PBP
Most of the criticisms can be summarised as “you didn’t talk about x” or “what about x?”.
Certainly not mine.
It was meant to have summarised reasons and evidence for the genetic origin of the gap
And this is fine, but it did not do that. Instead, it listed off reasons which could potentially make it plausible if we subscribe to other notions you believe but which you left unstated. You failed to connect the dots. If you’re going to make a case for something like group differences being genetic, merely saying that group differences correlate with g loadings and g loadings correlate with heritability does not do that for you. This is especially the case since the relationships are not strong enough to be transitive, and thus do not imply (look up the meaning of “imply” and compare it to the word “suggest”) any necessary level of co-consideration for these observations. You will have to do more to connect them.
Finding an invariant common pathway model would connect them since a higher-order g would necessarily lead to a Jensen effect because loadings, with invariance, must be collinear with intercepts. If this model fit, these things you’ve discussed would be observations within the same model, and thus the relationship between g, group differences, and heritability, could not be ignored. That is not what the existing evidence is.
If you wish to make this claim, go fit the model with the NLSY 79 and 97 or any dataset including the ability to fit this class of behaviour-genetic model. If you elect for a bifactor model with the ASVAB tests, note that you will not find a very strong Jensen effect because of the confounding role of psychometric s which I mentioned. You can also make the argument likelier by looking at longitudinal datasets like the ECLS-K or NICHD and showing that the longitudinally — and developmentally — consistent (major) portion of the difference between the groups is g and that this is the case within a model generating a Jensen effect.
no, not that the gap is fully genetic
I should not have to clarify that I never said there was anything to do with the gap being “fully genetic.” However, the title certainly suggests that it is “fully genetic,” or at least so “genetic” that it would be odd not to conflate the two. If you’re not going to claim the gap is “fully genetic,” try to figure an amount based on the sources you’ve cited, or note that the surveys you cited are consistent with experts believing in a low level of genetic influence on the gaps between ethnoracial groups.
That was not the point of the post, so it’s more like cherry-picking instead of criticism.
The point of the post, if the title is to be believed, is “why the Black-White IQ gap is almost certainly genetic.” If you’re saying that your cited evidence wasn’t relevant to the title, then I’m not sure what your post is about anymore. The idea that I’m cherrypicking is misplaced; what I’m doing is pointing out that the “evidence” you cite for your title point can only be construed as evidence given unstated assumptions and arguments which you’ve failed to include.
Shuey 1966 does do a meta-analysis, so I don’t see the point in mentioning the need for a meta-analysis.
At best, what Shuey did can be described as “an attempted review of prior studies.” She missed many and there was no meta-analysis of their results, nor was there even an attempt to compare expectations and observations. The point of mentioning the need for a meta-analysis is that these results have not been meta-analyzed — ever — and they should be, especially if they’re going to be used as evidence for something like this. If they’re evidence for this, then the results should conform to predictions. Do they? I don’t see that being a strong conclusion based on Shuey’s data.
I didn’t say admixture studies are direct evidence
To quote your post:
Direct genetic estimates have also been used and the usual correlation was found.
If you meant only to say that direct genetic estimates of admixture have been used, that is fine. But, you continued and said that “the usual correlation” was found. What is “the usual correlation?” and how does the first study of the admixture-IQ relationship since the 1970s comport with “the usual?” There’s no calculating of expectations in that article, but if you use the values in their supplement, it sure looks like the observed within-Black correlation is greater than the expected value with 100% heritability. How exactly does this work? Should a value of 100% be considered “usual?” Are you making a statement to the effect that “the difference is 100% due to genes based on this result, but it’s also enhanced due to environments?” Is the shared environment reinforcing differences? How would this then square with evidence from adoption experiments? How, if this is the case, can the environmental and genetic influences be disentangled? There is absolutely no clarity here, and this is a major problem with extrapolations based on naïvely produced expectations.
On variance between and within siblings
What you’ve cited is not about “variance between and within siblings,” it’s about a sibling control model. This is a conceptual confusion: the “variance within” sibling pairs is known to be constrained compared to the population because sibling traits are correlated, but this is completely unrelated to the colourism test.
Some valid criticism is that I was too hasty to jump to conclusions “the obvious conclusion is that the gap is genetic”. Sure, but what I meant is that the likely explanation is the genetic one.
Confusion again! “Likely” is a probability statement, and you have not made any of your evidence understandable in terms of a probability! Just estimate the between-group heritability using some of this evidence once, or talk about expected amounts of genetic effect, the proportion of the gap due to genes, or transitivity from behaviour-genetic analyses to group differences — something, which establishes any degree of likelihood, at the very least!
The rest of the post is either “You didn’t mention x”
I didn’t include any of this. You must be reading a different post. If you think asking you to state what a given source says about group differences is equivalent to “You didn’t mention x,” that’s just foolish.
it fails to misunderstand some arguments made
I agree. I do fail to misunderstand.
I hope you will go back and make probability statements understandable in light of some sort of theory or interpretation of results which does make that possible, instead of leaving them bereft of content. There is no “crux of what [you’re] getting at” if there’s no connection between anything written and the argument — that “the Black-White IQ gap is almost certainly genetic.”
P.S.
With values of approximately 0.6 for the correlation between group differences and g loadings and around 0.3 between g loadings and heritabilities, you end up with too little to justify a necessary relationship and, thus, too little to justify an argument for the relevance of this finding to understanding the origins of group differences instead of individual differences within groups.
