Stephen J Gould in "The Mismeasure of Man " is worth reading, for a deeper analysis.
I suggest in some of ny essays that the correlations presented among testing subfactors appear to be relatively weak and theory says that they are inappropriate computationaly in any case. Maybe the theory on computation is inappropriate or just plain wrong, but that is an argument for the mathematicians. It become more and more hard to fatham the more you look at it.
I have stated a few times that empiricism trumps theory, but I am not sure that a case has been made by the psychometrcians on IQ testing that I find plausible.
"Statistics claiming to apportion “percent of variance explained” to heredity or environment at the group level are logically and practically suspect, and even if they had some group-level meaning, they do not apply to individuals."
Why? How heretibility of IQ differs from heretibility of other traits?
Also g is not total competence, it's general intelligence
Group-level “percent variance explained” is a property of a statistical model describing how scores vary across a population; it is not a decomposition of an individual’s mind or abilities. Applying it to an individual is the ecological fallacy—inferring personal proportions from aggregate data.
Heritability of IQ differs from heritability of traits like height because:
IQ scores are constructed from ordinal, culturally bound tasks, not direct physical measurements.
Environmental and cultural influences on test performance are large, varied, and interact with genetics in complex, non-additive ways.
The g factor extracted from IQ tests is a narrow statistical abstraction, not total competence. It omits many skills and abilities that matter in real-world functioning.
Thus, heritability estimates for IQ are methodologically fragile, culturally contingent, and conceptually narrower than for traits grounded in direct, physical measurement.
“Group-level “percent variance explained” is a property of a statistical model describing how scores vary across a population; it is not a decomposition of an individual’s mind or abilities. Applying it to an individual is the ecological fallacy—inferring personal proportions from aggregate data.”
Are we talking about polygenic traits? Can't you say the same about all of them?
“The g factor extracted from IQ tests is a narrow statistical abstraction, not total competence. It omits many skills and abilities that matter in real-world functioning.”
But g isn't meant to be total competence. It's meant to measure general intelligence. The trait that makes competence among many different intellectual tasks correlate.
Stephen J Gould in "The Mismeasure of Man " is worth reading, for a deeper analysis.
I suggest in some of ny essays that the correlations presented among testing subfactors appear to be relatively weak and theory says that they are inappropriate computationaly in any case. Maybe the theory on computation is inappropriate or just plain wrong, but that is an argument for the mathematicians. It become more and more hard to fatham the more you look at it.
I have stated a few times that empiricism trumps theory, but I am not sure that a case has been made by the psychometrcians on IQ testing that I find plausible.
"Statistics claiming to apportion “percent of variance explained” to heredity or environment at the group level are logically and practically suspect, and even if they had some group-level meaning, they do not apply to individuals."
Why? How heretibility of IQ differs from heretibility of other traits?
Also g is not total competence, it's general intelligence
Group-level “percent variance explained” is a property of a statistical model describing how scores vary across a population; it is not a decomposition of an individual’s mind or abilities. Applying it to an individual is the ecological fallacy—inferring personal proportions from aggregate data.
Heritability of IQ differs from heritability of traits like height because:
IQ scores are constructed from ordinal, culturally bound tasks, not direct physical measurements.
Environmental and cultural influences on test performance are large, varied, and interact with genetics in complex, non-additive ways.
The g factor extracted from IQ tests is a narrow statistical abstraction, not total competence. It omits many skills and abilities that matter in real-world functioning.
Thus, heritability estimates for IQ are methodologically fragile, culturally contingent, and conceptually narrower than for traits grounded in direct, physical measurement.
“Group-level “percent variance explained” is a property of a statistical model describing how scores vary across a population; it is not a decomposition of an individual’s mind or abilities. Applying it to an individual is the ecological fallacy—inferring personal proportions from aggregate data.”
Are we talking about polygenic traits? Can't you say the same about all of them?
“The g factor extracted from IQ tests is a narrow statistical abstraction, not total competence. It omits many skills and abilities that matter in real-world functioning.”
But g isn't meant to be total competence. It's meant to measure general intelligence. The trait that makes competence among many different intellectual tasks correlate.
I agree with rest