So… I think the correlation between IQ and WORDSUM is mostly mediated by education (i.e., in terms of Stuff That Makes Stuff Happen, there’s an arrow from IQ to education and one from education and WORDSUM—there’s also one directly from IQ to WORDSUM but it’s thinner). So I’d expect that the third graph in that article to show an effect more extreme than if you used IQ instead.
So… I think the correlation between IQ and WORDSUM is mostly mediated by education (i.e., in terms of Stuff That Makes Stuff Happen, there’s an arrow from IQ to education and one from education and WORDSUM—there’s also one directly from IQ to WORDSUM but it’s thinner). So I’d expect that the third graph in that article to show an effect more extreme than if you used IQ instead.