phill said:
You are claiming that any environmental factor which does not have a random distribution among populations makes a heritability estimate impossible?
Of course not. Where would you get that? I said it means you have to be very careful about how you define those populations, and completely thorough in dealing with confounding variables, to have the slightest hope of discovering valid genetic correlations even - let along causes.
And since this care has not been taken, and the due diligence not performed, the relationship between population genetics and population cognitive abilities is at the present time speculative.
phill said:
So, for example, national heritability of height estimations are also impossible because of varying availability of milk? This is nonsense. Heritability estimate methods are designed to account for this, by looking at the relationship between the variable in question and the outcome eg. Japanese in America.
Interesting example. Recent work by John Komlos and others have pretty much blown the long accepted and "scientific" estimates of the heritability of height at the population level. Komlos himself is on record as observing that as far as his work shows there is no significant variance in inherited height among the human populations on this planet - that aside from a couple of known physical subtypes of humans such as the Bushmen tribes in central and southern Africa, we are all genetically about the same height. The illusion was created by an uneven distribution of overlooked environmental factors.
phill said:
No, genetic clusters are formed independently of racial information and individuals are matched to the clusters post facto.
- - -
But defining race by ancestry or genetic similarity we do actually get clusters which match "social" races and indeed races inferred from non-metric cranial traits a la Blumenbach.
You are missing the point. That only gets you the individual, and their ancestral community - not the race. You can discover that one person's ancestry is a coastal community in Norway, and another person's is a coastal community in Portugal, and a third is descended from a coastal community in Sierra Leone. Evenly spaced along the eastern Atlantic ocean. But you cannot, from the cluster membership alone, assign two of them to one race and the third to another. You have to find out, first, what race those clusters belong to. And that will require sociological information.
The "genetic clusters" are gene pools, usually geographically located. They are not races. All the sociological races feature - within them - many such clusters. You can only identify the race associated with a particular cluster by gathering racial information - sociological racial information - about that community or geographical area.
Or as I put it: most people know what race other people think they belong to. So if you have a gene cluster, and you want to assign it to a "race" for some reason, the way to do it is by asking the members of that cluster what race they are. Or asking their neighbors. Or looking at photographs of the people who live there. Or guessing based on geography and history. Or guessing based on the sociological race assigned to some cluster you think is similar. Or some such method. You can't make the assessment from the genetic data alone, because there are no genetically defined races. No one has taken the human genetic heritage and divided it up into races based on its internal relationships and natural clustering. The species is divided into races first, sociologically, mostly on essentially trivial criteria known to be subject to rapid evolutionary change such as skin color, and only then are the "genetic clusters" assigned to one race or another - regardless of their genetic similarity to each other.
sculptor said:
Denying any genetic or regional differences in the test results is indeed denying the science of psychology!
Accepting the data as genuine, then leads to seeking an understanding of "Why?".
More to the point, "accepting" that the differences in the test results have been shown to have an identifiable degree or percentage of genetic cause on the basis of such horseshit evidence is denying all of the sciences except psychology. Including statistics.
For example: Anyone who "accepts" evidence for the genetic distribution of cognitive abilities on a US racial population scale, from research that did not control for childhood lead exposure, stereotype bias, and genetic vulnerability to lead exposure and stereotype bias (which is research I believe not to exist yet), is fooling themselves.
And when faced with otherwise competent, educated, professional people so willing to fool themselves in such obvious ways, the person seeking understanding is indeed led to ask "Why?".