[AstralStorm]

Did he even bother checking error bars due to multiple test correction when verifying the correlations in pooled data?

We could ask Cochrane what they find. They tend to use many more studies than 14 as input with clear inclusion and exclusion criteria.

I suspect the author misuses statistics and fails to check internal and external validity of his measures.

How did he even get the real IQ points when the test had been changed and regraded about every 20 years? How does he convert the numbers to g? Is it general or slanted? Is his measure verified to be consistent even in a single generation across cultures? Etc.

Unfortunately, the other paper is behind the paywall.