Taleb is wrong about IQ (2019)
jsmp.dk
jsmp.dk
Taleb's original medium post deals with other issues aside from correlation with on-job performance and wealth. Specifically the purported correlation between IQ and race. Therefore, the claim that 'Taleb is wrong about IQ' doesn't seem to follow from the limited analysis provided.
The article makes no attempt to make any analysis about differences in wealth or education. It is VERY SPECIFICALLY addressing some of Taleb's claims about IQ and is VERY SPECIFIC that that is the goal.
Not sure this is the case at all. The conclusion states:
> Conclusion
> All the claims from the article that I looked at, that can be interpreted as something specific and tested in a real data set, turned out not to be correct.
It's true that Taleb's article is a bit of a rant, in his specific style, and that it might not get its quantitative claims precisely right. However, I think that his core points are valid and that the current paper, by focusing on relatively minor quantitative claims, doesn't actually provide relevant counter-arguments.
To be clear, not following the scientific method does not mean your hypothesis must be wrong. But to convince the community at large, that's what you should do.
- Let us return to the point of the correlation test-retest. Unlike measurements of height or wealth, which carry a tiny relative error, many people get yuugely different results for the same IQ test (I mean the same person!)
- National IQ is a Fraud. From engaging participants (who throw buzzwords at you), I realized that the concept has huge variance, enough to be uninformative. [...] Some people use National IQ as a basis for genetic differences: it doesn’t explain the sharp changes in Ireland and Croatia upon European integration, or, in the other direction, the difference between Israeli and U.S. Ashkenazis.
- If, as psychologists show (see figure) MDs and academics tend to have a higher “IQ” that is slightly informative (higher, but on a noisy average), it is largely because to get into schools you need to score on a test similar to “IQ”. The mere presence of such a filter increases the visible mean and lower the visible variance.
> But to convince the community at large, that's what you should do.
I don't agree. I don't think that only scientific (or at least scientific-looking) arguments should be considered valid. Why would that be the case? There are a lot of empirical observations that we can draw useful knowledge from, without resorting to mathematical theories and quantifiable experiments. Not to mention that science itself is far less of a rigorous field than it may seem at first sight.
Because we humans are excellent at convincing ourselves of seeing or believing things that are simple untrue[0]. It turns out that numbers - while they can still be messed with knowingly or unknowingly - are a very useful tool for not only combating your own biases but also convincing others that our conclusions avert these.
Empirical observations are certainly immensely useful - for forming a hypothesis. If the hypothesis can successfully predict certain non-trivial effects, that elevates it over hypotheses that cannot do so.
> I don't think that only scientific (or at least scientific-looking) arguments should be considered valid.
As I said, arguments can be valid without being scientific. But mere human perception is a very fallible thing to build research, results and, eventually, policy on.
There are important caveats to what the author's done here:
1. Taleb's real point is that IQ is mostly useful for avoiding (extreme) incompetence rather than selecting for competence. Taleb constructed his argument using statistics which are either incorrect or underspecified, which invites the kind of analysis the author made. But that also means the author isn't substantially engaging with what Taleb means, because Taleb is abrasive and doesn't care to be specific.
2. You should be skeptical of any statistical analysis which leads with Pearson's r (or Spearman's rho, or whatever correlative measure) or the associated R^2. It is a top line measure which is useful but incomplete. You may miss out on the nonlinear relationship or nonperiodic relationship entirely, either of which could hold most of the variance. There are also many ways in which the variance of two things can be associated without there being a predictive relationship. Simply stated: as a statistician would tell you, if all we needed was R^2 then there would be no papers, only abstracts.
A little bit of statistics can be a dangerous thing. I'm not defending Taleb per se; rather I'm saying that both sides are arguing past each other and we probably shouldn't draw a conclusion from either.
It seems we could put this a little stronger, as the article claims to refute all the quantifiable evidence for the core point. As far as I can tell, it does so.
I also don't understand how your first point works. It contradicts itself. Look, either statistics (because this is a problem with statistics, not with IQ specifically) can predict individuals or not. So either IQ is useful in classifying both success and failure. Or it can't be used for either. Generally using statistical values for predicting individual success or failure is, first, very unreliable in the individual case, and second, not considered "just", for good reason. If you statistically show a group of 10 people has a 90% chance of being criminal, it is considered impossibly hard to pick out the innocent from among them using statistics. Therefore the correct way to respond to statistics is to group people by IQ bands, not to give specific individuals advantages (which, granted, is often done). Of course, this will magnify differences over time, which you may not consider socially just.
Your second argument essentially states that "measures that don't take time into account ... don't take time into account". It is wrong insofar that IQ usually does take time (time since birth) into account, but nothing else. That's not a coincidence, that's because they tried other factors and they didn't matter, but obviously a tautology is always right. The point is wrong because it's useless, not because it's untrue. There are many other factors that could be evaluated against IQ, and there has been a ton of research that did so, and never found much correlation.
And frankly, it boils down to this: "being good at X" correlates across X. That's to a large extent what IQ is. Everybody has seen that incompetence extends beyond any individual skill an individual might or might not have. Research shows that competence does too. It makes sense to some extent: having exceptional skill at one particular thing allows you to advance quicker when it comes to other skills. When those skills are related this is considered "trivially true" (e.g. learning piano to learn violin quicker), but what's harder to see is that this is in reality almost universally true, and this is in fact what our school system turns out to be based upon. Skills you learn in school aren't generally meant to be generally useful, but to allow you to learn "real" skills that can't be effectively learned in school much faster.
No, statistical relationships are not symmetric in general. You shouldn't expect equal absolute predictive power for events two sigma from the mean in opposite directions if the random variable doesn't have a normal distribution. IQ is normally distributed, but many obvious and conventional measures of success are clearly not.
> Your second argument essentially states that "measures that don't take time into account ... don't take time into account".
This isn't what I'm intending my argument to "essentially state". My second argument is just that drawing conclusions from "top line" statistics like R^2 is dangerous and potentially misleading.
Usually cognitive abilities have a step skew to the right. And even IQ fails the sniff test for normal distributions as you can imagine that the difference between 70 and 100 is far greater then the difference between 100 and 130.
I can use that information to predict a wide variety of things with very high confidence, but I can’t use it to tell you how tall any individual human will become. That doesn’t make it wrong nor useless.
IIRC, one thing that we have studied pretty thoroughly, is that IQ — that is, the number IQ tests spit out — goes up when people are under the effects of stimulant drugs during the test. (Even neurotypical people, whose executive functioning is just fine without such drugs.)
This is one major reason that IQ is considered a separate thing from the g factor.
Essentially, g factor is "trait intelligence", while IQ is "state intelligence."
While IQ might be altered by nurture/environment/situation, g factor cannot be, by definition. g factor is supposed to be what you get down to when you remove all the state-intelligence confounders. That's why it's a more interesting number to study.
-----
To couch IQ vs. g factor in a cute analogy: think of ability scores in Dungeons & Dragons, vs. savings throws in that same game.
g factor is essentially your INT ability score. It's defined at character creation, and it's effectively static (for regular humans; the only things that alter it in the game are supernatural.)
But you don't use your INT ability-score directly for anything in the game. Instead, you use your INT modifier — which has your INT ability-score as its base value, but then aggregates on top all other temporary or semi-permanent bonuses/maluses from magical effects, equipment, etc.
If you do a number of "saving throws vs. INT" at a calibrated difficulty level, and average them together, what you'll get is a measure of your character's current INT modifier, not their INT ability score. Their IQ, not their g factor.
(The analogy is surprisingly even tighter than this: people who go into INT-based classes get a proficiency bonus on their INT modifier, somewhat like people who do lateral-thinking puzzles all day get a proficiency bonus on IQ tests. And there are situational modifiers to difficulty, just like there are confounders in the questions to an IQ test.)
But all that and a lot more is discussed in Stephen J. Gould's The Mismeasurement of Man, a text to read before reading random posts with graphs on the internets.
I was actually looking into some of Gould's work just a few days ago; mostly on punctuated equilibria, but The Mismeasurement of Man is on my reading list too :D
I don’t see that claim addressed very well here.
To wit, consider how most famous serial killers have high or very high IQ (120 and above). The reason is simple: you don't get to kill that many people if you're dumb ... because you get caught quickly.
Go below 100 though and very low IQs seem to be predictive alone of various poor outcomes.
So it’s more useful as a diagnostic for deficiency than a predictor of exceptionality.
Nurture accounts for 20-43% of the differences in adult IQ scores:
"Twin studies of adult individuals have found a heritability of IQ between 57% and 73% with the most recent studies showing heritability for IQ as high as 80%."
Invert, always invert.
Looking at 'nurture' as a positive action may not cover all the cases of increased IQ. Generally, yes, I think that investing more time into others is going to increase the IQ score (for many reasons). I think this is a great thing we all should invest in.
However, there is the edge case where a lack of nurture will also increase it. The adversity is overcome and the person becomes more 'IQ-y' as a result. Ramanujan come to mind as a perfect example of a person overcoming large adversity to achieve great status.
These edge cases aren't particularly repeatable, of course, but they are instructive. These people are the very end of the 'long tail' of these statistical constructs and give great insight into what it takes besides pure luck.
https://en.wikipedia.org/wiki/Srinivasa_Ramanujan#Early_life
Of course these tests may be improved upon, but so far I haven't seen any solutions offered to account for children in my situation if quantitative measures of aptitude are eliminated.
The author performs a regression on the WLS dataset. I did not know what that is so I followed the links provided:
>> The Wisconsin Longitudinal Study (WLS) is a long-term study of a random sample of 10,317 men and women who graduated from Wisconsin high schools in 1957.
So, in other words, it is a small dataset that is not representative of most humans on the planet. It is also specifically populated by high school graduates which is likely to introduce some unknown bias to the dataset- and any results drawn from it in particular with regards to IQ.
In short, please don't abuse statistics, or if you do, don't do it in an article that starts by accusing another text of being "just a rant".
It's linked in the second paragraph
Remember kids, IQ was developed to predict which children would perform best in school so that they could be streamed, and that's what it's best at.
I don't think that the author is abusing statistics here at all, as the variance explained by IQ on grades is always pretty large (in psychology, 35% explained is super, super good), and if I really cared (i.e. you reply and ask for this) I could probably dig up another few datasets that show the same or higher association.
Assuming the R2 here is indeed correct how do you use this to make real life hiring or admission decisions ? Just weight it at 35% ? What do you do for the other 65% then ?
I am not doubting the validity of the correlation. I just want to know how one can use this info properly.
If I was interviewing an engineer that said something like this, I wouldn't hire them. It's somewhat unclear what you even think "random chance" means in a regression context.
Any model with non-zero R2 is better than random guessing.
I believe it was developed to predict those which would perform worse so they could be further examined for specific issues and extra help.
10k is definitely large enough to draw conclusions. I think your concern about generalization is much more relevant.
To your point about high school grads, though, it is absolutely worth noting that in 1957, “no child left behind” wasn’t a thing.
In 1950, 36% of whites and 13% of blacks over age 25 had a high school diploma.
In 2010, it was 92% and 84%.
Then, graduation was generally a vocational choice by household and student rather than a performance bar. Today, it’s an externalized requirement, largely for the system to be funded: you will be graduated despite yourself.
A few other diffs:
1954: Brown vs. Board of Education desegregation ordered
1951 spending per student: $215 for 155 days of school. 2010: $10,615 for 180 days of school. Average teacher salary then and now: $4K vs. $39K.
1955’s most popular youth book: Charlotte’s Web. 2012: Hunger Games.
I don't understand how selecting a non-representative sub-population controls for anything! Same goes if the data was from Maryland or New York City.
In any case, disentangling confounding factors and isolating the influence of IQ should be the goal of any research (though of course I can't ask that of an "analysis" published on a blog). If a dataset seems to reveal some deep relation between IQ and other outcomes because that dataset is free of confounding factors, that doesn't tell us what happens in the general case, where the confounding factors apply. In other words, there's a danger of finding what one is looking for because one is looking for it where it exists, when the goal of looking is to make sure it exists everywhere.
Sure wealth is an interesting thing to look at societally, but it offers no insight into human psychology.
Wealth is not a feature internal to an individual, it's something defined by a society/particular social structure and as such, it doesn't make a lot of sense to compare it to IQ (an individual difference variable measured by a test).
Wrong. The original goal of IQ testing was to predict which students needed special care. It is still very useful (albeit expensive) metric for that.
I just mentioned wealth here just to demonstrate that just because something correlates and predicts something, it doesn’t make it interesting. We know how wealth is inherited, and how people of wealth are privileged and generally more successful (at least if we measure success the same way as IQ proponents do). We know the mechanics of how wealth works and we know it is more societal then psychological. I took wealth as an example here just to weather the possibility the IQ is correlated and has this predictive power because of similar mechanics, i.e. not descriptive of psychology, but society.
You are entirely correct, turns out I misremembered.
Wealth is definitely not inherited the same way as IQ, in that it doesn't show any regression to the mean. We also know that adopted children from a lower-IQ family will show an IQ between their birth and adoptive parents, which is again very different from wealth.
I do agree that IQ is not a perfect metric, but can you describe why you think IQ is more reflective of society than anything else?
Assets/wealth is actually pretty uncorrelated with IQ among adults.
Saying that wealth is the same thing as IQ is a conservative position, IMO.
Why aren't those two sides of the same coin? I mean, smart people can generally figure out how to accrue and maintain wealth.
Someone will say by way of counter-example that, let's say the Kardashians are not smart. But they seem awfully clever at manipulating their publicity in ways that bring them wealth.
So it really seems to have more to do with whether one has a narrow conception of intelligence as being able to, say, find new theories in QCD or derive relativity, rather than being able to figure out how to accomplish things beneficial to themselves. Which might be D-K in effect, but who knows?
That said, if wealth and IQ correlates—which I’m not sure it does—you cannot say which causes which, or even if either causes the other. It is just as plausible that wealthy folks have better training taking tests which will lead to higher test scores in general, including IQ scores.
https://rpadgett.butler.edu/ps320/coursedocs/Richardson-Norg...
Now if you manage to convince me that there is a correlation, twin studies is not enough to convince me of the causation. Twin studies are controversial as they do not provide a random sample and have a number of inherit biases.
And even if you find a way to convince me that the causation is not wealth → IQ, it is still not enough to convince me that IQ causes wealth. You still have to eliminate any third variable that causes both.
This is all after you’ve convinced me that IQ is a useful and interesting metric (see nibling comments), as well as useful and unbiased, all of which is up for debate in academic studies as of yet. Good luck.
I wasn't talking about Twin studies. It is a fact that among adopted kids their intelligence correlates almost as much with their biological parents as for non adopted kids. Adopted kids is such a big group that there is no lack of data.
We have, actually! It's just that they've all been implemented already. Iodine supplementation was huge. (Childhood Iodine deficiency results in an adulthood IQ deficit of ~15 points)
The next big IQ improvement looks like it might be iterated embryo selection [0] which is currently possible and will soon be practical, but has a whole host of ethical issues attached to it. (Is it right for the children of the rich to get another advantage like this? (Not a question for the Talebs of the world who think IQ is bunk, but one I'm sure many will be asking) Is throwing away dozens of viable embryos ethical? Are we 'playing god'? And further down the line, is using CRISPR-Cas9 to directly edit the genome of your descendants OK?)
Like the entirity of personel selection for many, many years was based on IQ like measures, and it was the base for all of the work ETS have done on standardised testing (item response theory and all that).
Can you explain why you do not believe IQ to be rigorous?
>Can you explain why you do not believe IQ to be rigorous?
That's a much longer and more complicated argument than what I'm saying, which is simply that IQ is not anywhere near the most essential nor the most studied question in psychology. Classical conditioning is another. Attachment and critical periods in developmental psychology have received a lot of attention. In particular, pretty much any effect which generalizes in a simple and reliable way from humans to animals is easier to experiment on and has a stronger evidence base than IQ: instinctual drift, childhood amnesia, most of the theory of perception, etc.
The actual content and priorities of psychology as a field of scientific research is quite different from the few topics which get debated with unreasonably high frequency on the Internet.
Full disclosure, I have a doctorate in psychology (but have been out of the field almost a decade now), so am well aware of the "actual" psychological research.
Personally, for my money, IQ lead to the development of most of modern psychometrics, which I think is probably psychology's largest contribution to intellectual knowledge, but that's probably my quant bias showing.
Sure, behaviourism has had a bunch of successes, but it's only necessary for psychology, and definitely not sufficient.
Classical conditioning I'll give you, but I definitely didn't hear much about either instinctual drift or childhood amnesia relative to IQ.
> theory of perception, Which one? are you talking about the information processing model or the gestalt models? I'd probably agree with you on Gestalt, but not on IP (which I think is the steam-engine metaphor of the mind of our time).
I'd agree that it's not the most essential question, but it's definitely one of the most studied.
Even if you stumble on such facts accidentally, it can very well damage your academic prospects (and sometimes non-academic prospects as well ... in places like Silicon Valley).
Many years ago I had a friend who was studying an unrelated topic at University (he was trying to figure out how to test for future programming aptitude via a purely paper test). He was administering the paper tests to students with no programming experience before their first programming class, and then studying the correlation between various paper tests and final grades in programming projects in courses (including grad courses) over time.
He found some quite good paper tests that correlated quite nicely with future grades on programming projects ... but a faculty advisor pointed out that the results of the paper test were also strongly correlated with some politically sensitive student categories. My friend abandoned the project because when it was pointed out like that, it was quite clear he was producing results that could be exploited to create a selection criteria he was not comfortable with endorsing.
[for reference, if using IQ as an indicator for performance, 2 standard deviations below the mean would be 70 IQ]
Taleb: "The argument by psychologists to make IQ useful is of the sort: who would you like to do brain surgery on you/who would you hire in your company/who would you recommend, someone with a 90 IQ or one with 130 is ...academic. Well, you pick people on task-specific performance, which should include some filtering. In the real world you interview people from their CV (not from some IQ number sent to you as in a thought experiment), and, once you have their CV, the 62 IQ fellow is naturally eliminated. So the only think for which IQ can select, the mentaly disabled, is already weeded out in real life: he/she can’t have a degree in engineering or medicine. Which explains why IQ is unnecessary and using it is risky because you miss out on the Einsteins and Feynmans."
Einstein and Feynman would certainly not be rejected if you automatically rejected people with below average IQs, so I don’t really get this part of the argument. We have no number for Einstein and a supposed 125 for Feynman but he did extremely well (the best in the nation) on Putnam, so his nonverbal (or at least “quant”) score was likely very high.
A piece of intuition is that if you select a population based on a criteria, the correlation between that criteria and the target in the selected subpopulation is usually greatly diminished, and sometimes reversed.
Imagine you know that being very tall is very good for basketball... so you select very tall people to be players. You will likely find that among the selected players height is not very correlated with performance, or even inversely correlated because your initial selection wiped the correlation out and potentially noise or other effects (like agility) which are correlated with being short begin to dominate the differences within that population.
This also holds for correlated traits-- e.g. if you select very tall players it may wipe out or reverse the correlation shoe-size has with performance.
As a result if you first select people for an intelligence requiring task specific capability and also select for IQ you're likely selecting for noise, and potentially selecting for people who are just really good at taking tests and which weren't actually as good at the task specific capability as a randomly selected person in your first selection would be.
I agree overall that task-specific selection is better, I was just questioning the example given.
So when people say "Feynman had iq of 125" it is very different from someone having 125 IQ today.
Think about the USA. The year and city where you're born/raised can outweigh your IQ when trying to predict income, life expectancy, etc.
Dyslexia is a good use of IQ testing; either you do worse than expected (because of the difficulty reading) or better (compared to your grades) and in either case further investigation could lead to a diagnosis and useful help.
How about people just be humble about how smart they are, and start using their "intelligence" to help others and be a good example.
Yeah it's a super popular trope among racists on the internet.
That said, I find the leftist rejoinder "IQ tests are racially biased" to be pretty unconvincing given that most of the tests are non-verbal. Much better to focus on the large confounding environmental factors than to deny there is any sort of intelligence gap at all.
When I was younger, I would have said that there's no such thing as innate intelligence.
If I'm better than some people at math or programming, that's entirely because they didn't pay attention. Or they didn't spend the same amount of time and effort on it as I did. Or they just aren't being careful, or aren't trying to really understand. They're doing something wrong, or they just don't care enough. That's what I would have said.
It's quite a comfortable view, really: Other people had their fair shot to be exactly where I am. They chose not to put in the time required, or their learning techniques were tactically sub-optimal. I was in school a bit before being nerdy was cool. Others certainly didn't show me any mercy when I was near the bottom of the pecking order in the realm of social interactions or athletics, so why should I go easy on those who are close to the bottom of the barrel when it comes to intellectual achievements or salary comparisons?
We humans play games for status. If you want me to accept that I come out near the bottom of some of those games, it's only fair for me to require you to acknowledge my superiority in others.
But I don't say it anymore.
I've met a number of people who just don't seem to be able to think in certain ways. Who will never be able to be good programmers, no matter how much training they get or how much time they spend.
There might be something innate to intelligence. It has real effects, so it should be measurable.
Is it IQ? I don't know. Let's call it, I don't know, INT (after the D&D stat).
What if there really is such a thing as INT? What if it really is something that people of some races, on average, have a greater amount of than people of other races?
Years ago, I'd have said "Of course that isn't the case, all humans are created equal."
Now I say "I don't know if this is the case. The evidence I'm familiar with is consistent with the possibility, but there are other explanations."
Let's suppose it's a simple physical fact that people of a certain race are (on average) taller. And it's a simple physical fact that (on average) being tall makes you better at basketball.
If you run a basketball team and you want to be successful, you will hire the best players without thinking about their race. And you will end up with a team where people of the tall race are over-represented relative to the population.
Now let's turn it around.
Let's suppose it's a simple physical fact that people of a certain race have (on average) higher INT. And it's a simple physical fact that (on average) higher INT makes you better at programming.
If you run a tech company and you want to be successful, you will hire the best programmers without thinking about their race. And you will end up with a team where people of the higher-INT race are over-represented relative to the population.
The world I observe is compatible with the explanation that people of some races have (on average) lower INT.
To be sure, there are other possible explanations: It could be bad schools, bad peer groups, bad home life. It could be living in older buildings with lead and asbestos leading to health problems that affect the brain.
Heck, there could be bona fide racists out there running tech companies who cackle maniacally as they throw an applicant's perfect resume in the trash simply because they're not a white male.
The point is, if you want to face reality rather than live in a comfortable fantasy, you have to confront the possibility that the racists are right about one thing: It's possible that some races really do have less intelligence than others.
It's just a possibility. But I haven't seen it conclusively disproven. If it's true, what should you personally do? What should we as a society do? I don't know. And it's kind of a terrifying question to ask, really.
Honestly, if it's true, I don't think it would be a truth that American society could handle. We would tear ourselves apart. Maybe that's what's happening now.
I think that the way we as humans decide to structure society determines whether life is fair, not natural circumstance.
Just because you are born smart does not mean you are born more deserving.
No amount of exercise will give me the garden hoses for veins that someone I know has, but I can still work to reach my own potential. Should I still be given a job that requires holding a jackhammer over my head for 8 hours a day because it's not fair?
Why is it socially acceptable to be taller or stronger or faster, but not to be smarter or more studious?
Exactly - "to deserve" is a nonsense notion in these contexts! Tall people in the NBA might lead to a more beneficial outcome (more interesting games?) but it does not arise because tall people are more "deserving" of being in the NBA.
Since we get to choose how to structure the human sphere of the "game" of life, it seems better to optimize for an outcome that is fair, ie. equally attentive of people's moral worth.
If that requires giving more $ to some who are more talented in order to better off everyone, then life does seem fair because our society is structured w.r.t. everyone's benefit.
If you are interested in knowing about the thing itself, and not just criticisms of the thing, I invite you to pick up, for example, Stuart Ritchie's book https://www.amazon.com/Intelligence-That-Matters-Stuart-Ritc...
And before you get upset about the title, the book is from a survey series called "All That Matters". It's not claiming that _intelligence_ is "all that matters".
I think most of the HN audience is well aware of the positive aspects of IQ. It's just that the power structure here protects the vocal minority who espouse solely the negative side. Like so much of media/social media/politics/culture/etc, it's censorship that tips the scale and skews discourse towards the extreme.
What do you consider it's positive values? All I'm aware of is how it's used to judge people. Unless you consider that a positive thing.
1. We can quantitatively study environmental factors that influence intelligence. For example, we know that even low concentrations of lead in the blood are linked to lower IQ. Knowing what can hurt our mental abilities allows us to protect ourselves via environmental policy.
2. It's not uncommon that a child who is inattentive at school is high IQ and is understimulated by going at the normal classroom pace. Testing allows identification and targeted enrichment opportunities.
Also IQ is a sub- component for general intelligence. There are so many other factors. Claiming it's useless or useful is pointless. It probably depends on how relevant that sub component is for the role and task at hand and therefore requires a case-based understanding.
It's like claiming that good body strength is useless. For a programmer that's probably true, but for a professional home mover or pyramid builder, results will be different.
This is the bit that is more important to me and something I dont see in the IQ debates. Some "smart" people are fast learners, others quick witted, some have great memory, others can distill lots of conflicting data, some are great communicators, others just work really hard. One IQ number is supposed to reflect all of that?
Essentially, each new test is normed against the previous tests, so the Stanford-Binet test would have been the first, and then when a new measure is developed, we'd administer both tests to a sample (hopefully many samples) and use the correlation between the tests to determine whether or not or new test worked.
It's not really a great method, but it's the best we've got for most individual difference variables.
I also tend to go off on inner rants like Feynman about round manholes, because the tests do not always seem to have unambiguous solutions.
They do a much better job of getting estimates of IQ which are robust to culture, which was a real problem for the Stanford-Binet ones (and lead to some belief that the IQ of poor black children was lower because they weren't familiar with some of the analogies).
If DeepMind developed a system that could ace Raven's Progressive Matrices questions, I don't think that we would say "machines are now as smart as human geniuses." Turing's Imitation Game still seems like a valid challenge to me but is perhaps too stringent. Someone who is not literate in English would fail to convince the questioner even if otherwise intelligent.
"General" human intelligence appears to have a very complicated and particular shape. If there is such a thing as universal intelligence (in the sense that there are universal constants), the boundaries of human intelligence in this space look like a gerrymandered voting district more than a symmetrical geometric object.
I'd like to see a test that correlates well with our colloquial notions of human intelligence but that also generalizes to non-humans (including machines). Maybe those are fundamentally incompatible goals. Maybe the outcome of AI research is finding that colloquial notions of intelligence resemble "real" universal intelligence only as closely as the intuitive physics of throwing and catching a ball resembles the unification of quantum mechanics and general relativity.
Many other types of tests were tried for this, because being able to predict people's performances has extremely important implications, but the ones that did not predict anything were forgotten, while the ones that worked, like the current generation of IQ tests, became widely accepted (despite being attacked every now and then by a few researches, as in this case - which is a good thing as we need to make sure our current theories are as good as they can be).
* IQ is fixed; a test at, say, 16 will give you your intelligence for the rest of your life. (16 is arbitrary because I am uncomfortable with psychological testing of young children, but you can make the argument to roll this back to 6 or 8 if you like.)
* IQ is hereditary.
* Skin color (and ethnicity in general) is correlated with IQ. Sex is correlated with IQ variance.
(Am i missing anything here?)
Here's a clarified from of Kantian categorical imperative: Would you like to live in a society based on those principles? (I forgot to add, "given human history".)
IQ testing is very valuable at the low end, to identify individuals that need some form of help. I, personally, am a little uncomfortable with other policy extensions.
I personally think that this re-evaluation is absolutely necessary, because the current model of social inequality, in which any differences between individuals and groups are attributed to environmental or cultural factors, cannot hope to ever lead to constructive policies aimed towards helping all people - regardless of their genetic potential for intelligence - to achieve their best. Even worse, certain policies which are now being advocated - such as the elimination of standardized testing - are actively harmful.
To answer your categorical imperative: yes, I would much prefer to live in a society where genetic differences in the intelligence potential between individuals and groups is acknowledged, because it would lead to far more sensible social policies which would benefit all - high IQ or low IQ.
Finally, even if you consciously choose not to care about IQ, eventually reality will rear its ugly head and force you to take it into consideration if current social policies cause a lowering of average IQ. If there is such a drift towards a lowering of IQ of the general population, whether through higher reproduction rates of the low IQ segments of the population or immigration, many of the amenities which you take for granted, such as a functioning infrastructure, which can only be supported by a population which is above a certain IQ threshold, will start to deteriorate to such a degree that the IQ problem will become impossible to ignore.
By including groups there, are you going so far as to say, without further information, that "Group X statistically has attribute Y and individual Z is in Group X, therefore Z has (or perhaps should be treated as having) attribute Y?
Further, what specific "sensible social policies" would you have in mind? Reproductive assistance for higher-IQ segments? Inhibition for lower-IQ segments? Testing for immigration? What else?
All three of the policies you have proposed sound good to me. But I would already be more than satisfied if we stopped the current trends towards trying to "equalize through force".
As a concrete example, right now there is a pressure for many high performing high schools across the USA to drop standardized testing in favor of quotas and lotteries, because of the low enrollment numbers of blacks and hispanics. The justification usually given for this is that, starting from the axiom that all groups are of equal ability on average, the discrepancy in numbers can only be due to some kind of racist bias in the methods of evaluation, namely standardized testing (a theory which is unable to explain the fact that asians are even more overrepresented than whites at those schools). I personally believe that this is a wrong interpretation of the numbers, that the reason for the discrepancy is to a great degree due not to some kind of bias in merit-based admission policies, but due to differences in average ability, and that the inevitable consequence of the lowering of standards which will result from the introduction of quotas and lotteries will be the destruction of the outstanding academic quality of these institutions.
I think the main value of IQ on a societal level, besides identifying individuals that need help, is in recognizing that people just innately have differing abilities and, without any external influence necessary, will have different outcomes. There is a narrative being pushed very hard currently that differences in outcome between groups automatically implies discrimination, and fundamentally it's hard to argue with that logic unless you're able to accept that there are just innate differences between people and groups that cause them to have different abilities and make different choices.
Perhaps, but that is mostly due to the historical fact that differences in outcome between groups have been caused by discrimination and "innate differences between people and groups" has been the primary argument in favor of such discrimination---with the "differences" being significantly bogus.
As a result, suggestions that the current status quo[1] is right and just on a scientific basis begin to smell of extraordinary claims that require extraordinary evidence.
Don't get me wrong, but historically and, as far as I've seen, current IQ-related policy suggestions seem to come down to that. If IQ research were to determine that Aboriginal Australians had a median IQ of 160, I rather suspect the whole issue would disappear overnight.
No, not usually automatically, though that's usually the straw man that opponents focus on; the usual actual position is that that is taken as a presumptive cause where there is a clear history of discrimination against the group, and there is not clear evidence that the difference in outcome at issue has a non-discriminatory explanation.
By "linked with eugenics" I mean IQ research was used to provide the scientific basis for eugenics policies, including being designed to demonstrate that undesirable groups had lower intelligence.
The example shows that an employer gains a significant improvement (450%) in avoiding candidates two sigma below the mean, with only a minor improvement (16%) in obtaining candidates above the mean. A better example (if it exists) would show a very strong improvement in finding candidates at least one sigma above the mean.
Low IQ people in low responsibility positions don’t pose much risk, I’d actually think high IQ in highly responsible positions may pose a higher risk: think of hiring a high IQ malefic narcissist who would ruin your business just for fun
Ouch... “Explains” is used here in a very strict technical sense whereas I guess Taleb uses it much more roughly (just the statement “between 5-13%” must make you pause).
As it stands, an R^2 of 0.356 is not actually a refutation of Taleb's "5 - 13%". Taleb probably shouldn't have quantified his point without some rigor behind it, but the author's single counterexample doesn't substantively respond to it.
Using R^2 this way is like saying someone's wrong in an online debate by citing a logical fallacy. I guess, sure, you successfully rebutted them in a shallow sense. But nothing of value has been achieved (yet). We haven't arrived at a better understanding of the truth.
Not that I'm defending IQ. I find it a weird measure, and certainly not one in which to base all trust, but it does seem to correlate reasonably with many outcomes. That other factors also have an influence on these outcomes, and that IQ is not purely measuring intelligence, does not discredit it as a (limited) predictor. Taleb's article would be much stronger if he could point at better predictors.
“Without further information “ is too technical to what Taleb is arguing.
A school like Harvard admits rich people who are maybe 1 SD above the mean and poor people who are 3-4 SD above the mean.
It's a form of "achievement laundering". You might be 22 years old and the only remarkable thing about you is that your parents are rich (either they achieved something or their parents achieved something or got lucky or stole or...) You took classes that weren't graded harshly with some really bright kids, and for the benefit of both groups, rich kids and bright kids are blended together so that newspaper commentators and other pundits will look at them with 80/20 vision and think they are looking at a "meritocracy".
A threat to this is that some alumni families have a strong sense of entitlement and insist on their spawn with inferior IQ (and often inferior moral character in general) get admission.
Some of these children benefit a lot from test prep, some benefit a little, other ones can at best make up a story like "I must have skipped a row and put all the right answers in the wrong rows".
So there is a lot of hate for standardized tests and circa 1970-1980 somebody must have hired a public relations agency, done some polling, figured that you couldn't get sympathy for low-IQ rich kids but maybe you could get sympathy for low-IQ black people and that makes it a little easier for people like Stephen J. Gould to get funding to drop what they are doing and write on the topic of "racial bias in IQ testing".
---
Effective political movements usually have a combination of an elite sector that supplies resources and a mass sector that supplies footsoldiers. The elite is too small to win at anything on it's own, and it is next-to-impossible for mass groups to organize in the long term:
https://en.wikipedia.org/wiki/The_Logic_of_Collective_Action
---
An interesting phenomenon is that you'll often see something in People magazine to the effect that "(Rosanne Barr|Sylvester Stallone|...) has a 148 IQ" and wonder who it is that trains celebrities to ace this this test:
https://en.wikipedia.org/wiki/Raven%27s_Progressive_Matrices
which is has to be, because with just 60 questions, somebody with an IQ > 100 or so should be able to memorize the answers. Maybe you can't really a fault an actor for this because an actor doesn't have to be a genius, they just have to play one on TV.
https://www.iqcomparisonsite.com/Occupations.aspx
The people around a set like an actor to be "smart" in the following sense: you show up to work and know your lines, the director or somebody else tells you what to do and you don't need to be told again. You need to be a "quick learner" of the random sorts of skills you need to do your work. As an actor you can outperform the person you become because your lines are written for you, so you can put 100% of your effort into everything else.
I have a somewhat high IQ, I'm good at hard math (I'm sure better than Sly Stallone) and abstract thought and my words/minute rate of writing has been legendary at times, but I struggle as an actor.
Stallone's success had to do with the breakout success of his movie "Rocky", which he wrote. That in turn had to do with convincing other people to take a chance on him and his script, and that gets you into a whole interpersonal area where the role of IQ is really controversial.
If I wrote fiction I'd be dangerously tempted to be like
https://en.wikipedia.org/wiki/Robert_L._Forward
I am fascinated with how many "Starship(s)" full of ammonia it would take to fill the atmosphere of a small space colony, what you do with the hydrogen and so forth. Stallone is much better at relating to the average person than I am, and that's what makes him a "genius".
What? Essentially all of "success" (which IQ is correlated with) involves the "interpersonal area". If IQ isn't strongly related to successful interpersonal relations, something is broken somewhere.
There was that time my wife and I got from the airport to the mountains in the Dominican Republic with instructions to use various forms of transport I'd never heard of, ran away from motorcycle taxis, shook hands with a police officer (along with everyone on the bus), walked into what we thought was a bus stop but was really a sports betting establishment, and caught the last bus to Ocoa which was a good thing because the only Spanish I knew I learned from reading half of Don Quixote and shopping at Home Depot -- it would have been awkward to find a hotel and regroup the next day.
A person working across a cultural barrier to do a more complex task over a longer time has similar 'knife edges' to performance which will make everything fraught.
See also
https://en.wikipedia.org/wiki/Bamboo_ceiling
"Based on publicly available government statistics, Asian Americans have the lowest chance of rising to management when compared with African Americans, Hispanics and women in spite of having the highest educational attainment."
Otherwise, it seems like a strange dismissal. For one, the roles that an actor plays, and even the persona they have in public don't necessarily match their private persona where their intelligence might be more evident. For two, while luck and other personality factors surely play a significant role, it would make sense that IQ would be a differentiator amongst any group with similar characteristics.
So among good-looking talented actors/actresses with similar demographics, I'd expect IQ to correlate at least somewhat with their success. It seems odd to think otherwise.
For sure [1].
But it is actually pretty easy to distinguish these groups of people. Was the person on the Putnam team? Did the person take Math 55 (or Honors Analysis at Chicago)?
You could argue that these are noisy signals because people from unprivileged backgrounds have to do some catching up before they can crush a Math 55.
[1] https://nymag.com/intelligencer/2019/09/david-e-shaw-college...
A freshmen-only class like Math 55 implies that at least the high school experience was somewhat privileged.
There's still some inherited factors (who has better access to tutors at high school and who can move to a different city for 5 years without having to work). But at least you don't have to get a mortgage to afford a degree.
I think it's precisely because of recent communist history (in 1989 99% of population was equally as poor) that the private universities couldn't overtake the public ones. Not enough rich kids to make a whole university out of them - so instead the trick was to pass exams well enough to outcompete the other 49 candidates for a place at the best public universities.
Are there any public universities you'd describe as globally prestigious?
Outside the US, Oxbridge, ETHZ and NUS would be some examples, where globally prestigious means ~top 20. If you broaden to top 100 there are quite a lot of European/Australian/Canadian/Chinese ones.
as a high-iq plebe, i've known many patricians. i could crush all of them on school-work, yet most of them would easily best me when it comes to wisdom and life-choices.
in other words, i wouldn't be so sure about whose reputation is being laundered here.
If you read the biographies of these high-IQ celebrities, ,s you will see that their non-acting accomplishments generally match with their IQs. They attended prestigious schools, as was the case with Sharon Stone, or tested well , as was the case with James Woods, who attended MIT.
Well, I found The Mismeasure of Man persuasive. I guess you didn’t.
More recently there has been a bit of scandal related to thinly veiled bribes for admissions involving a few celebrities. Where I went to school, ARAMCO endows the petroleum engineering department. Saudi students tended to sit in a V-formation during exams, with the brightest of them at the tip. Professors all knew what was going on, but you can’t just kick out kids whose sponsor pays your six-figure salary.
Frankly, I don’t think academic credentials are a useful proxy measurement of abilities for the reasons you brought up.
Is that what you think affirmative action is? Wow.
The problem now is that a huge amount people who think of themselves as wealthy are thinking in nominal dollar terms, but there are so many of such people today, and they're so conformist in their investment and lifestyle choices that when they retire (likely all at the approximately same time, in approximately the same way), there is no way that the real economy will be able to support their retirement without driving up salaries/incomes and inflation.
Economic conformity is fragility.
But part of the blame rests with Taleb for arranging his arguments in a series of colloquial blog posts with exaggerated claims of both the relevance of this debate to psychology and unjustified speculation about what IQ actually measures — which would never survive peer review. He's invited low-quality debate like this rather than contributing to the development of standards and techniques that could improve the field, maybe out of laziness or just the desire for immediate publicity.
If IQ didn't matter we would see tons of people say things like "I have an IQ of 85 and still am a software engineer at Google, IQ is total bullshit!", but we don't.
> “IQ” is a stale test meant to measure mental capacity but in fact mostly measures extreme unintelligence (learning difficulties), as well as, to a lesser extent (with a lot of noise), a form of intelligence, stripped of 2nd order effects
Right before the conclusion, the author shows us this chart comparing incomes in IQ cohorts: https://www.jsmp.dk/posts/2019-06-16-talebiq/iq_files/figure...
Maybe it's because I'm not a high-IQ, R-programming brain-genius, but it seems like the low-IQ distribution has a very distinct shape with a low mean and leftward skew, whereas the mid and high-IQ distributions look pretty similar, with a modest advantage to the high-IQ cohort. That seems to confirm Taleb's claim that IQ mostly measures unintelligence, and has a much smaller impact on the higher end.
It's crap.
It's the only good thing that came from psychology.
If the army uses it, it can't be so bad.
It stays the same, your whole life.
You can train it.
It's right, but it doesn't help you much because if you look at the world, the people with the highest IQ aren't the most successful.
What's right?
This cohort is most certainly comprised entirely of programmers.
PS. I am wondering whether it correlates with the belief in IQ testing itself :)
Yes, doing pointless soul crushing work makes you the money you need to create the next generation of automatons that do the same.
I’d rather think “if my family had money, I would’ve gotten in” than “I’m just not good enough”.
These silent double standards are a curse. Saying it out loud is the first step to stopping it.
So you can get people with higher intelligence and low conscientiousness being overtaken by less smart people high in conscientiousness who actually finish the task.
My understanding is that with a higher IQ you finish a task quicker . So smarter people tend to get to success quicker but doesn't mean that lower IQ folk are denied successes if both groups have similar levels of conscientiousness.
IQ is not the strongest predictor of success.
So if you do well on tests, then you probably would do well on the IQ test as well as tests in school, which means you are more likely to have attained a higher level of educations, which is associated with a higher income.
To actually prove to me that IQ is a thing, you would have to show me that what you call IQ is innate. Meaning that you can't do better on an IQ test if you study, which I don't believe is true. All written tests I have ever taken have tricks or patterns that you can exploit after some practice. And then you'd have to show me that IQ is a better predictor than some other rote test to tease out general test taking ability or ability to sit still.
A lot of people just work off the basis that IQ is a thing that we can measure, which I don't believe to be true, hence the pseudoscientific swindle
"Twin studies of adult individuals have found a heritability of IQ between 57% and 73% with the most recent studies showing heritability for IQ as high as 80%."
This does not follow. IQ is a thing because they test children on it without warning and without any prep and their performance on these tasks they’ve never seen before in their lives predicts their performance in school better than many other tests.
You can’t say the same thing about, say, a language test for a child unfamiliar with the language because the kid won’t understand it at all. The IQ test is designed to be as unrelated to experience as possible in order to measure the capacity for abstract thinking and the ability to solve novel problems when you first encounter them.
Since you’re undermining a test of your ability to solve something novel, studying for an IQ test is cheating.
I believe this is a myth. There are many ways to prep children for this kind of test, and the wealthy in America have access to an early education system that trains children for standardized tests.
I have anecdotal evidence because I have two children who were raised in different countries. My older child did daycare and preschool in Finland, before we moved to USA. My younger child is now three and in private daycare in Manhattan.
In Finnish daycare, three-year-olds spent their days playing. In affluent New York, they are clearly being trained to perform on tests. I’m not exactly happy about that because I believe the Finnish system creates more balanced individuals on average. But I guess I’ll see the results of this experiment firsthand in 5-10 years.
I suspect the test-oriented early education available to wealthy Americans is such a large confounder that it overwhelms innate ability at this point.
There is no evidence for that and twin and adoption studies show evidence to the contrary: that adopted children and especially twins raised separately over/underperform their adoptive families and perform similarly to their biological relatives.
The only thing we know that can overwhelm innate ability is crushing poverty and the malnutrition and other extreme stresses that come with it. Otherwise, the effect of home environment on IQ is minimal.
It’s different, of course, with language acquisition. Kids learn more vocabulary the more you speak to them and read to them. Kids with wealthy upbringings can have much more opportunity to acquire language than otherwise. The extent to which that helps with abstract pattern recognition, mental rotation, etc is limited. And we have plenty of examples of kids with logical and spatial reasoning that far exceeds their language skills. It’s such a common phenomenon that we have a name for these kids: nerd.
I don't think IQ is a perfect predictor for basically anything. It's a very myopic metric that helps explain, but any statistician today is going to use dozens of variables in their model, not just one.