true philosophy is all about. I love to tick off French literary intellectuals by naming Poincaré as my favorite French philosopher.

    « Him a philosophe? What do you mean, monsieur? » It is always frustrating to explain to people that the thinkers they put on the pedestals, such as Henri Bergson or Jean-Paul Sartre, are largely the result of fashion production and can’t come close to Poincaré in terms of sheer influence that will continue for centuries to come.

    Many claim that Poincaré figured out relativity before Einstein—and that Einstein got the idea from him— but that he did not make a big deal out of it. These claims are naturally made by the French, but there seems to be some validation from Einstein’s friend and biographer Abraham Pais. Poincaré was too aristocratic in both background and demeanour to complain about the ownership of a result.

    Poincaré was the first known big-gun mathematician to understand and explain that there are fundamental limits to our equations. He introduced nonlinearities, small effects that can lead to severe consequences, an idea that later became popular, perhaps a bit too popular, as chaos theory.

    What’s so poisonous about this popularity? Because Poincaré’s entire point is about the limits that nonlinearities put on forecasting; they are not an invitation to use mathematical techniques to make extended forecasts. Mathematics can show us its own limits rather clearly.

    I like Binet the more I read.

  404. Sparkon says:

    Binet: “…initiative…”

    At last! Only 411 comments to get there.

    But what about curiosity, or determination?

    These are valuable personal characteristics or attributes, but hard to test for, no?

    Here’s another: originality.

    Pro tip/hint: Most people are copycats.

    I call it: Monkey see, monkey do.

    • Replies: @CanSpeccy
  405. @CanSpeccy

    If you select the top 10% of STEM graduates from high quality universities, the IQ test has already been given. Always use short-cuts if you can.

    • Replies: @CanSpeccy
  406. CanSpeccy says: • Website
    @James Thompson

    If you select the top 10% of STEM graduates from high quality universities, the IQ test has already been given. Always use short-cuts if you can.

    Not sure about that. When I was an undergraduate, a star of our university debating society, a red-haired Eng. Lit. major of Czech extraction, was overheard in the cafeteria line-up to declare, “I just took an IQ test and scorerd 95.”

    Whether he graduated in the top ten percent of his class, I don’t know, but he might have. As Alfred Binet commented regarding his intelligence scale:

    This scale properly speaking does not permit the measure of the intelligence, because intellectual qualities are not superposable, and therefore cannot be measured as linear surfaces are measured, but are on the contrary, a classification, a hierarchy among diverse intelligences; and for the necessities of practice this classification is equivalent to a measure. (Binet and Simon, 1980, pp. 40–41)

    There seems no reason to suppose, therefore, that an IQ of 95 is inconsistent with the ability to graduate at a good university in the top 10% in a particular area of specialization.

    Moreover, there are different ways of getting good grades that don’t all mean the same thing. I have an older sister who won a big university scholarship on the strength of, among other things, A grades in physics, a subject in which she showed little interest and about which she seemed to have little comprehension. This was no impediment to examination success, however, since, as she remarked after the exam, “I could see the pages of the textbook before my eyes”. She was certainly a bright student, but just not the same way a good physics student is.

    This article by Joel Mitchell, of the School of Psychology at the University of Sidney, offers a learned critique of the idea that IQ is a measure of intelligence.

    In a comment, hitherto unremarked upon, Alfred Binet, well known for constructing the first intelligence scale, claimed that his scale did not measure intelligence, but only enabled classification with respect to a hierarchy of intellectual qualities. Attempting to understand the reasoning behind this comment leads to an historical excursion, beginning with the ancient mathematician, Euclid and ending with the modern French philosopher, Henri Bergson. As Euclid explained (Heath, 1908), magnitudes constituting a given quantitative attribute are all of the same kind (i.e., homogeneous), but his criterion covered only extensive magnitudes. Duns Scotus (Cross, 1998) included intensive magnitudes by considering differences, which raised the possibility (later considered by Sutherland, 2004) of ordered attributes with heterogeneous differences between degrees (“heterogeneous orders”). Of necessity, such attributes are non-measurable. Subsequently, this became a basis for the “quantity objection” to psychological measurement, as developed first by Tannery (1875a,b) and then by Bergson (1889). It follows that for attributes investigated in science, there are three structural possibilities: (1) classificatory attributes (with heterogeneous differences between categories); (2) heterogeneous orders (with heterogeneous differences between degrees); and (3) quantitative attributes (with thoroughly homogeneous differences between magnitudes). Measurement is possible only with attributes of kind (3) and, as far as we know, psychological attributes are exclusively of kinds (1) or (2). However, contrary to the known facts, psychometricians, for their own special reasons insist that test scores provide measurements

    and

    Anyone who knows what scientific measurement is, also knows that psychometric testing is not measurement in the same sense as that term is used in physical science to describe assessment of quantitative attributes like distance, mass, or temperature. While some psychometricians realize this, most do not and they typically regard tests as instruments of scientific measurement.

    Mitchell’s critique of psychometrics is worthy of consideration in its entirety.

    • Replies: @res
  407. CanSpeccy says: • Website
    @Sparkon

    Here’s another: originality.

    That IQ is a measure neither initiative nor originality is perhaps why the US army uses it to evaluate recruits. The army sure don’t want no smart arse types figuring how to avoid getting killed as required.

  408. res says:
    @CanSpeccy

    There seems no reason to suppose, therefore, that an IQ of 95 is inconsistent with the ability to graduate at a good university in the top 10% in a particular area of specialization.

    If you honestly believe this it explains a great deal. Thanks.

    • Replies: @James Thompson
    , @CanSpeccy
  409. @res

    Agree. “Overheard in a cafeteria” hardly evidence.

    • Replies: @CanSpeccy
  410. CanSpeccy says: • Website
    @James Thompson

    Agree. “Overheard in a cafeteria” hardly evidence.

    LOL.

    But Mitchell’s critique (and Binet’s) of the idea that one can measure intelligence on a linear scale, seems compelling to me, and supports the notion that an IQ of 95 may not be a bar — nor an IQ of 195 be a guarantee of capacity — to graduating in the top 10% of an Eng. Lit. class.

    Would depend of course on the IQ range of the class, and with more than half the population receiving a “higher” education, the average IQ of many a class cannot be much above the population mean. Indeed, some must surely be below the population mean.

    • Replies: @James Thompson
  411. CanSpeccy says: • Website
    @res

    Look, if you have a scale based on non-commensurable variables — which is precisely how Binet, described the test that he invented — a high score on the scale is not inconsistent with a low score on one or a small number of the incommensurable characteristics upon which the overall test result, or IQ, is based. Conversely, a low score on the scale is not inconsistent with a high score on one or a small number of characteristics. Everyone knows that there are mathematical savants, morons for example, who can factor huge numbers in their heads, or recite pi to a billion digits. So what is so inconceivable about a literary or verbal savant, or in the case of what one might call a near-savant, a person with an IQ of 95, who nevertheless talks and writes better than 90% of his Eng. Lit. class-mates?

    • Replies: @res
  412. @CanSpeccy

    As regards Binet, I like his comments. A test score is not a summary of the steps taken to solve a problem, just a description of the result. Buz Hunt was critical of IQ scores on that basis, but I think that is expecting too much, and not really what one expects of a score. All examinations would fail that criteria.
    I do not wish to over-sell IQ, but the personality measures on which I placed hopes four decades ago have not contributed as much as I expected. Currently, despite measurement concerns,IQ is the front runner. I would prefer a ratio scale a la SS Stevens, but one can get that with Digit Span and Digit Symbol. I really don’t see this as a deal breaker. Odd, I know, that one can get so powerful a predictor in two minutes.

    • Replies: @CanSpeccy
  413. res says:

    I would prefer a ratio scale a la SS Stevens, but one can get that with Digit Span and Digit Symbol.

    Could you please elaborate on this? Have you covered it in an earlier post?

    In particular I am interested in how that converts to a ratio scale and how your two minute predictor compares to Wordsum overall.

    • Replies: @James Thompson
  414. res says:
    @CanSpeccy

    Look, if you think finding a one in a million case says anything about the vast majority of people then there really isn’t anything more to say.

    And try finding a real documented case rather than a bundle of hearsay and speculation.

    • Replies: @CanSpeccy
  415. JLK says:

    Chances are, most of the policy experts did the best they could at the time. I’m still getting around to my singular g critique, but as I pointed out earlier, g and Bolshevism aren’t as dissociated as some may have you think.

    Don’t discount the danger of the smoke-filled rooms.

    • Replies: @utu
  416. CanSpeccy says: • Website
    @res

    And try finding a real documented case rather than a bundle of hearsay and speculation.

    There is a difficulty here, in that one must rely on the work of an IQ-ist, and as seems evident from the comments here, IQ-ists are consistent in their determination to prove the non-existence of such a real documented case.

    However, given the low correlations observed among scores for individual cognitive capacities, the emergence of individuals of the kind I described is inevitable. No doubt one of the clever people here, Utu for example, could, given a Pearsonian intercorrelation matrix for cognitive capacities, tell you the frequency with which one might expect to find someone with a verbal capacity at the 90th percentile, versus a mean for all cognitive capacities at the 40th percentile. I doubt that the odds are as low as one in a million, as you assert. But even if they are, my case is proved.

    • Replies: @res
    , @JLK
  417. CanSpeccy says: • Website
    @James Thompson

    I do not wish to over-sell IQ, but the personality measures on which I placed hopes four decades ago have not contributed as much as I expected. Currently, despite measurement concerns,IQ is the front runner.

    The front runner in defining the human intellect with one number!

    • Replies: @JLK
  418. utu says:
    @JLK

    Don’t keep us up in suspense for too long.

    • Replies: @JLK
  419. JLK says:
    @utu

    I was actually following up on my last past in the second thread. I’m surprised that you haven’t followed up on the Rimfeld study that I linked, as it served your arguments well. You were out in front on this.

    • Replies: @utu
  420. res says:
    @CanSpeccy

    I doubt that the odds are as low as one in a million, as you assert.

    I was wondering if you would call me out on that bit of hyperbole ; ) In all seriousness I’d guess it’s more like something between 1 in 1,000 and 1 in 100,000. Much depends on how you define “good university.”

    find someone with a verbal capacity at the 90th percentile, versus a mean for all cognitive capacities at the 40th percentile.

    Remember that you asserted “ability to graduate at a good university in the top 10% in a particular area of specialization”. That is a much higher bar than “verbal capacity at the 90th percentile.” Most of your “good university” is going to meet that 90th percentile verbal mark and then you are taking the top 10% of them.

    And as far as “proving your case” goes, see the first paragraph (which you did not quote) of my comment you were responding to.

  421. utu says:
    @JLK

    Sorry that I did not acknowledge it but I looked it up after you linked. I did not know of Rimfeld before. I may comment later if I get a chance to read it 2nd time.

  422. JLK says:
    @CanSpeccy

    tell you the frequency with which one might expect to find someone with a verbal capacity at the 90th percentile, versus a mean for all cognitive capacities at the 40th percentile.

    I’m not sure about the magnitude, but such skews no doubt exist. High verbals often have a head start over high spatials and high maths because their skills are more readily apparent to those around them. Glib salespeople usually make a better first impression than the phlegmatic engineers who design the products, and often more money, to the chagrin of the latter.

    Which skills have the most social and economic utility? It would be a great topic for an article, and the balance of the arguments may have shifted over the years as the Internet changes the way things are done.

    • Replies: @res
  423. @res

    Search bar out of order at the moment.
    Don’t have a Wordsum / 2 min predictor correlation, but should have one!

  424. res says:
    @JLK

    This blog post from Steve Hsu is an interesting take on this topic: http://infoproc.blogspot.com/2011/06/high-v-low-m.html

    Here is a graphic from SMPY (cohort 2, 1 in 200 level) showing typical M-V-S (S sign and magnitude are represented by the arrows) profiles for different majors and occupations.

    This is a nonrepresentative group of people. From Steve’s comments at the end of http://infoproc.blogspot.com/2011/06/human-capital-mongering-m-v-s-profiles.html

    Because of the selection criteria I wouldn’t be surprised if the SDs are large in this population. Many of the SMPY qualifiers could have relatively average V scores and vice versa for SVPY. So the variation between the highest and lowest scores in each ability could be larger than in the general population.

    • Agree: JLK
    • Replies: @JLK
  425. JLK says:
    @CanSpeccy

    The front runner in defining the human intellect with one number!

    It’s pretty clear to me that at least a few additional mental faculties exist that are not being probed by conventional aptitude tasks. Most of the reasoning questions on IQ tests are short, discrete and solvable by linear thinking. They are a completely different endeavor from the work of, for example, an intelligence analyst who has to read hundreds of documents, assimilate the information and synthesize analyses.

    In such analytical fields, I’ve noticed that some people are more accurate and efficient than others by orders of magnitude. Their brains seem to unconsciously organize the information into interrogable constellations without much effort, recognize patterns and associations, and spit out theories that it would take others much more time to arrive at using a series of linear comparisons. If they could do it at all.

    Another thing I’ve noticed is that some people have the ability to block read. They can glance at a paragraph or page for a second and know basically what it says without actually moving their eyes line by line.

    I’ve read that some geniuses have flashes of gestalt intuition. I’m not an expert in the mapping of the brain, but there is probably some unconscious processing going on in an area behind the prefrontal cortex, coupled with more than typically efficient connection between those two parts. A little bit like distributed computing, where the two computers work independently until there is an information dump. I think the subject is fascinating.

    • Replies: @CanSpeccy
  426. JLK says:
    @res

    That’s actually a very politic way for Hsu to represent it, as the social science/humanities/legal people will probably have a hard time figuring out how to read the Z-axis. 😉

    It speaks well for his emotional intelligence.

    Hsu’s Z-axis is essentially what is missing from the PSAT/SAT format, along with the compression of the X-axis.

  427. CanSpeccy says: • Website
    @JLK

    It’s pretty clear to me that at least a few additional mental faculties exist that are not being probed by conventional aptitude tasks.

    Which is what the dictionary definition of intelligence tells us: that there are many aspects to intelligence, or the synonymous, cleverness. These range from talking a blue streak, like any graduate of Oxford University, to being able to make a girl smile, or walk a tight-rope over Niagara Falls. Such diverse faculties must engage different mental processes, and indeed it appears that the brain is divided into a multitude of modules, each evolved through selection over countless generations to perform a specific task with relative accuracy and speed.

    And something similiar is now said to be evident in the emerging systems of artificial intelligence. The advancement of AI does not depend on applying to every problem just one process ever more rapidly. Rather, AI systems advance through innovations that, increasingly, the systems themselves create, evaluate and further evolve. That being the case, it will be meaningless to attempt to assess the power of AI systems on a linear scale. Systems will emerge to solve particular problems or sets of problems and it will make sense to evaluate them only in relation to their speed and accuracy on those tasks.

    Humans have a need to solve many diverse problems, so it may make some sense to evaluate their overall intellectual competence on a single scale that combines scores on incommensurable variables. But one should not lose sight of the independent variability of the separate cognitive capabilities, which at the extreme can be vast, as in the case of savantism. And many a genius has more than a faint resemblance to an idiot savant, not withstanding the occasional emergence of astonishing all-rounders.

    • Replies: @CanSpeccy
    , @JLK
  428. CanSpeccy says:
    @CanSpeccy

    AI systems advance through innovations that, increasingly, the systems themselves create, evaluate and further evolve.

    What this means is that the development of AI has become a process of in silico evolution taking place at close to the speed of light in a “brain” with essentially limitless resources — or at least with the resources of Google, Amazon, DARPA and the Chinese Government.

    The result will be many sorts of intelligence each adapted to particular objectives, some objectives conceived by AI itself as necessary to its attainment of human-selected goal. In some cases, AI will likely formulate objectives we could not have conceived of and would never have agreed to. Maybe that’s how the arms race will end: an AI directed simultaneous nuclear strike on Washington, Moscow and Beijing, with a stern warning to all other capitals to abandon hope of advantage by military means.

  429. JLK says:
    @CanSpeccy

    And many a genius has more than a faint resemblance to an idiot savant, not withstanding the occasional emergence of astonishing all-rounders.

    Some of those all around skills I could live without. At least Rain-man wasn’t a liar.

  430. @J

    Yes, but that would be assortive mating anyway.

  431. “Then of course there is the question of ability being measurable only in relation to motive and goal, and unless all individuals and groups can be said to be identical in these respects (which would curiously contradict the HBD premise that groups differ in all traits), measuring ability is problematic…the elephant in the room no one wishes to address…

    Anyways this is just the tip of the iceberg…

    The move towards abstract thinking and vast generalizations was historically useful for a period in time, but today, we are in a position where constructing a more intelligent approach to reality would mean a move away from the abstract and general and towards the concrete, and the loss in precision would be compensated for by a gain in richness of detail and better conformability to actual patterns in reality, and the inability to establish eternal laws and find Platonic ideas would be offset by a realistic flexibility in responding to reality.”
    AaronB

    I’ll add something to this wonderfully intelligent observation.

    Most of the people I have met with a high IQ are kind of stupid.

  432. Novelists, for instance, are not known for having high IQs.

    Now, I know all you sophomoric fools are Asperger “STEM” worshipers (even using the acronym “STEM”–except to mock it–is a sign of idiocy), and you think novels and novelists are just dumb, lazy, goldbricking shirkers, who have the temerity to refuse to participate in the general crusade of all mankind to destroy all mankind and the world it lives in by means of (“STEM”) technology.

    So you try to write a novel that anybody would want to read.

    Then you’ll know where the clown hat belongs. On your own heads.

  433. In my experience, most “scientists” are rather simple-minded in their understanding of what “Science” is. They actually think it exists. I mean, like a pear or a dumbass exists.

Current Commenter
says:

Leave a Reply - Comments on articles more than two weeks old will be judged much more strictly on quality and tone


 Remember My InformationWhy?
 Email Replies to my Comment
Submitted comments become the property of The Unz Review and may be republished elsewhere at the sole discretion of the latter
Subscribe to This Comment Thread via RSS Subscribe to All James Thompson Comments via RSS