It appears in larger populations because it's a weak proxy for other things.
It disappears in smaller populations because it's a weak proxy for other things.
The same relationship would exist with the MSRP of the car the test subject arrived in or the cost of their socks.
In a small population study you'd see no real correlation between how someone did a test and the price of their car.
But over say 100,000 people you would because the cost of someone's car weakly correlates with a number of things that are, in aggregate more advantageous for testing and so the confounding phenomena would arise.
In fact, the circumference of someone's head probably has about the same correlation with the price of their socks as it does with whatever intelligence metric your want to use because it's all measuring the same thing - social class, opportunities, economic arrangements, things like that.
It's a physical consequence of our unequal society that then gets used as if it's a biological reason for the perpetuation of the vary inequality that produced it.
Eugenicists called this biological determinism, it's an old idea that doesn't stand up to scrutiny.
> It's a physical consequence of our unequal society that then gets used as if it's a biological reason for the perpetuation of the very inequality that produced it.
The two population cohorts are 11-year olds and 17-year olds so regardless of what they say, we're still just measuring shoe size.
I'm guessing they published these two groups because their adult cohorts showed zero difference and that's not a publishable paper.
In fact, they admit it
"Two previous studies taking this approach, each with a sample size of about 35 sibling pairs, failed to find a significant correlation"
Right before they say "chill, let's cheat":
"We reexamined this empirical issue in a Human Connectome Project (HCP) (Van Essen et al., 2012; Van Essen et al., 2013) sample of 1,022 young adults"
^ it's a shoe size study from here out so of course they find something.