The Nature of Human Intelligence (Sternberg, 2018)

[richardbaxter]

Here is a good counter argument to the effect of 50k years of divergent evolution on observed racial differences;
Flynn, J. (2018). Intelligence, Society, and Human Autonomy. In R. J. Sternberg (Ed.), The Nature of Human Intelligence (pp. 101~115). Cambridge University Press.

[ofd]

What is the argument?

[barn]

If I found the right one, it's this (from here) :

Abstract

IQ scores are volatile indices of global functional outcome, the final common path of an individual’s genes, biology, cognition, education, and experiences. In studying neurocognitive outcomes in children with neurodevelopmental disorders, it is commonly assumed that IQ can and should be partialed out of statistical relations or used as a covariate for specific measures of cognitive outcome. We propose that it is misguided and generally unjustified to attempt to control for IQ differences by matching procedures or, more commonly, by using IQ scores as covariates. We offer logical, statistical, and methodological arguments, with examples from three neurodevelopmental disorders (spina bifida meningomyelocele, learning disabilities, and attention deficit hyperactivity disorder) that: (1) a historical reification of general intelligence, g, as a causal construct that measures aptitude and potential rather than achievement and performance has fostered the idea that IQ has special status and that in studying neurocognitive function in neurodevelopmental disorders; (2) IQ does not meet the requirements for a covariate; and (3) using IQ as a matching variable or covariate has produced overcorrected, anomalous, and counterintuitive findings about neurocognitive function. (JINS, 2009, 15, 331–343.)

While here you can check out, Robert J. Sternberg's bio. (have a gander at his "Administrative Experience")

[barn]

Hi thinkers and alike,

What a great interview the FDR team has just put out, fully packed and really interesting...

(quote from the video's description: )

 "Human intelligence is an important construct in psychology, with far-reaching implications, providing insights into fields as diverse as neurology, international development, and sociology. Additionally, IQ scores can predict life outcomes in health, education, work, and socioeconomic status. Yet, students of psychology are often exposed to human intelligence only in limited ways. ... We found that 79.3% of textbooks contained inaccurate statements and 79.3% had logical fallacies in their sections about intelligence."

The study can be found, here.

Good mind expanding,

Barnsley

[barn]

Supplementary for this thread, worth a read-through, following up.

(Genetics of Brain Structure and Intelligence by Paul Thompson and his team)

take you (pdf) there

[richardbaxter]

The central argument provided here for a substantial independence of measured general intelligence (IQ) from genetics is the Flynn effect. There are also anecdotes/examples provided of children being raised with european parents, and how this reduces intelligence differences typically observed across race.

Another important evidence is that intelligence/IQ genetic contribution measurements are different in low socioeconomic environments; https://journals.plos.org/plosone/article/figure?id=10.1371/journal.pone.0030320.t001 (with these differences being observed across racial groups also; http://science.sciencemag.org/content/174/4016/1285).

Copying here my analysis provided https://youtube.com/watch?v=jvwG367izwg;

Note (7:30) it is important to distinguish superiority from intelligence. The claim that high intelligence implies superiority is philosophical not scientific. Intelligence is just one trait that nature selects for depending on the environment.

General intelligence as measured by IQ has high heritability (it is typically measured to be approximately 70% genetic, increasing with age). Measurements of trait heritability differ however depending on the environment in which they are measured, for example in low socioeconomic environments the relative genetic contribution to intelligence differences is much lower; being almost entirely dependent on the opportunity to access education/resources. A significant component of general intelligence as measured by IQ is environmental/cultural (the Flynn effect suggests that average intelligence has been increasing over the last century). The question of whether average differences in IQ (reflective of actual general intelligence) across racial groups are genetic is presently unestablished. The reason being that it is difficult to partial out environment differences (for example differential treatment by parents/peers based on the race of the subject and/or the culture of the group). Once the genetics of intelligence differences has been mapped at a molecular level (DNA) it will be possible to answer.

[barn]

Hi @richardbaxter

I'm still going through the massive update you've provided, maybe it's there... so, I apologise for being impatient but if you don't mind I'd like to know:

How extensively have you looked into twin-studies?

[richardbaxter]

Twin studies are used to estimate genetic/environmental contribution to traits, and I am familiar with their strengths and limitations. Do you have a question regarding them?

Like all statistical tests, they are only valid under the assumptions of the model being tested. Copying here a summary of the prototypical method for reference;

Genetic determination/contribution of a trait in twin studies is measured relative to that of the environment (e.g. society/home/etc). Twin studies consist of both identical (monozygotic) and non-identical (dizygotic) twins.

In a simple ("ACE") model, the correlation (r) of a measured trait between two persons (their phenotype) is subject to the similarity of their genotype (how much genetic material they share), along with;

A) the heritability of the trait,
C) their common/shared environment, and;
E) their unique environment

A high correlation (r) of a trait means that both persons either exhibit it or do not exhibit it, and a low correlation means that it is completely random as to whether they both exhibit it. The unique environment (E) has the potential to significantly affect the correlation of the trait between the persons. The shared environment (C) does not have the ability to significantly affect the correlation between the persons. The heritability of any trait will vary on its own accord (some traits are completely determined by the environment while others are almost completely hard coded, e.g. eye colour), but the capacity of the heritability (A) of the trait to affect the correlation will depend on how much genetic material is shared between the persons (e.g are they identical twins, fraternal twins, siblings, cousins, etc). The total contribution of all components to the correlation will (by definition) never vary, and so in this model A, C, and E can be set to equal 1 (ie, A + C + E = 1).

It might now be worth imagining some specific trait, say extraversion.

1. The correlation (r_MZ) of the trait between monozygotic (MZ) twins in a family is adversely affected by E) only, their unique environment. This is the case because their genetic material is identical, and they are assumed to share a common environment (home). Note this is mathematically equivalent to saying that the correlation is wholly determined by A) the heritability of the trait and C) their common environment, as E = 1 - (A + C). Therefore, we set r_MZ = A + C.
2. The correlation of the trait between dizygotic (DZ) twins is similarly determined by A) the heritability of the trait and C) their common environment, but they only share 50% of their twin's genetic material, and as such we set r_DZ = 1/2xA + C.

We can now solve for A, E and C. Assume a study has been conducted on a large sample of twins and both r_MZ and r_DZ have been measured.

i)
r_MZ = A + C. (1.)
C = r_MZ - A

r_DZ = 1/2xA + C (2.)
C = r_DZ - 1/2xA

Substitute 1. rearranged into 2. rearranged;
r_MZ - A = r_DZ - 1/2xA
r_MZ - r_DZ - A = - 1/2xA
r_MZ - r_DZ = 1/2xA
2x(r_MZ - r_DZ) = A
A = 2x(r_MZ - r_DZ)

ii).
r_MZ = A + C (1.)
r_MZ = 1 - E
E = 1 - r_MZ

iii)
A + C + E = 1
C = 1 - A - E
Substitute in ii);
C = 1 - A - (1 - r_MZ)
C = - A + r_MZ

The total genetic component (heritability) of the trait is A. In the case of extraversion, this is commonly measured to be approximately 0.5 (50%).

When interpreting heritability estimates from twin studies an important factor one needs to consider is the gene-environment (GxE) interaction effect. The realisation of a genetic predisposition in one's phenotype is subject to their environment. For example, heritability (genetic component) estimates of IQ are usually quite high (e.g. 0.8), but in low socioeconomic environments they can become negligible.

In the case of non-separated (at birth) twin studies there is serious confound potential, because visually and/or behaviourally (note appearances are strongly genetically determined, and personality is ~50% genetic) similar identical twins may be treated more similarly by their parents (shared environment) than visually/behaviourally discrepant non-identical twins. This is particularly pertinent for traits which have a low base (incidence) rate, because it is really difficult to obtain high enough sample sizes of separated twins who demonstrate such traits. This is not the case for intelligence however.

Twin studies also can't rule out a prenatal hormonal (not necessarily genetic) effect. 75% of monozygotic (identical) twins share the same placenta, but ~0% (sometimes fused) of dizygotic (non-identical/fraternal) twins share the same placenta. This means that their shared in utero chemical exposure will differ depending on whether they are monozygotic or dizygotic.

Again, it is important to note that even in the case where a genetic component of a trait has been identified (established), it doesn't necessarily mean that the trait (its propensity or existence) will be determined by the individual's genetics in every environment. This is because of gene-environment interaction (not to be confused with exclusive environmental effects which affect every individual of a species the same regardless of their genome). Some traits, despite being demonstrably genetically determined, will in many environments be largely (or even exclusively) determined by that environment - for example due to the presence of an environmental catalyst. To use a biological analogy; a DNA readout promoter/blocker. In reality this might be something more like education. Therefore, the absolute genetic component determination calculations (ie x%) are only meaningful for the environment of the population being studied. It might be that there exist (existed) cultures which are so different than ours (that in which the genetics of the trait is being studied) that the trait would never arise, always arise, or be entirely dependent on an environmental factor which is taken for granted as "normal" in our culture - i.e. be measured as 0% genetically determined in their own culture.

[barn]

Hi @richardbaxter

 

On 09/25/2018 at 3:58 PM, richardbaxter said:

Twin studies are used to estimate genetic/environmental contribution to traits, and I am familiar with their strengths and limitations. Do you have a question regarding them?

A follow up question? No, not that I can think of. For now that is.

Thanks.

I see more provided in your response than what I originally, at first, thought I'd wanted to ask. (seems pretty intentional, re: on the clarity of your stance too... makes for comparing things easier (usually). I like when people choose that, thanks again.)

 

refs.

About twin studies (,in general)

 

Supplementary:

Gregor Mendel (curious just how many leading scientific minds were supported, I guess 'cared for' by the Church back then)