> "the idea that an idiot savant who can tell you if a ten-digit number is prime in seconds but cannot reliably drive to a supermarket to buy groceries a genius based upon the results of a test for mathematic ability? The truth of that is a turd in the IQ measures all punchbowl."
Alas, I think that may be a strawman (perhaps accidentally,…
> "the idea that an idiot savant who can tell you if a ten-digit number is prime in seconds but cannot reliably drive to a supermarket to buy groceries a genius based upon the results of a test for mathematic ability? The truth of that is a turd in the IQ measures all punchbowl."
Alas, I think that may be a strawman (perhaps accidentally, in good faith). I've read a lot about IQ, and I have never come across a single source which asserts that "IQ measures all". There are virtually zero people arguing that position. A fair presentation would be that IQ, properly administered, presents a surprisingly good estimate of g, or general intelligence. "General intelligence" or skill at abstract thinking, is never, ever presented as the only kind of intelligence or smarts. It appears, from hundreds of experiments, to be one kind of intelligence which pays off in many domains, and thus has a substantial effect on outcomes. Never the only effect, and perhaps not even 50%, but nevertheless it's one of the more well researched and validated component traits.
In other words, you, me and every psychometrician in the business are agreeing that IQ tests provide a good but not perfect measure of *just one form* of intelligence, albeit an important one. We all know there are other kinds of intelligence not measured by IQ tests, and that there are other character traits which can be as important at those intelligences in influencing outcomes. IQ tests have never pretended to measure all forms of intelligence.
---
You give an example of your Asian wife not being as good at mental arithmetic as you are. That would refute one of the following two hypotheses. However, the actual assertion that psychologists make is the other one:
(1) All Asians are better at math than all non-Asians, without exception
(2) The aggregate mean measured mathematical ability of Asian-Americans is higher today than that of other commonly used American Racial/Ethnic groups, but there is a great deal of overlap in the distributions, and the variation between individuals is much greater than the difference in group averages.
Of course, statistics about "Asian-Americans" mash up a great variety of cultures from Pakistan to Siberia to the Philippines. It would be more meaningful to compare based on national origins. Some of the stats are even stronger regarding Chinese-Americans or Japanese-Americans, but Hmong-Americans tend to be behind the average. (There are also differences between Americans from different European countries, or different African countries; it's not just Asians). The big categories can only reflect trends broad enough to still show a signal after aggregating many "Asian" subpopulations - in proportion to their prevalence in the US. In the US, the stats are dominated by the proportion of East Asians (CJK) and South Asians (India and nearby) in the pool being sampled, both of whom tend to score highly in some areas.
(Side note: as a math major, albeit long ago, I will note that performing simple arithmetic calculations is a quite different cognitively from actual mathematics, which involves deriving or applying some very advanced theorems. From the inside of mathematics, this is like suggesting that being good at pounding nails is equivalent to being a skilled architect, or uses the same skillset. Or considering good penmanship (or typing) to reflect the same skills as being a good novelist or tech writer. Many people can be brilliant at math and bad at manual arithmetic, or vice versa. (I'm one of these; I've generally found math easy to learn, but I'm middling at performing arithmetic without mistakes; I'd rather use a calculator) The tests which Asian-Americans have a statistical tendency to do well at are not tests of manually adding, multiplying and dividing numbers but are instead about understanding and being able to manipulate and apply mathematical abstractions. I would have no expectation as to whether Asian-Americans are better or worse at doing manual arithmetic. Also: I use "Asian-American" because I have only seen statistics for them; I make zero projections from that onto non-Americans living in Asia. I have no experience or knowledge about that, I'm at most rather mildly interested so I've never investigated math tests in Japan, for example. Sorry pet peeve, we will return you to your regularly scheduled program).
I am well aware that the idea that IQ measures all is not an explicitly expressed idea, but it seems implicit in the emotion around an association of race and intelligence.
There is, as you note, a difference in being able to do basic math in your head and understanding higher math. At peer reviews of new electronic circuitry design, I worked with an engineer who liked to work thru circuits looking at component values and mentally (and announce results) calculate gains, filter roll-offs, etc. While useful for discussion, it always seemed a bit like showing off. A room full of people with at least a knowledge of calculus, linear algebra, etc. but rapid mental math calculations were not the norm. The design calculations were performed with Spice, and spreadsheet formulas where mental math was not relied upon. A different skill, as you note.
As for big categories, tiny groups tend toward expanding their size to gain relevance while at the same time clinging to a degree of exclusiveness. My wife shrinks Asian to Oriental (eyes: epicanthal folds) or even Thai, Chinese, etc. depending on the underlying reason for even mentioning a distinction.
> "the idea that an idiot savant who can tell you if a ten-digit number is prime in seconds but cannot reliably drive to a supermarket to buy groceries a genius based upon the results of a test for mathematic ability? The truth of that is a turd in the IQ measures all punchbowl."
Alas, I think that may be a strawman (perhaps accidentally, in good faith). I've read a lot about IQ, and I have never come across a single source which asserts that "IQ measures all". There are virtually zero people arguing that position. A fair presentation would be that IQ, properly administered, presents a surprisingly good estimate of g, or general intelligence. "General intelligence" or skill at abstract thinking, is never, ever presented as the only kind of intelligence or smarts. It appears, from hundreds of experiments, to be one kind of intelligence which pays off in many domains, and thus has a substantial effect on outcomes. Never the only effect, and perhaps not even 50%, but nevertheless it's one of the more well researched and validated component traits.
In other words, you, me and every psychometrician in the business are agreeing that IQ tests provide a good but not perfect measure of *just one form* of intelligence, albeit an important one. We all know there are other kinds of intelligence not measured by IQ tests, and that there are other character traits which can be as important at those intelligences in influencing outcomes. IQ tests have never pretended to measure all forms of intelligence.
---
You give an example of your Asian wife not being as good at mental arithmetic as you are. That would refute one of the following two hypotheses. However, the actual assertion that psychologists make is the other one:
(1) All Asians are better at math than all non-Asians, without exception
(2) The aggregate mean measured mathematical ability of Asian-Americans is higher today than that of other commonly used American Racial/Ethnic groups, but there is a great deal of overlap in the distributions, and the variation between individuals is much greater than the difference in group averages.
Of course, statistics about "Asian-Americans" mash up a great variety of cultures from Pakistan to Siberia to the Philippines. It would be more meaningful to compare based on national origins. Some of the stats are even stronger regarding Chinese-Americans or Japanese-Americans, but Hmong-Americans tend to be behind the average. (There are also differences between Americans from different European countries, or different African countries; it's not just Asians). The big categories can only reflect trends broad enough to still show a signal after aggregating many "Asian" subpopulations - in proportion to their prevalence in the US. In the US, the stats are dominated by the proportion of East Asians (CJK) and South Asians (India and nearby) in the pool being sampled, both of whom tend to score highly in some areas.
(Side note: as a math major, albeit long ago, I will note that performing simple arithmetic calculations is a quite different cognitively from actual mathematics, which involves deriving or applying some very advanced theorems. From the inside of mathematics, this is like suggesting that being good at pounding nails is equivalent to being a skilled architect, or uses the same skillset. Or considering good penmanship (or typing) to reflect the same skills as being a good novelist or tech writer. Many people can be brilliant at math and bad at manual arithmetic, or vice versa. (I'm one of these; I've generally found math easy to learn, but I'm middling at performing arithmetic without mistakes; I'd rather use a calculator) The tests which Asian-Americans have a statistical tendency to do well at are not tests of manually adding, multiplying and dividing numbers but are instead about understanding and being able to manipulate and apply mathematical abstractions. I would have no expectation as to whether Asian-Americans are better or worse at doing manual arithmetic. Also: I use "Asian-American" because I have only seen statistics for them; I make zero projections from that onto non-Americans living in Asia. I have no experience or knowledge about that, I'm at most rather mildly interested so I've never investigated math tests in Japan, for example. Sorry pet peeve, we will return you to your regularly scheduled program).
I am well aware that the idea that IQ measures all is not an explicitly expressed idea, but it seems implicit in the emotion around an association of race and intelligence.
There is, as you note, a difference in being able to do basic math in your head and understanding higher math. At peer reviews of new electronic circuitry design, I worked with an engineer who liked to work thru circuits looking at component values and mentally (and announce results) calculate gains, filter roll-offs, etc. While useful for discussion, it always seemed a bit like showing off. A room full of people with at least a knowledge of calculus, linear algebra, etc. but rapid mental math calculations were not the norm. The design calculations were performed with Spice, and spreadsheet formulas where mental math was not relied upon. A different skill, as you note.
As for big categories, tiny groups tend toward expanding their size to gain relevance while at the same time clinging to a degree of exclusiveness. My wife shrinks Asian to Oriental (eyes: epicanthal folds) or even Thai, Chinese, etc. depending on the underlying reason for even mentioning a distinction.