Friday, January 19, 2007
BELL CURVE RELOADED ?
Thanks to reader Dominic C. and eddie, I was alerted to a trio of articles by Charles Murray, the influential libertarian public intellectual and author of The Bell Curve, Human Accomplishment: The Pursuit of Excellence in the Arts and Sciences, 800 B.C. to 1950 and numerous other books. The articles are short and vigorously argued:
"Intelligence in the Classroom"
"What's Wrong with Vocational School?"
"Aztecs vs. Greeks"
Murray writes about IQ and public education from the perspective of the psychometrician and social policy analyst. This is not to say he is wrong - most of Murray's points of argument are well grounded in peer-reviewed consensus and are old hat to experts, even if they sound controversial to laymen - simply that his position is derived from the predictive reliability garnered from looking at the aggregate mean scores of points on the bell curve. In setting national policy with an eye to cost-benefit ratios this is a perfectly appropriate perspective for discussion; indeed, many of Murray's concerns about increasing spending on vocational and gifted educatuion would be in the best interest of millions of school children.
The problem lies primarily in the rigid and deterministic conception of "g" used by Murray, rather than a more fluid, neurobiological one that accepts "g" as one of several major drivers in cognition or as one factor in a complex system of intrinsic processing power, efficiency gained through practice and environmental stimulus to a brain that appears to have a massively modular structure. For example, creativity and insight, some of the most valuable aspects of human thought, overlap but not directly correlate with IQ scores. The ability to handle increasing complexity of variables and mental processing speed are also not always in sync, either.
While Murray is generally correct if you construe his points narrowly, a certain degree of humility about our lack of knowledge about human brain operations is in order. While we see, for example, correlative evidence of brain activity with mental and physical tasks in MRI studies, we have much to learn about what these patterns of activity represent. Intuitive thinking, insight, "fingertip feeling" are not quantifiable with the same ease or reliability as are verbal or mathematical reasoning or even spatial-pattern recognition.
Food for thought.
LINKS TO RELATED POSTS AT:
Gene Expression
Eide Neurolearning Blog ¶ 9:10 PM
Thanks to reader Dominic C. and eddie, I was alerted to a trio of articles by Charles Murray, the influential libertarian public intellectual and author of The Bell Curve, Human Accomplishment: The Pursuit of Excellence in the Arts and Sciences, 800 B.C. to 1950 and numerous other books. The articles are short and vigorously argued:
"Intelligence in the Classroom"
"What's Wrong with Vocational School?"
"Aztecs vs. Greeks"
Murray writes about IQ and public education from the perspective of the psychometrician and social policy analyst. This is not to say he is wrong - most of Murray's points of argument are well grounded in peer-reviewed consensus and are old hat to experts, even if they sound controversial to laymen - simply that his position is derived from the predictive reliability garnered from looking at the aggregate mean scores of points on the bell curve. In setting national policy with an eye to cost-benefit ratios this is a perfectly appropriate perspective for discussion; indeed, many of Murray's concerns about increasing spending on vocational and gifted educatuion would be in the best interest of millions of school children.
The problem lies primarily in the rigid and deterministic conception of "g" used by Murray, rather than a more fluid, neurobiological one that accepts "g" as one of several major drivers in cognition or as one factor in a complex system of intrinsic processing power, efficiency gained through practice and environmental stimulus to a brain that appears to have a massively modular structure. For example, creativity and insight, some of the most valuable aspects of human thought, overlap but not directly correlate with IQ scores. The ability to handle increasing complexity of variables and mental processing speed are also not always in sync, either.
While Murray is generally correct if you construe his points narrowly, a certain degree of humility about our lack of knowledge about human brain operations is in order. While we see, for example, correlative evidence of brain activity with mental and physical tasks in MRI studies, we have much to learn about what these patterns of activity represent. Intuitive thinking, insight, "fingertip feeling" are not quantifiable with the same ease or reliability as are verbal or mathematical reasoning or even spatial-pattern recognition.
Food for thought.
LINKS TO RELATED POSTS AT:
Gene Expression
Eide Neurolearning Blog ¶ 9:10 PM
Comments:
<< Home
Very well put.
We've known for thirty years, at least, that two factors predict school success: school behavior and intelligence. These factors are roughly equally powered, and one appears to be much more easily altered than the other.
This should be clear on reflection. College students tend to be those with some combination of intelligence and hard-work. Merely putting in effort lets you overcome a lot in school, as in life.
Thus, even if g was the only "intelligence," Murray would still be going too far.
In that way, these articles are similar to The Bell Curve. They become controversial for straightforward talk, while the actual academic questions are brushed aside.
I would imagine that domain-specific fingertip feeling is testable, by the way. Implicit association tests may be one method...
We've known for thirty years, at least, that two factors predict school success: school behavior and intelligence. These factors are roughly equally powered, and one appears to be much more easily altered than the other.
This should be clear on reflection. College students tend to be those with some combination of intelligence and hard-work. Merely putting in effort lets you overcome a lot in school, as in life.
Thus, even if g was the only "intelligence," Murray would still be going too far.
In that way, these articles are similar to The Bell Curve. They become controversial for straightforward talk, while the actual academic questions are brushed aside.
I would imagine that domain-specific fingertip feeling is testable, by the way. Implicit association tests may be one method...
More of this?
Bollocks man.
The overdrawn argumentation based on a "g" that itself is overdrawn has a certain stench to it.
As does his approach to race.
No, this sort of neo-Victorian vulgar social Darwinish thinking is neither very interesting, nor very illuminating. The questions he says he wants to attack are served by a more fluid conception of intelligence, and a better grounding in modern population genetics - to cite merely two issues he gets wrong.
Bollocks man.
The overdrawn argumentation based on a "g" that itself is overdrawn has a certain stench to it.
As does his approach to race.
No, this sort of neo-Victorian vulgar social Darwinish thinking is neither very interesting, nor very illuminating. The questions he says he wants to attack are served by a more fluid conception of intelligence, and a better grounding in modern population genetics - to cite merely two issues he gets wrong.
# posted by 
: Saturday, 20 January, 2007
: Saturday, 20 January, 2007
Hi Dan,
I agree with you on the value of task persistence. Murray overstates what can and cannot be learned at a given IQ level, which is part of his " hard" interpretation of of "g" as a variable.
To use his calculus example, that's quite hard for someone at an average IQ level but with a longer time frame, slower pace, more practice, it can be mastered, if not enjoyed.
OTOH Dr. Von teaches some select groups of 7-8 grade students in the summer at NU ,who learn calculus in about a day in order to proceed into his modern physics course. Not many people have that kind of ability to handle advanced mathematics at so young an age but that is what can be done with outlier subpopulations.
Hi Col-
Motivation behind raising this data and what sociological conclusions drawn from them are different matters from the validity of psycholetric testing itself. Or concepts like "g". What Murray is doing though, is leaving out the appropriate caveats.
I don't hold with drawing conclusions about individuals based on aggregate means of the groups to which they might belong ("race" itself, being a fairly nonsensical category in scientific terms). I've had thousands of various kinds of IQ and other standardized test scores of students I have worked with pass through my hands which has given me some experience in considering the limitations or usefulness of such testing.
A fluid interpretation of intelligence is a more acurate picture, I would agree -though I though that was clear from my criticisms of Murray's " hard" interpretation of "g".
I agree with you on the value of task persistence. Murray overstates what can and cannot be learned at a given IQ level, which is part of his " hard" interpretation of of "g" as a variable.
To use his calculus example, that's quite hard for someone at an average IQ level but with a longer time frame, slower pace, more practice, it can be mastered, if not enjoyed.
OTOH Dr. Von teaches some select groups of 7-8 grade students in the summer at NU ,who learn calculus in about a day in order to proceed into his modern physics course. Not many people have that kind of ability to handle advanced mathematics at so young an age but that is what can be done with outlier subpopulations.
Hi Col-
Motivation behind raising this data and what sociological conclusions drawn from them are different matters from the validity of psycholetric testing itself. Or concepts like "g". What Murray is doing though, is leaving out the appropriate caveats.
I don't hold with drawing conclusions about individuals based on aggregate means of the groups to which they might belong ("race" itself, being a fairly nonsensical category in scientific terms). I've had thousands of various kinds of IQ and other standardized test scores of students I have worked with pass through my hands which has given me some experience in considering the limitations or usefulness of such testing.
A fluid interpretation of intelligence is a more acurate picture, I would agree -though I though that was clear from my criticisms of Murray's " hard" interpretation of "g".
Well, I am not an academic, so I like to get to the pragmatic essentials.
(i) Murray's grasp of population genetics is either (a) utterly impoverished due to ignorance or stupidity or (b) closed off due to prejudice;
(ii) his statistics is oft suspect or queerly framed;
(iii) his use of or framig of "g" runs against what one can say in re intelligence, as he seems to use it, in bio-genetic terms and rather abstracts away from complex environmental (e.g. nutrition, early childhood, etc) factors to imply rather strong conclusions with a strong Victorian Neo Social Darwinist hint.
Taking these things together, and taking his ... rather political approach and writing, I see him frankly as a vulgar 19th c. style polite bigot dressing up social darwinist ideas in polite language.
Pity, it's not as if he's a droolig knuckle dragger, or devoid of valid points, but my familiarity with human population genetics and the related statistics, as well as the background and his stridency leads me to conclude he is useless. There are better persons and arguments in the same frame without his strong stench of archaic racialism and vulgarly deterministic thinking.
(i) Murray's grasp of population genetics is either (a) utterly impoverished due to ignorance or stupidity or (b) closed off due to prejudice;
(ii) his statistics is oft suspect or queerly framed;
(iii) his use of or framig of "g" runs against what one can say in re intelligence, as he seems to use it, in bio-genetic terms and rather abstracts away from complex environmental (e.g. nutrition, early childhood, etc) factors to imply rather strong conclusions with a strong Victorian Neo Social Darwinist hint.
Taking these things together, and taking his ... rather political approach and writing, I see him frankly as a vulgar 19th c. style polite bigot dressing up social darwinist ideas in polite language.
Pity, it's not as if he's a droolig knuckle dragger, or devoid of valid points, but my familiarity with human population genetics and the related statistics, as well as the background and his stridency leads me to conclude he is useless. There are better persons and arguments in the same frame without his strong stench of archaic racialism and vulgarly deterministic thinking.
# posted by : Saturday, 20 January, 2007
: Saturday, 20 January, 2007
Col, on what grounds are you attacking his population genetics?
While I don't think Murray would deny the environment matters (with a culture) to g, I believe he'd probably agree with the consensus that the environment explains perhaps half of observed variation.
While I don't think Murray would deny the environment matters (with a culture) to g, I believe he'd probably agree with the consensus that the environment explains perhaps half of observed variation.
"factors to imply rather strong conclusions with a strong Victorian Neo Social Darwinist hint."
I'll give you that Murray gives off a Herbert Spencer fan club aura in his delivery.
Can't speak to his grasp of population genetics as I do not have a copy of The Bell Curve handy; I read it when it first came out and then loaned it to a colleague who was better at borrowing books than returning them. So I'll let Dan handle that point.
For me, while IQ has valid use in determining cognitive floors ( you can test well underneath your ability but not equally far beyond it if you are using any kind of a decent test) on an individual basis, neural plasticity should not be discounted lightly.
I'll give you that Murray gives off a Herbert Spencer fan club aura in his delivery.
Can't speak to his grasp of population genetics as I do not have a copy of The Bell Curve handy; I read it when it first came out and then loaned it to a colleague who was better at borrowing books than returning them. So I'll let Dan handle that point.
For me, while IQ has valid use in determining cognitive floors ( you can test well underneath your ability but not equally far beyond it if you are using any kind of a decent test) on an individual basis, neural plasticity should not be discounted lightly.
On the grounds of actual population genetics, including modern insights on allelic expression.
Really simple.
Nothing in modern population genetics supports the archaic victorianism of the Bell Curve.
If you boys want to run around w the Bell Curve author, go right ahead but I recognize archaic racialism and vulgar vicotianesque neo-social darwinism when I see it.
Interesting and productive things will come out of ongoing genetic research on individual genetic expression and trait distribution in the future. Nothing interesting comes out of the Bell Curve and its stat games and reification of "g."
Politics is its game, I have no use for it.
Really simple.
Nothing in modern population genetics supports the archaic victorianism of the Bell Curve.
If you boys want to run around w the Bell Curve author, go right ahead but I recognize archaic racialism and vulgar vicotianesque neo-social darwinism when I see it.
Interesting and productive things will come out of ongoing genetic research on individual genetic expression and trait distribution in the future. Nothing interesting comes out of the Bell Curve and its stat games and reification of "g."
Politics is its game, I have no use for it.
# posted by : Sunday, 21 January, 2007
: Sunday, 21 January, 2007
By the way, with respect to the articles, what I take away is an old man pining for an elitist and somewhat dirigiste type system, rather than letting free choice dictate, and effort select. I prefer to let students' free choices dictate their paths.
# posted by : Sunday, 21 January, 2007
: Sunday, 21 January, 2007
Col, still not sure what you mean by either "modern population genetics" or "archaic victorianism" in your sentence, "Nothing in modern population genetics supports the archaic victorianism of the Bell Curve."
Among TBC's controversial claims
1. The human race is composed of identifiable breeding populations whose attributes' means vary to each other
2. Intelligence is well defined by a single variable, g, which remains stable over life
3. g lays on a unimodal normal distribution for any population
One of these? All? Something else?
Among TBC's controversial claims
1. The human race is composed of identifiable breeding populations whose attributes' means vary to each other
2. Intelligence is well defined by a single variable, g, which remains stable over life
3. g lays on a unimodal normal distribution for any population
One of these? All? Something else?
I'm not particularly inclined to blither on about The Bell Curve as I find it most uninteresting, but with respect to population genetics and
TBC's controversial claims you list:
1. The human race is composed of identifiable breeding populations whose attributes' means vary to each other
This is by all available data clearly and evidently not true (presuming you mean in your summary, there are coherent populations with discrete genetic attributes, as in 'black race.') The past decade of population genetic mapping (variation of alleles distributions across populations) rather clearly indicate that the classic race model is incoherent, the origin populations for the recent African inputs to American "black" populations are among the most genetically diverse (i.e. variable) in the world, and generally no large scale population group lacks alleles present in another (with the exception of some small isolated populations with apparently recent mutations).
I have other things to do than trot out my old hobby from my days of working on gene engineering business, but generically speaking all evidence runs against the idea that there are coherent large populations which are across numerous genes have coherent allelic patterns.
But that was poorly expressed and I am too lazy to edit.
Moving on:
2. Intelligence is well defined by a single variable, g, which remains stable over life
There simply isn't genetic level understanding of this, and of course I am sure you are well aware of the robust objections to the "g" concept as such.
Put this in the area of modesty and humbleness w respect to current data would require one not to make strong arguments.
3. g lays on a unimodal normal distribution for any population
See above.
One of these? All? Something else?
In effect, Murray has a pre-conceived conclusion on which he hangs his collected data, rather than building up.
Vulgar and archaic social darwinism trying to dress up for a new era.
I do not question in any way that some -often significant- portion of individual capacity comes from genetic heritage, nor that small population groups (e.g. on the tribe level) will be found that are genetically coherent to be lightly differentiated (although human genetic variability is very low). That does not lead me into vulgar Victorian like social darwinism.
TBC's controversial claims you list:
1. The human race is composed of identifiable breeding populations whose attributes' means vary to each other
This is by all available data clearly and evidently not true (presuming you mean in your summary, there are coherent populations with discrete genetic attributes, as in 'black race.') The past decade of population genetic mapping (variation of alleles distributions across populations) rather clearly indicate that the classic race model is incoherent, the origin populations for the recent African inputs to American "black" populations are among the most genetically diverse (i.e. variable) in the world, and generally no large scale population group lacks alleles present in another (with the exception of some small isolated populations with apparently recent mutations).
I have other things to do than trot out my old hobby from my days of working on gene engineering business, but generically speaking all evidence runs against the idea that there are coherent large populations which are across numerous genes have coherent allelic patterns.
But that was poorly expressed and I am too lazy to edit.
Moving on:
2. Intelligence is well defined by a single variable, g, which remains stable over life
There simply isn't genetic level understanding of this, and of course I am sure you are well aware of the robust objections to the "g" concept as such.
Put this in the area of modesty and humbleness w respect to current data would require one not to make strong arguments.
3. g lays on a unimodal normal distribution for any population
See above.
One of these? All? Something else?
In effect, Murray has a pre-conceived conclusion on which he hangs his collected data, rather than building up.
Vulgar and archaic social darwinism trying to dress up for a new era.
I do not question in any way that some -often significant- portion of individual capacity comes from genetic heritage, nor that small population groups (e.g. on the tribe level) will be found that are genetically coherent to be lightly differentiated (although human genetic variability is very low). That does not lead me into vulgar Victorian like social darwinism.
# posted by : Sunday, 21 January, 2007
: Sunday, 21 January, 2007
Col,
We agree on #2 and #3. I think the same level of openness should be applied to #1, as well.
Lewontin, and others, are right to say that about 85% of genetic variation on randomly selected allelles is between individuals while only 15% is between populations. One can correctly say, as you did, that there are not "discrete" races. However, it's a jump to go from there to the idea that there are not races (or at least race-like variation) of humanity.
Lewontin's calculations did not account for the clustering of variation (a visual at Wikipedia). Evidence for the existence of large-scale breeding populations can be found from the genes currently under selection, which tend to be within racial groups more often than between them (see gnxp for more). Whether or not these are "races," or "clines," or whatever is under debate, and whether or not these changes effect intelligence is unknown.
Further, it is possible that there are Old World clines and New World races. Even if genetic variation was smoothly distributed in the old world, immigration to the new world may not have been smoothly distributed. Thus, even if in Europe "French" and "German" allelles blend into each other, if most of the French came from around Orleans while most of the Germans came from around Hannover, New World French and German may form "races" even if in Europe they only form "clines."
As to which of these possiblities (no variation, clines, races, clines/races, something else?) is true: we don't know. Murray's position is as valid (tentative, perhaps incorrect) as Gould's mismeasure-of-man thinking.
However, it is as incorrect to decry those who observe group-level genetic diversity as "social darwinistic Victorians" (or whatever) as it is to decry those who deny it as politically-correct New Leftists.
We agree on #2 and #3. I think the same level of openness should be applied to #1, as well.
Lewontin, and others, are right to say that about 85% of genetic variation on randomly selected allelles is between individuals while only 15% is between populations. One can correctly say, as you did, that there are not "discrete" races. However, it's a jump to go from there to the idea that there are not races (or at least race-like variation) of humanity.
Lewontin's calculations did not account for the clustering of variation (a visual at Wikipedia). Evidence for the existence of large-scale breeding populations can be found from the genes currently under selection, which tend to be within racial groups more often than between them (see gnxp for more). Whether or not these are "races," or "clines," or whatever is under debate, and whether or not these changes effect intelligence is unknown.
Further, it is possible that there are Old World clines and New World races. Even if genetic variation was smoothly distributed in the old world, immigration to the new world may not have been smoothly distributed. Thus, even if in Europe "French" and "German" allelles blend into each other, if most of the French came from around Orleans while most of the Germans came from around Hannover, New World French and German may form "races" even if in Europe they only form "clines."
As to which of these possiblities (no variation, clines, races, clines/races, something else?) is true: we don't know. Murray's position is as valid (tentative, perhaps incorrect) as Gould's mismeasure-of-man thinking.
However, it is as incorrect to decry those who observe group-level genetic diversity as "social darwinistic Victorians" (or whatever) as it is to decry those who deny it as politically-correct New Leftists.
Col,
Errr...how did you move from genetics to developing market financial expertise ? That's quite a hop, skip and a career jump!
Well done on the directorship thing, BTW.
Errr...how did you move from genetics to developing market financial expertise ? That's quite a hop, skip and a career jump!
Well done on the directorship thing, BTW.
Well, on the substance, I disagree with the cite on a number of areas, but frankly am not motivated to pursue this further. I will just observere that while I am probably about 3-5 yrs out of date, my last understanding of what the data showed fairly excluded the idea that any of the classic "races" could represent a coherent genetic grouping. Further to that, given the extremely diverse genetic heritages of American blacks, I find the proposition there must be (and it must be the preferred explanation) some coherent genetic explanation when one has evident socio-economic factors at work to fairly reek of an agenda.
I take The Bell Curve as a package, and that's why in the aggregate of its argumentation I label it vulgar Victorianesque social darwinism. If one abstracts away from its entirity, pretends the argumentation was not pushing the boundaries of what could be supported at the time, and takes each in isololation, well, of course you can quibble.
In any case, rather evidently our opinions are not far apart on the substance.
As for Mark's question:
Business. I am not a scientist. I worked on the business end of genetic engineering in management. Obviously that required a certain engagement with the science, off of which since I was bored with plants and the like, I developed a hobby looking at human population genetics literature.
In any case, after Monsanto fucked gene engineering with their lawyer driven idiocy, I decided to move into finance. Or specifically private equity.
Post a Comment
I take The Bell Curve as a package, and that's why in the aggregate of its argumentation I label it vulgar Victorianesque social darwinism. If one abstracts away from its entirity, pretends the argumentation was not pushing the boundaries of what could be supported at the time, and takes each in isololation, well, of course you can quibble.
In any case, rather evidently our opinions are not far apart on the substance.
As for Mark's question:
Business. I am not a scientist. I worked on the business end of genetic engineering in management. Obviously that required a certain engagement with the science, off of which since I was bored with plants and the like, I developed a hobby looking at human population genetics literature.
In any case, after Monsanto fucked gene engineering with their lawyer driven idiocy, I decided to move into finance. Or specifically private equity.
# posted by : Monday, 22 January, 2007
: Monday, 22 January, 2007<< Home
