From your lips to god’s ears

Oh that makes total sense actually. Very cool! Hopefully people take notice and we can leave heritability and polygenic scores in social sciences behind us

Every day brings worse news no matter which country it’s about

Awesome! I had a stray thought along these lines once, so it's cool such an in depth assessment of how well it actually works! Is there a risk it ends up picking up pop strat/structure captured by GRM in the same way those can inflate GREML heritability?

I think the other plot is more damning, honestly. They can claim that "ancestry differences in genetics" drive the relationship, but a dummy continent variable shouldn't take that signal out of the genetic scores. It's also a big red flag that scores don't predict within regions at all

Well, HLO is just a nice World Bank dataset on learning outcomes/academic achievement. The IQ freaks won't care, but global development researchers might. The plot basically shows that African countries often seem systematically lower in IQ given their HLO, & NIQs show a weaker relationship to HLO.

It made me realize just how hollow the whole thing was when their response to that was "having low IQ is actually just different if you're white than if you're Black". Shameless ad-hoc reasoning and directly undermines their whole project if you take that to it's logical conclusions.

But there are even more basic problems that stem from outright fraudulent research practices and an obsession with eugenics and race science

Oh tons of problems, the tests aren’t necessarily from the UK but everyone is compared to the UK baseline. But It’s definitely not clear that the test scores mean the same thing in different cultural or economic contexts

At the end of the day almost all they ever do is type lm(race~IQ) into an R terminal, but there are increasingly complex looking things used to dress it up at more meaningful or rigorous. The genomic stuff is flashy enough to trick most people

Weird facet of the “analysis”. Scores are normed against the UK mean, which is 100

The problem was well stated by @sashagusevposts.bsky.social here and elsewhere. It's a part of the broader right wing ecosystem of scams and grifts, turning people in to marks willing to accept increasingly blatant cons.

The most basic of data exploration and quality checks would have identified this as a clear spurious correlation from confounded data. Same as with the National IQ case. It's deceptive and plays to people's confirmation bias. The goal is to get people to think every correlation you like is causal.

That Amsterdam group is really top notch. They're the reason I still have hope for psychometrics

The obvious question when Lynn's data is so unquestionably flawed, but other datasets "show the same patterns" is 'why not use the other datasets'. This seems to answer that question: they like the cooked numbers and want to them to keep saying Africans are inferior

This was the exact same problem that Jelte Wicherts and colleagues pointed out (among many others). Lynn cherry-picked studies to depress the IQ of African countries and there's noticeable divergence between datasets for Africa. jeltewicherts.net/wp-content/u...