Recent Psych Science paper used national IQ to predict some stuff:

I ran some simple probability calculations and the measure they used says that 43% of Africans are mentally retarded (IQ < 70).

Says that MAJORITY of people in some countries are mentally retarded.

WTF even...

Oh and that Lynn national IQ measure doesn't come from representative samples. In fact, @JelteWicherts found that a good predictor of which samples were included for national IQ in Africa was...drum roll....

Lower average IQ scores!

Add to that, tons of countries in the measure used by the Psych Science paper didn't actually have ANY samples of IQ data. So Lynn just took averages of surrounding countries.

44% of the African countries were imputed.

So it's bad data generating worse data.

Why would the guy who came up with this measure make such poor analytic choices? Hmmm....

Enough Lynn, what about PS paper? Statistically, I don't know what one could do in this situation to generate any sort of solid inference. And a multiverse analysis (which the PS paper did) isn't addressing the right problem: garbage in the spreadsheet.

And yet, they conclude:

Wild speculation unsupported by shaky (to be generous) data. I just...

I should probably channel this from Tweets into a formal commentary or something.

Anyway, here are 10 countries that this dataset says are mostly intellectually disabled, despite the fact that THEY HAVE NO DATA FROM THESE COUNTRIES. None.

Seems like those countries have something in common...

replace terms in OP:

intellectual disability <- mental retardation

My bad. Got riled.