You’d have to raise the bar really far before any actual gender-based differences showed up. It seems far more likely that the cause is a cultural bias against intellectualism in women (women will under-report IQ by 5ish points and men over-report by a similar margin, women are poorly represented in “smart” jobs, etc.). That makes women present themselves as less intelligent and makes everyone perceive them as less intelligent.
Does anyone know of a good graph that shows this? I’ve seen several (none citing sources) that draw the crossover in quite different places. So I’m not sure what the gender ratio is at, say, IQ 130.
La Griffe Du Lion has goodwork on this, but it’s limited to math ability, where the male mean is higher than the female mean as well as the male variance being higher than the female variance.
The formulas from the first link work for whatever mean and variance you want to use, and so can be updated with more applicable IQ figures, and you can see how an additional 10 point ‘reporting gap’ affects things.
Unfortunately, intelligence in areas other than math seem to be an “I know it when I see it” kind of thing. It’s much harder to design a good test for some of the “softer” disciplines, like “interpersonal intelligence” or even language skills, and it’s much easier to pick a fight with results you don’t like.
It could be that because intelligence tests are biased toward easy measurement, they focus too much on math, so they under-predict women’s actual performance at most jobs not directly related to abstract math skills.
Of course, if you use IQ testing, it is specifically calibrated to remove/minimize gender bias (so is the SAT and ACT), and intelligence testing is horribly fraught with infighting and moving targets.
I can’t find any research that doesn’t at least mention that social factors likely poison any experimental result. It doesn’t help any that “intelligence” is poorly defined and thus difficult to quantify.
Considering that men are more susceptible to critical genetic failure, maybe the mean is higher for men on some tests because the low outliers had defects that made them impossible to test (such as being stillborn)?
The SAT doesn’t seem to be calibrated to make sure average scores are the same for math, at least. At least as late as 2006, there’s still a significant gender gap.
Apparently, the correction was in the form of altering essay and story questions to de-emphasize sports and business and ask more about arts and humanities. This hasn’t been terribly effective. The gap is smaller in the verbal sections, but it’s still there. Given that the entire purpose of the test is to predict college grades directly and women do better in college than men, explanations and theories abound.
Not a rigorously conducted study, but this (third poll) suggests a rather greater tendency to at least overestimate if not willfully over-report IQ, with both men and women overestimating, but men overestimating more.
You’d have to raise the bar really far before any actual gender-based differences showed up. It seems far more likely that the cause is a cultural bias against intellectualism in women (women will under-report IQ by 5ish points and men over-report by a similar margin, women are poorly represented in “smart” jobs, etc.). That makes women present themselves as less intelligent and makes everyone perceive them as less intelligent.
Does anyone know of a good graph that shows this? I’ve seen several (none citing sources) that draw the crossover in quite different places. So I’m not sure what the gender ratio is at, say, IQ 130.
La Griffe Du Lion has good work on this, but it’s limited to math ability, where the male mean is higher than the female mean as well as the male variance being higher than the female variance.
The formulas from the first link work for whatever mean and variance you want to use, and so can be updated with more applicable IQ figures, and you can see how an additional 10 point ‘reporting gap’ affects things.
Unfortunately, intelligence in areas other than math seem to be an “I know it when I see it” kind of thing. It’s much harder to design a good test for some of the “softer” disciplines, like “interpersonal intelligence” or even language skills, and it’s much easier to pick a fight with results you don’t like.
It could be that because intelligence tests are biased toward easy measurement, they focus too much on math, so they under-predict women’s actual performance at most jobs not directly related to abstract math skills.
Of course, if you use IQ testing, it is specifically calibrated to remove/minimize gender bias (so is the SAT and ACT), and intelligence testing is horribly fraught with infighting and moving targets.
I can’t find any research that doesn’t at least mention that social factors likely poison any experimental result. It doesn’t help any that “intelligence” is poorly defined and thus difficult to quantify.
Considering that men are more susceptible to critical genetic failure, maybe the mean is higher for men on some tests because the low outliers had defects that made them impossible to test (such as being stillborn)?
The SAT doesn’t seem to be calibrated to make sure average scores are the same for math, at least. At least as late as 2006, there’s still a significant gender gap.
Apparently, the correction was in the form of altering essay and story questions to de-emphasize sports and business and ask more about arts and humanities. This hasn’t been terribly effective. The gap is smaller in the verbal sections, but it’s still there. Given that the entire purpose of the test is to predict college grades directly and women do better in college than men, explanations and theories abound.
Not a rigorously conducted study, but this (third poll) suggests a rather greater tendency to at least overestimate if not willfully over-report IQ, with both men and women overestimating, but men overestimating more.