The title is a bit misleading. The study is basically comparing the path of the top 1% to the top 2% of the IQ distribution and arguing that it doesn't matter that much in term of extraordinary achievements (nobel prize, etc).
If you compare the top 20% to the bottom 20% of the distribution I bet there will be a strong correlation to professional sucess.
There's an excellent rant somewhere about the pattern where people at the extreme margins of a trait proudly declare that the trait doesn't really matter for success.
It's the rocket scientist saying he knows lots of people smarter than him, and obviously hard work and precision are what really matter in rocketry - never mind that estimates put the average IQ in the physical sciences around 125. Or the pro basketball player saying that it's not enough to be the biggest, they key is to train hard and work as a team - never mind that the person saying it is the NBA-average height of 6'7". What these people are actually seeing is that once you have lots of a key trait, other traits gain relevance.
Or to take a less quantifiable example, we can look at something like musical skill. 10,000 hours of practice will (in)famously make you an expert, but Ericsson et al actually found that "time to expertise" is fairly constant in some fields, and massively variable in others. (I seem to remember examples varying by a factor of four?) Skill at a musical instrument can be learned by almost anyone, but some people will see vastly better return on their time investment than others. And if we're looking at professionals, then "just practice harder" ceases to be an option.
None of which means most people are excluded from most roles, or that success is tightly correlated with IQ. But there's a distinct pattern of assessing the importance of some trait after applying a hard filter on that trait, and it does everyone a disservice to pretend that's comparable to a population-level result.
the pattern where people at the extreme margins of a trait proudly declare that the trait doesn't really matter for success
I just noticed yesterday for the nth time a girl on facebook posting her pic with a caption saying how true beauty is on the inside. Can't help noticing it's only the most beautiful girls who do that, which I've always found puzzling. You've shed some light on why for me, thanks. I'll think twice before again writing on here how unimportant talent is in music! (I'm a musician.) I just don't know about that. My favourite story about talent, from Kevin Spraggett:
"I remember fondly one conversation I had a few years back with Boris Spassky. We were discussing 'THE' Victor Korchnoi...Boris and Victor had been bitter adversaries for more than 40 years at the time of this conversation, and they had played more than 60 times in official competitions..(including 2 candidates finals)... Boris [said] that Korchnoi had every quality necessary to become world champion but lacked one very essential quality...and it was precisely this quality that prevented him from attaining chess' highest title. I coaxed Boris on...He began to list Korchnoi's many qualities: ...Killer Instinct (nobody can even compare with Victor's 'gift')...Phenomenal capacity to work (both on the board and off the board)...Iron nerves (even with seconds left on the clock)...Ability to Calculate (maybe only Fischer was better in this department)...Tenacity and perseverance in Defense (unmatched by anyone)...The ability to counterattack (unrivaled in chess history)...Impeccable Technique (Flawless, even better than Capablanca's)...Capacity to concentrate (unreal)..Impervious to distractions during the game...Brilliant understanding of strategy...Superb tactician (only a few in history can compare with Victor)
...Possessing the most profound opening preparation of any GM of his generation...Subtle Psychologist...Super-human will to win (matched only by Fischer)...Deep knowledge of all of his adversaries...Enormous energy and self-discipline...
'But, Boris, what does Victor lack to become world champion?' Boris' answer floored me: "He has no chess talent !" And then he roared with laughter..."
I think that the filters you mention are an important mechanism here. These continuous variables like IQ or test scores get mapped onto binary ones (filters/gates) as you make it, or don't, into schools or jobs or other opportunities. It's not necessarily a linear or symmetrical mapping, either.
For example, when I was in graduate school, the chair was discussing how high GRE scores were not very strong indicators of 'success' in grad school (i.e. doing good research, publishing it, and getting out in a reasonable time). However, low GRE scores were much stronger indicators of 'failure' given these criteria.
IQ probably maps pretty linearly to GRE scores (for most of the small subset of the population who would ever take such a test), but with regards to scientific progress there are a series of thresholds that are more important (grad school, faculty hiring, securing funding, etc. etc.). These thresholds are really about opportunity and access.
For native english speakers, you mean? When i participated, GRE had 2 sections on english language: 1. about vocabulary(may not be relevant depending on your discipline) 2: reading comprehension(certainly useful). I don't think it maps that well to IQ. I certainly had to work on my vocabulary, which i don't have any use for and don't remember much at all.
You should write this rant up as a blog post! I'd read it.
One factor doesn't seem to me to get mentioned enough in these discussions - call it "agreeability" or "sanguinity." Basically, ability to successfully cope with neuroses and other social/behavioral challenges that seem to be more common among intelligent people.
I put some thought into the people I've met who seemed to me to be the greatest outliers in terms of intellectual achievement. One is a famous writer, another is a classics professor, and another is a math prodigy turned neuroscientist. The commonalities between the three are a combination of 1) high g, 2) obsessive focus, and 3) the aforementioned happiness/agreeability/ability to cope with neuroses. All three of the people I have in mind are neurotic in some ways, but can channel it toward constructive channels. Haven't looked into the research on this but I would be curious if you could recommend any study that addresses this component. I agree that the Gladwellian "just practice a lot" view doesn't seem to hold a lot of weight, but anecdotally it seems to me like raw intellectual ability isn't the determining factor either.
>>It's the rocket scientist saying he knows lots of people smarter than him
Well, that much is probably true: the very nature of his/her profession surrounds him/her with similarly smart people, some of whom are going to be smarter.
And there's also probably a big difference between the top 0.01% percent and the top 1%. The difference between the top 1% and top 2% isn't very important for most types of talent. Being in one of those categories for athletic talent, for instance, wouldn't get you into a professional sports league of any kind.
“Gladwell cites studies done by Arthur Jensen that indicate the IQ level needed to successfully attend and graduate from undergraduate college is 115, after which Jensen believe IQ becomes relatively unimportant. For comparison, Albert Einstein, a well-known and commonly agreed upon genius, had an IQ of 150. Studies by Liam Hudson indicate that, ‘a mature scientist with an adult IQ of 130 is as likely to win a Nobel Prize as is one whose IQ is 180.’”
He gives the example [paraphrasing from memory] of an individual with an IQ of 170 who was so intelligent it impaired his ability to relate to normal people and hold a job.
Worth noting in this discussion: Einstein never took an IQ test, at least none that historians know of. Pretty much any time I see a claim about a historical figure's IQ, it turns out to be an invented figure made on the basis of a totally unscientific "estimate." The tendency of pop writers about intelligence to use these invented historical IQs as a measuring stick makes me doubt the whole enterprise (though I'm not singling out Gladwell here, whose books I generally like).
> He gives the example [paraphrasing from memory] of an individual with an IQ of 170 who was so intelligent it impaired his ability to relate to normal people and hold a job.
High IQ doesn't do that (in fact, it makes it easier to understand other people, including those of average intelligence, if you are so inclined), though people with high IQ and a problem that does impair their social interactions may be inclined to prefer to attribute the social difficulty to their strength (IQ) rather than some problem.
Some of the problems that do that may be loosely correlated with IQ, despite being distinct from it, though.
Humans do seem to generally prefer the company of those with a sufficient degree of shared traits, interests and preferences. Being an extreme outlier in IQ, as in any other area, may just make it statistically harder to find such company.
Relating to somebody has two parts: understanding them and empathizing with them. The empathizing part is largely built on feeling similar to them, which could be hard with a big IQ difference.
It makes understanding people easier - much easier, in fact - but it doesn’t necessarily help you use that understanding any better.
Firstly, as others have observed, it makes it hard to relate to others when the majority of people you meet seem to have the intellect of a child - I don’t mean this unkindly, rather more that one ends up looking at most folks as making grievous and avoidable errors incessantly, and almost as overgrown children. You want to help, but you can’t, because nobody likes being patronised or made to feel inferior, but if you don’t help, they resent you for not having helped. It isn’t intractable, but it’s most readily avoided by avoiding interaction in the first place.
Secondly, intimidation. Ties back to making people feel inferior, and generally stymies social interactions. It’s tough, even in a marriage, as they either know you’re playing dumb, or feel like you’re beating them over the head with your intellect. Either way you end up hurting their self worth every time you notice something or think of something and unwisely decide to share it.
Thirdly, overshoot. I tend to assign far more intelligence and cunning to others than there is, and resultabtly second, third, fourth guess their mind-state, as it’s hard to comprehend the disappointingly facile motives behind behaviour.
Finally, expectations. Once you let on that you can solve the intractable while half-asleep, you’re expected to do it day in, day out, and find that the world beats a path to your door with their petitions and demands, which feeds more than a degree of misanthropy.
I’ve met maybe three people in my life who I’ve felt on a level with, and one of them is my brother.
The expectations part is the brutal one for me. Did it for years. Only started trying to not in the last year. Lost a few folks I thought were friends, but turned out to just be folks who were outsourcing their problemsolving to me.
150 IQ is extremely high ~1/1,125 relative to the general population and that score may or may not be that accurate. Which seems to disagree with your suggestion.
Winning a Nobel prize is a highly random process as someone needs to be in an 'open' field ripe for major discovery.
That's what I was thinking. Nobel prize is like a lottery that requires a high IQ and a desire to work in academia and research among other factors in order to buy a ticket.
You also have to be lucky enough not to have contemporaries with equally impressive work in more popular fields.
> Teller also said "von Neumann would carry on a conversation with my 3-year-old son, and the two of them would talk as equals, and I sometimes wondered if he used the same principle when he talked to the rest of us." [1]
Teller is best known as "the father of the hydrogen bomb".
"I have sometimes wondered whether a brain like von Neumann's does not indicate a species superior to that of man"
that's a pretty good indicator that the man was beyond brilliant and I think it's fair to say that they have the better authority than any of us here to make that classification.
Or George Pólya's quote
"Johnny was the only student I was ever afraid of. If in the course of a lecture I stated an unsolved problem, the chances were he'd come to me at the end of the lecture with the complete solution scribbled on a slip of paper."
>I don't think there would be too much disagreement if I said von neuman was one of the smartest 10 guys to live in the last century.
Obviously there won't be much disagreement on a platform dominated by computer scientists. If you went to forums for other fields they would also pick their favorites and see no reason to disagree that they are not only the best and brightest in the field, but also the best and brightest in general.
And to be clear I'm not saying he isn't. I don't really know.
Gladwell has a real problem with appealing-but-unsubstantiated claims, and Outliers is perhaps his worst offense.
The IQ content is less egregious than other flaws - his famous "10,000 hours" bit is so oversimplified that the authors of the study he cites have condemned Gladwell's treatment of their work. But even the IQ stuff is quite bad.
The first note would be that the initial statement here is a shockingly strong claim which goes far beyond Hudson's studies. (115 to 130 is the difference between 15% of the population and 2.3%.) And, of course, the followup reference to Einstein seems entirely self-defeating - how does it discount IQ to note that a famous genius was well above three sigma?
I don't have a copy of Outliers to reference, but the paraphrase of Jensen is also extremely surprising since he's widely criticized for over valuing the role of heritable intelligence in success. All I can find outside of his much-condemned racial work is a brief statement that he believed (in 1969) there were many unfilled jobs requiring an IQ upwards of 110. This is basically what I would have expected from his legacy - not a claim that higher IQ is irrelevant, but instead that +1 SD is enough to open many professional doors.
Meanwhile, the high-end claims raise several other questions.
Most damningly, any discussion of 170 and 180 IQs is garbage for the common 15 point SD. 5 sigma (IQ 175, depending on normalization) is a 1-in-1.7M event. No IQ test ever devised has been calibrated on that population; Raven's Advanced Matrices only goes to ~150, and other high-IQ tests are also seen to cap at 4 SD or lower. Test-retest and inter-rater variation are high enough that's it not entirely clear whether IQ tests can be calibrated out to 170+, at least without a massive test population.
I can't find Hudson's actual study, only replications of that quote, but it's a disaster coming and going. IQ 180 cannot possibly be a value he found in accurate studies, and a work on Nobel winners isn't necessarily relevant to high-IQ success anyway. Nobel winners are generally people whose research succeeded in a major way, not just those who did very good work, and often features mixed groups like "the theoreticians who made the theory" and "the experiments who confirmed the theory", where quite likely only one side is exceptional in the field.
Finally, the example of a very high IQ person leading a dysfunctional life fails on several levels. I'll let the Gladwell's use of anecdote slide, because studies on very high IQ children generally do find large rates of non-success, but those outcomes are heavily distorted. First, because childhood IQ testing is heavily age-adjusted, with the result that fast development can produce astronomical IQ scores in children who will be merely quite smart when their peers have caught up, both undermining the relevance of the result and creating unsustainable high expectations. Second, because it ignores the tendency of extreme outliers on one axis to be unusual on other axes also. High IQs might impede socialization, but they're also correlated with mood and anxiety disorders, torsion dystonia, and a variety of other conditions which have obvious negative effects on life outcomes. For these purposes, causation doesn't matter - it's enough to note that without correction, any study of success-by-IQ will be distorted by these conditions. We can imagine a case where success in certain fields is dominated by those very high IQs people who don't get nasty side-effects, while the overall performance for high IQ people is unremarkable.
None of which is meant as an attack on you, or even a claim that the people at the peak of e.g. physics are higher-IQ than the field's average. I just think it's worth noting how fractally wrong Outliers manages to be in even a brief statement like this one.
While I agree with the intentions of your comment, it fails to be a persuasive rebuttal. Not saying you’re wrong, just saying a more productive comment would have responded to the specific claims of Gladwell’s in the comment, versus an ad hominem dismissal.
There may be a big difference between the top 0.01% of g and the top 1% of g (or there may not). But IQ is only a loss abstraction of g, and it almost certainly lacks the resolution to say how top 1% of IQ and top 0.01% of IQ actually correlate with g.
(In contrast, you can be relatively certain that someone in the top 1% of IQ has a higher g than someone with bottom 50% of IQ).
I'm not really assuming it's very precise. It's more that there's a huge difference between 99th percentile levels of talent and 99.99th percentile levels of talent. For instance, I am easily better than 99/100 males at basketball. I'm 6'6", reasonably athletic, and I played semi-competitively when I was young. But I am nowhere near being better than 9,999/10,000 males. And no amount of hard work or persistence is going to change that.
The difference between 99th and 99.99th percentile is 1.4 standard deviations, a common IQ test ought to be able to be accurate to that. Otherwise it couldn't measure the difference between 100 and 120 IQ (0 to 1.4z) which it clearly can.
I don't think your reasoning is valid. A test might be unable to discriminate accurately at the extremes not because it's uniformly too inaccurate but because it doesn't have enough range.
Toy model: the test consists of one question that everyone in the top 1% can do and no one in the bottom 99% can do; one question that everyone in the top 2% can do and no one in the bottom 98% can do; ... one question that everyone in the top 99% can do and no one in the bottom 1% can do. This test discriminates very nicely and accurately throughout its range of applicability, but it will do no good at all from distinguishing a top-0.01% person from a merely top-1% person.
(Just as a tape measure 2m long will let you compare people very accurately by height provided they're no taller than 2m, but will be much less useful for people taller than that.)
Modern tests are delivered by computer and typically are adaptive. This means that as you answer questions correctly you get asked increasingly difficult questions until you get some wrong.
This means that you aren't limited to asking the same questions to everyone so you can have appropriate discrimination through the range.
> a common IQ test ought to be able to be accurate to that
Accurate to what, even? The very notion of IQ is fuzzy, so naturally any test trying to measure the value would inherit that fuzziness.
The difference between 100 and 120 may be statistically identical to 99 and 99.99, but the practical difference is vastly different. At a certain point, the IQ test is "defeated", and any value above a certain threshold is nose.
If you yell into a microphone, the recording will come out distorted and inaccurate. Yet if you speak normally into one, the recording will sound rather true to life. This is because the microphone has been tuned to a certain level of sensitivity, and when that threshold is exceeded, what it records is clipped.
A similar principle holds for many types of human tests in education, psych, etc.
they are very not precise. Most give a 80% confidence interval of +- 8 Points (or 1/2 STD). Also the different tests measure slightly different types of intelligence/put emphasis on other things. There's only ~0.8 correlation between modern IQ tests.
Yes, in fact, because it's generally calibrated to the whole human population and is defined on an assumed bell curve, the differences in scores in the outliers are somewhat meaningless
Well, yes. Comparing top .01% to top 1% is akin to comparing top 1% to general population - NOT akin to comparing top 1% to top 2%. So a big difference should be expected/not surprising at all.
I genuinely wonder sometimes if no one on HN has actually bothered to go through and take a full intelligence test.
You get a multi-page report back with measurements of all types of cognitive functions. An IQ number is given, but it quickly becomes apparent that number is just a summary measure. Just as a 3.95 GPA doesn't tell what subjects a student is good in, a high IQ doesn't tell you what areas of mental strengths and weaknesses someone has.
That 3.95 GPA student may have had to spend 4 hours a day studying math, while being able to write beautifully structured long form English essays in half that time.
Maybe that student did great in stats and had a horrible time in discreet mathematics. Or vis versa.
Summary numbers are useful in that large variances can indicate overall trends, that 4.0 GPA student did better all up, but that student with a 3.0 GPA may very well be better in a given subject than the 4.0 GPA student.
Likewise, someone with a high IQ can still have weaknesses, it is just that to get the really high score, any weaknesses will have to be offset by strengths in other places.
If you are reading this an are interested in finding out an objective scientific measurement of what you are good and bad at, pay for a full psychological intelligence evaluation. It takes a long time, the testing is long enough that you probably want to do it over multiple days, insurance won't cover it, and it is rather expensive (I think I paid around $1500 for mine, maybe more).
At the end of it all you'll get an IQ number, and a better understanding of why one number does not represent a person's capability.
I am more railing against the discussion of IQ as the definitive way to measure intelligence. Everyone down thread is arguing "Yes IQ! No IQ! Confounding factors!"
It is as if people argued about someone's GPA in college while not paying attention to the subject's major, study habits, or natural abilities.
There is no single, easy to post about online, number.
Intelligence is not single faceted. Posting and interpreting an entire psych tests worth of numbers is possible, but it is harder to have silly Internet arguments over subtly and nuance around a topic that requires research and education.
The level of discourse talking about "IQ" reaches is the same as discussing a new vehicle by saying it has 300 horse power, and not mentioning if it is front wheel drive, rear wheel drive, or how many wheels the vehicle even has.
Sure if the delta is large enough then it doesn't matter, but there is a serious trees/forest problem that happens with these topics of discussion.
Intelligence is not single-faceted. However the evidence is overwhelming that those facets are far from uncorrelated. If you know that person X is very good with mathematics, you can guess with reasonably high probability that any given other facet of intelligence you may care to name is also likely to be high.
People are diverse and brains are complicated and you can always find someone who is generally very intelligent but can't string three words together in writing, or someone who is generally incredibly stupid but has one particular measure very high ("idiot savant"). But in the general population all the intelligence measures are highly correlated, which means that something like an IQ number can exist and have significant, if not necessarily totally definitive, truth to it.
I feel like the only person confused is you. You're talking about trees (IQ isn't intelligence) when everyone else is talking about the forest (IQ and outcomes).
That correlation does not mean that people's IQ caused their success or lack thereof. I would bet both are caused by economic circumstances, among other things, especially access to resources like good education.
If you compare the top 20% to the bottom 20% of the distribution I bet there will be a strong correlation to professional sucess.