4 ms·

>the "quotient" is a ranking mechanism for solving a certain type of paper puzzle, not an actual thing. Sure, except it strongly correlates with pretty much ev

>the "quotient" is a ranking mechanism for solving a certain type of paper puzzle, not an actual thing. Sure, except it strongly correlates with pretty much every single measure of fluid intelligence you can think of. And with a billion real life outcomes, besides. It just so happens that the best statistical distillation of intelligence measures happens to take the form of "paper puzzles". We can, I guess, do a 12 hour barrage of 50 different intelligence tests, but what's the point, when one does the job. >So two hypotheses There are plenty of other hypotheses.

notahacker 4y ago
> Sure, except it strongly correlates with pretty much every single measure of fluid intelligence you can think of. And with a billion real life outcomes, besides Sure, test scores correlate with other test scores (surprisingly badly) which correlate with other proxy measures of intelligence. Quelle surprise! There's a massive difference between acknowledging a test is somewhat useful in ranking peers and agonising over whether one cohort getting scores in a particular test a couple of points lower than another cohort from a completely different background represents some real world shortcoming of the latter cohort in anything other than that particular set of puzzles. The research establishing a lack of support for supposed "dysgenic" trends (though absence of evidence isn't necessarily evidence of absence) because the drop happens within families might be superficially interesting, but we're looking at tiny points of difference in a noisy series. There's typically more variation between Weschler and Raven tests adminstered to the same individuals than the measured generational "IQ decline" the headline is encouraging us to agonise over. When all the points of data other than IQ tests suggest that subsequent generations haven't struggled intellectually compared with their forebears, there's not much reason to suspect the trends in paper puzzle scores are a signal and not just noise.

notahacker 4y ago

> Sure, except it strongly correlates with pretty much every single measure of fluid intelligence you can think of. And with a billion real life outcomes, besides Sure, test scores correlate with other test scores (surprisingly badly) which correlate with other proxy measures of intelligence. Quelle surprise! There's a massive difference between acknowledging a test is somewhat useful in ranking peers and agonising over whether one cohort getting scores in a particular test a couple of points lower than another cohort from a completely different background represents some real world shortcoming of the latter cohort in anything other than that particular set of puzzles. The research establishing a lack of support for supposed "dysgenic" trends (though absence of evidence isn't necessarily evidence of absence) because the drop happens within families might be superficially interesting, but we're looking at tiny points of difference in a noisy series. There's typically more variation between Weschler and Raven tests adminstered to the same individuals than the measured generational "IQ decline" the headline is encouraging us to agonise over. When all the points of data other than IQ tests suggest that subsequent generations haven't struggled intellectually compared with their forebears, there's not much reason to suspect the trends in paper puzzle scores are a signal and not just noise.