3 ms·

Yes, his review is scathing. "I started to laugh in the methods section. The authors constructed their key measure (called “internal consistency”) to make it es

Yes, his review is scathing. "I started to laugh in the methods section. The authors constructed their key measure (called “internal consistency”) to make it essentially impossible for them to not get the results they got. Or put another way, the study was basically rigged (although not necessarily consciously so), as subsequently shown by three other groups also published in PNAS." So he accuses those he disagrees with of having committed science fraud. Not only that, this fraudulent science has passed peer review and been published in PNAS. It is odd that it has not been retracted if it was so laughable as Cahill claims! He also claims that these "ideologues" have such a strong grip on academia that they will destroy your career if you oppose them: "Senior colleagues warned me as an untenured professor around the year 2000 that studying sex differences would be career suicide." It is smearing, which he offers no evidence for, hoping that the readership will already "know" it is true. Maybe he is right, maybe he is not, I can't tell. If he is right then I hope more neuroscientists speak up about the bullying and that journals like PNAS stops publishing junk science.

int_19h 8y ago
I don't think he's claiming that the study is outright fraudulent, but rather that the data in it cannot be summarized simply as "male and female brains cannot be easily distinguished" without using faulty methodology to analyse it. The accuracy of the data itself is not disputed, so the study is still valuable for that reason alone. Indeed, one of the other studies that he cites in support of his claims used that same data to deconstruct the model. And it really does seem to be as blatantly fudged as he claims: "Via simulations, we systematically varied sample characteristics such as the magnitude of sex differences and correlations among variables. Some of the simulated scenarios were intentionally unrealistic, involving uniformly strong sexual dimorphism and/or extremely high correlations between variables (up to r = 0.90). Despite this, the proportion of “internally consistent” profiles remained low in all conditions. ... To further reinforce this point, we applied Joel et al.’s methods to facial morphology features in three species of monkeys (crab-eating macaques, grivets, and tufted capuchins). Our goal was to see what percent of individual monkeys would display internally consistent species-typical profiles of features. Across comparisons, only 1.1–5.1% of the monkeys showed consistent “species-typical” profiles ..." https://www.pnas.org/content/113/14/E1965.full https://www.pnas.org/content/113/14/E1965.full

int_19h 8y ago

I don't think he's claiming that the study is outright fraudulent, but rather that the data in it cannot be summarized simply as "male and female brains cannot be easily distinguished" without using faulty methodology to analyse it. The accuracy of the data itself is not disputed, so the study is still valuable for that reason alone. Indeed, one of the other studies that he cites in support of his claims used that same data to deconstruct the model. And it really does seem to be as blatantly fudged as he claims: "Via simulations, we systematically varied sample characteristics such as the magnitude of sex differences and correlations among variables. Some of the simulated scenarios were intentionally unrealistic, involving uniformly strong sexual dimorphism and/or extremely high correlations between variables (up to r = 0.90). Despite this, the proportion of “internally consistent” profiles remained low in all conditions. ... To further reinforce this point, we applied Joel et al.’s methods to facial morphology features in three species of monkeys (crab-eating macaques, grivets, and tufted capuchins). Our goal was to see what percent of individual monkeys would display internally consistent species-typical profiles of features. Across comparisons, only 1.1–5.1% of the monkeys showed consistent “species-typical” profiles ..." https://www.pnas.org/content/113/14/E1965.full https://www.pnas.org/content/113/14/E1965.full