Update, 2011: The original images in this post were apparently lost due to server update…
Oh dear, this is dry stuff, but hopefully at least somewhat important- from the Wikipedia page on Sex and Intelligence:
Deary et al. (2003) performed an analysis of an IQ test administered to almost all children in Scotland at age 11 in 1932 (>80,000). The average IQ scores by sex were 100.64 for girls and 100.48 for boys. The difference in mean IQ was not significant. However, the standard deviation was 14.1 for girls and 14.9 for boys. This difference was statistically significant. In the sample studied, 49.6% are girls and 50.4% are boys. Because of the difference in variance between the sexes, however, girls are in excess by 2% in the middle IQ range of 90–115. At the extreme IQ ranges, 50–60 and 130–140, boys make up 58.6% and 57.7% of the population (gaps of 17.2% and 15.4%) respectively. That is, boys were overrepresented amongst the lowest and highest IQ groups. It is generally observed that males tend to hit the most positive and negative performance results of many tests.
In order to make the point even clearer I made the following plot that shows the ratio of men (and women) with a given IQ based on that data (pink points plot the female ratio).
Thus really high IQ people are likely to be men (a sooper genius with an IQ 160 is a man with approximately .7 likelihood as shown by the black circles which represent the ratios for men with a given IQ on the x-axis). However, it is definitely questionable how good that one very old study is. I haven’t seen the IQ test that was used so I can’t honestly say if it seems reliable but based on my personal observations I suspect it wasn’t terribly good and underestimates the real difference between men and women. In addition, as some researchers have noted (Rushton?), it is quite possible that teenage girls have an advantage over teenage boys since they mature faster which might make even the averages different later on.
Anyway, wikipedia quotes a second study which found such differences even in teenagers:
Deary et al. (in press) compared IQ scores from 1292 pairs of opposite-sex siblings from the US National Longitudinal Survey of Youth. Siblings were used to control for background factors that differ between families. They describe finding a 1 IQ point sized difference in mean scores favoring males, which was significant in this sample. They describe finding larger differences in variance, with nearly twice as many males as females scoring in the top 2% (the IQ equivalent of scores above 130).
Thus, Wikipedia is on the side of Larry Summers. Based on hard data, women are a minority in the group of cognitively gifted people. Also, less women are total morons, but in a modern worldIsuspect a single genius can accomplish much more good than a single moron can cause damageasa single technologicalbreakthroughcreatedbya genius of onemancanbeusedbymillions to make the world more prosperous. In addition, geniuses probably favor positive-sum-cooperation while people of average or less than average intellect seem to prefer zero-sum competition.Thishoweverhasmademethinkthatthecurrentpractice
some grave consequences (also the practice of supporting retards but not gifted people common inpublicschoolsin many places). If the few highly gifted men are discriminated against, societies suffer. Of course, the gifted people might manage to prosper even if they are discriminated against, but that doesn’t mean any artificial obstacles in their path won’t be highly expensive for human well-being if their existence means less time to spend on doing some real work proper for a genius.