I disagree with your blunt formulation of intelligence as ‘IQ’. An example: Lewis Terman (yes the father of Frederick Terman who has a building named after him at Stanford) followed a bunch of kids with high IQs—average of 151. As described in the article, William Shockley (have you heard of him?) didn’t have a high enough to be one of the ’Termite’s. But, as every electrical engineer will tell you, Shockley went onto invent the bipolar junction transistor at Bell Labs. (What’s ironic is that Shockley himself adopted a static (unchangeable) view of human ability, became a racist proponent of eugenics and was shunned in the academia in his late career.)
In fact, Barry Schwartz claims that the success rate of the Termites was as good as a random sample of individuals but I can’t quite find links for this claim. (It’s in the book Outliers by Malcolm Gladwell however.)
A discussion of IQ would take us far afield but I suggest you also check out Flynn effect for instance.
AnlamK, I agree that IQ is the measure of a mere shadow of actual ability. When describing a single individual, their IQ does provide a partial indicator as to their competence, but does not even begin to describe a human being.
In more macroscopic terms, however:
(1) People with an average IQ lower than X will not be able to perform task Z which requires IQ much greater than X.
(2) Having a higher IQ than needed for task Z does not make you much better at it, but may qualify you for another job, Q, which is more demanding.
(3) Contrary to what many people think, average IQs can be compared somewhat across cultures; they can be compared somewhat to a common point of reference; and the resulting average IQ measures can differ greatly across countries. Countries that have lower average IQs do not perform nearly as well, which I attribute mainly to argument (1), above.
There are a variety of reasons I can think of that would cause people with extreme IQ’s (150+) to perform randomly on average. In general, my interpretation is that the higher IQ hurts them more than it helps them. With an extreme IQ, it is hard to find an environment in which to develop social skills, and it is harder to enjoy random company. Meanwhile, there are few occupations that require such a high IQ. Instead, the most high-end occupations can be performed by people whose IQs are less extreme, but are more socially developed, and would thus be preferred for those occupations over the maladjusted extreme ones.
I would be very surprised to learn, however, that a study of subjects with merely above average IQs—rather than extremely high ones—didn’t show them to have markedly improved outcomes over subjects whose IQs are below average.
A mostly static view of human potential is probably correct. If you’ll agree that other animals don’t have the same potential as humans because of their genetic differences, then it is far fetched to assume that there are no differences in potential among humans, too.
A point to consider—isn’t IQ grounded to intelligence via improved life outcomes (wealth, education etc.); so if high intelligence levels start to lose correlation with improved outcomes, wouldn’t that make the extreme end of IQ results less and less correlated with actual intelligence?
The notion that abilities de-correlate at high range is known as Spearman’s law of diminishing returns . A simple analogy from sports: ability to run marathon positively correlates with the ability to sprint, in the general public, but among the world class athletes, you run into genetic variations which trade sprinting performance for marathon performance and vice versa.
Another point is that IQ tests have to be neutral with regards to the background skills or knowledge, which has a very unfortunate side effect of not measuring performance of the mechanisms involved in forming or applying skills and knowledge. By the way, on a Gaussian prior, poor correlation implies very substantial regression towards the mean.
wouldn’t that make the extreme end of IQ results less and less correlated with actual intelligence?
It seems much more plausible that at extreme intelligence, the correlation to life outcomes starts to break down. Once you earn enough money to live comfortably, it probably leads to more life satisfaction to spend your time on leisure, rather than earning more, and in particular, the cleverer someone is the more we might expect them to realize that’s the tradeoff.
I meant that the correlation between life outcomes and intelligence breaks down, but the correlation between intelligence and IQ likely remains strong; it sounded to me like you were questioning the IQ-intelligence link because IQ-life outcome broke down at high IQ levels.
As I said, it was my understanding that the correlation between IQ and life outcomes was well-established, and that IQ tests are designed and adjusted to ensure the correlation remains strong. This is a thing, right?
Thus, the hypothesis that the correlation between intelligence and life outcomes breaks down at high intelligence levels suggests that such adjustment would cease to produce IQ-to-life-outcomes correlation.
(Alternately, this whole system may break down somewhat at high levels anyway—I don’t know how much difficulty the relative rarity of really high IQ ratings has introduced.)
I disagree with your blunt formulation of intelligence as ‘IQ’. An example: Lewis Terman (yes the father of Frederick Terman who has a building named after him at Stanford) followed a bunch of kids with high IQs—average of 151. As described in the article, William Shockley (have you heard of him?) didn’t have a high enough to be one of the ’Termite’s. But, as every electrical engineer will tell you, Shockley went onto invent the bipolar junction transistor at Bell Labs. (What’s ironic is that Shockley himself adopted a static (unchangeable) view of human ability, became a racist proponent of eugenics and was shunned in the academia in his late career.)
In fact, Barry Schwartz claims that the success rate of the Termites was as good as a random sample of individuals but I can’t quite find links for this claim. (It’s in the book Outliers by Malcolm Gladwell however.)
A discussion of IQ would take us far afield but I suggest you also check out Flynn effect for instance.
AnlamK, I agree that IQ is the measure of a mere shadow of actual ability. When describing a single individual, their IQ does provide a partial indicator as to their competence, but does not even begin to describe a human being.
In more macroscopic terms, however:
(1) People with an average IQ lower than X will not be able to perform task Z which requires IQ much greater than X.
(2) Having a higher IQ than needed for task Z does not make you much better at it, but may qualify you for another job, Q, which is more demanding.
(3) Contrary to what many people think, average IQs can be compared somewhat across cultures; they can be compared somewhat to a common point of reference; and the resulting average IQ measures can differ greatly across countries. Countries that have lower average IQs do not perform nearly as well, which I attribute mainly to argument (1), above.
There are a variety of reasons I can think of that would cause people with extreme IQ’s (150+) to perform randomly on average. In general, my interpretation is that the higher IQ hurts them more than it helps them. With an extreme IQ, it is hard to find an environment in which to develop social skills, and it is harder to enjoy random company. Meanwhile, there are few occupations that require such a high IQ. Instead, the most high-end occupations can be performed by people whose IQs are less extreme, but are more socially developed, and would thus be preferred for those occupations over the maladjusted extreme ones.
I would be very surprised to learn, however, that a study of subjects with merely above average IQs—rather than extremely high ones—didn’t show them to have markedly improved outcomes over subjects whose IQs are below average.
A mostly static view of human potential is probably correct. If you’ll agree that other animals don’t have the same potential as humans because of their genetic differences, then it is far fetched to assume that there are no differences in potential among humans, too.
A point to consider—isn’t IQ grounded to intelligence via improved life outcomes (wealth, education etc.); so if high intelligence levels start to lose correlation with improved outcomes, wouldn’t that make the extreme end of IQ results less and less correlated with actual intelligence?
The notion that abilities de-correlate at high range is known as Spearman’s law of diminishing returns . A simple analogy from sports: ability to run marathon positively correlates with the ability to sprint, in the general public, but among the world class athletes, you run into genetic variations which trade sprinting performance for marathon performance and vice versa.
Another point is that IQ tests have to be neutral with regards to the background skills or knowledge, which has a very unfortunate side effect of not measuring performance of the mechanisms involved in forming or applying skills and knowledge. By the way, on a Gaussian prior, poor correlation implies very substantial regression towards the mean.
edit: cut-n-paste error in URL.
It seems much more plausible that at extreme intelligence, the correlation to life outcomes starts to break down. Once you earn enough money to live comfortably, it probably leads to more life satisfaction to spend your time on leisure, rather than earning more, and in particular, the cleverer someone is the more we might expect them to realize that’s the tradeoff.
Which is what I said, yes.
I meant that the correlation between life outcomes and intelligence breaks down, but the correlation between intelligence and IQ likely remains strong; it sounded to me like you were questioning the IQ-intelligence link because IQ-life outcome broke down at high IQ levels.
As I said, it was my understanding that the correlation between IQ and life outcomes was well-established, and that IQ tests are designed and adjusted to ensure the correlation remains strong. This is a thing, right?
Thus, the hypothesis that the correlation between intelligence and life outcomes breaks down at high intelligence levels suggests that such adjustment would cease to produce IQ-to-life-outcomes correlation.
(Alternately, this whole system may break down somewhat at high levels anyway—I don’t know how much difficulty the relative rarity of really high IQ ratings has introduced.)