The graph above comes from Helmuth Nyborg's paper, Sex-related differences in general intelligence g, brain size, and social status, which I discussed in The Persecution of Politically Incorrect Scientists. In the paper, Nyborg looks at many other studies to try to determine whether sex differences in IQ exist and, if so, which sex does better on which subtests and by how much. He comes to the conclusion that women do slightly better in verbal tests, men better on spatial, and in sum, men have an IQ advantage over women of between 3 and 8 points, his best number being around 7 points. Men also have a wider dispersion of IQs than women.
The result, as the graph shows, is that at an IQ of 145, which could reasonably be deemed an elite level, men outnumber women by a factor of 8. Nyborg concludes that this must be highly significant for male dominance in "higher educational and socio-economic spheres", such as academia, science, and business. Add the male dominance in IQ at the highest levels to the fact that men are predisposed to take more risks and work longer hours, and I'd say that the feminist case for discrimination against women vanishes.
Of course, Nyborg was dealt a "Larry Summers" for showing this.
If women have a narrower distribution, and it seems people at the far left tail (criminals, mental incompetents, the insane) are a lot less likely than those at the right end of the tail to be sample, then the difference in means might have to do with sampling bias.
ReplyDeleteThere are other potential sampling biases too, women outnumber men in older populations, and due to the Flynn effect and declining cognitive ability with age this may be due to age effects.
I'll take a closer look at the paper later, I wonder if I can actually get my hands on the data. I'd like to see a comparison of estimated medians as well as means, as well as maybe the means of subset (IQ within 1 std. dev of overall average).
"women do slightly better in verbal tests, men better on spatial": so according to how you weight the two sorts of results, you can conclude that men are more intelligent on average, or less. And any weighting you choose is essentially arbitrary, surely? That's why the early IQ researchers weighted them to attribute the same mean to each sex, was it not? The discovery of different standard deviations for each sex goes back to those early researchers, doesn't it?
ReplyDeleteThe paper gets into a discussion of test-weighting and bias, but it's an extremely technical discussion. Read the paper if interested.
ReplyDeleteOn biological grounds it is clear that men and women would be expected to have different average intelligence (not identical) - and pretty much all the evidence is that there is a higher proportion of men than women of both higher and lower intelligence.
ReplyDeleteBut the exact size of the difference in average intelligence is not really meaningful because it would be affected by so many things.
It is rather like differences in average height between men and women - men are obviously taller than women, but the exact percentage difference varies according to place, population and circumstance.
Nobody gets hung-up over the exact number - because, being so labile, it is not really of much significance.
Intelligent women of course know all too well already, from their everyday life experiences, that it is so.
ReplyDeleteIt's the non-intelligent women who are objecting to these facts.
If scores on verbal subsections need to be balanced with scores on other subsections, does this mean Raven's Progressive Matrices is incomplete as an IQ test? I thought it was good - highly g-loaded - and that it had no words on it (at least none in any of the questions).
ReplyDeleteWithout Raven's, a lot of the cross-sex calibrating is going to be really difficult. Within a gender, IQ is correlated heavily with both bad behavior (negatively) and high academic achievement (positively). Yet women are vastly better behaved - if you took a mixed group of men and women you could probably completely obscure the IQ-crime correlation, the IQ-alcohol abuse correlation, etc. - all you'd have to do is mix women with an IQ averaging 95 ... with men averaging 105.
Put another way, IQ protects everyone from going to the devil, but being born a girl is pretty good protection too. Academic achievement, on the other hand, goes the other way, for reasons Summer mentioned.
Do you want to marry an intelligent woman? Sure!
Do you want to marry a woman who is as much of a geek and a monomaniac as the average MIT grad student? Heck NO.
So the indicators we use to prove the value of IQ are best within-gender. And wordless tests like Raven's prove* that, for all our bad behavior, men have the slight edge in raw brainpower.
* Or so I thought. I'm no expert.
"..the female g lead disappeared after Jensen eliminated the unusually large number of test items favouring females in the General Aptitude Test Battery..": I snort at the idea that you can reject a study because of an "unusually high" number of test items favouring females - what on earth has usualness got to do with science? I think that this chap is trying to use IQ or g to do a job that it's too crude for. Here's an analogy. Suppose you invent the WQ - the Wonderfulness Quotient. You apply it to two populations, call them A and B. In the subtest that looks at mathematics, the As score better on average; on the singing subtest, the Bs. You can now weight the two mean scores in any way you like, to prove either that Bs are more wonderful or As. Someone says "we need more subtests". So now you introduce a verbal reasoning test and a dancing test. Again, As do better on average on the first, Bs on the second. And on and on it goes - spatial reasoning versus sprinting, etc, etc. The problem is that the two populations have intrinsically different mean levels of ability at different things, and to combine the results into a WQ just obscures the usefulness of the detailed results, and the WQ still depends on weightings that are essentially arbitrary. It's the subtest results that would be consistent with the Olympic medals for B sprinters and the Nobel prizes for A physicists - the WQ has no relevance. Similarly for men and women - only the most foolish of people could suppose that it's remotely likely that the two sexes have identical mean abilities at different activities. From those different means - and different standard deviations - you probably can explain many of the differences you see in society. You don't need to combine all these subtest results into a single Quotient to explain the evidence of your eyes - indeed, constructing a single Quotient risks diluting the most relevant of the subtest results (whichever those might happen to be).
ReplyDeletePut another way, IQ protects everyone from going to the devil..
ReplyDeleteNot really. It was higher IQ Germans after all that brought us the ovens of the Holocaust, and it was higher IQ Russians (relatively speaking) that brought us tens of millions dead via gulags, purges and other "necessities" of the revolution. It is the same white Russians and Eastern Europeans that kill 2 babies in the womb for each live birth- the highest rate of abortion in the world (Loveless and Holman 2006). And higher IQ whites are proportionately the greatest consumers of pornography including child pornography (Tewksbury 2005). So High IQs are no indicator of virtue- far from it- the opposite "to the devil" pattern is well established.
ConservativeSwede is right. Most of the people that I find interesting intellectually are men, while I enjoy to talk to women about more wordly stuff since I can ramble more without having the person I talk to open up a newspaper. And all the smart women I know, are in the same position.
ReplyDeletedearieme, besides communicating ability, I find men, in my own life, vastly better at mostly everything. And even at communicating, if you want to stick to facts and solutions, men are better. It's just the part of the emotional communication that women handle a lot better.
TT, high IQ Russians? What the heck? In my country we had an idiot who couldn't even speak properly as dictator! lol. And people get abortions here because of the poverty and having to support everybody through welfare.
Nyborg's subjects seem to have been only Danes. I'd be curious to know if the patterns he found are present in other ethnic groups/races. Hasn't someone suggested that the average IQ of (West?) African women is higher than that of (West?) African men?
ReplyDeleteEileen said...
ReplyDeleteNyborg's subjects seem to have been only Danes. I'd be curious to know if the patterns he found are present in other ethnic groups/races. Hasn't someone suggested that the average IQ of (West?) African women is higher than that of (West?) African men?
There is every possibility that that is true. It make sense from an evolutionary perspective. Brains are expensive organs and if other selection pressures are not there (like females selecting for intelligence in their mates because the survival of their offspring depends on it) then you can expect to see IQ in the unselected sex decline.
See Peter Frost's blog for more interesting discussion about selection among different groups.
So, for women to earn the same as men, we must reproduce with stupid men. NICE! This must have been the plan of feminism all along. jk
ReplyDeleteThat different standard deviations in intelligence between men and women account for the overrepresentation of men at both the top of highly cognitive professional endeavors and in the criminal underclass strikes me as plausible and likely. That women have a 7 point IQ deficit compared
ReplyDeleteto men is problematic.
Think about it: a 7 point difference, while small compared to the black-white difference, approaches the Hispanic-white difference, which I take to be around 9 points. Yet the Hispanic IQ manifests itself in all kinds of both academic and social indicators, and obviously so. In contrast, I struggle to think of a way in which the female deficit shows up in disparate male-female outcomes among the broad middle of the population.
The entire point of IQ testing is to
predict these outcomes. Assuming the researchers did their math right, I would be forced to conclude that IQ really is sexually biased in that it underpredicts female performance, OR that the female deficit is compensated for by other traits, higher conscientiousness for example.
Greek sign poster, women do commit crime, just that not the one you think of. We commit less crime because our ways are manipulative, not coercitive. For example, half, if not more of rape allegations are false. A lot of DV allegations are false. We lie and manipulate to reach our goals, we don't bash heads in. So we are underrepresented in crime due to these reasons.
ReplyDeletephi,
ReplyDeleteHave you taken into account rampant affirmative action which has landed and kept women in jobs that should have gone to men? The disparate male-female outcomes you speak of (presumably having a job?) are a function of government intervention. They are not unbiased variables.
Further, from what I have read via Catherine Hakim's economic analyses, in Denmark and Sweden, 3/4th of employed women work in the public sector. It appears most women in highly feminist countries work in the non-competitive sector in job designations that otherwise wouldn't have existed.
Thus one cannot use a job as an unbiased indicator of disparate male-female outcomes or of female (or male) performance for that matter.