This paper, to be published in Intelligence, naturally caught my eye. The British psychologists Richard Lynn and David Longley review published studies of the gentile-Jew IQ gap and report analyses of other data not specifically directed at this question. A summary of their conclusions:

• The mean IQ of American Jews is estimated to be 110. This represents an upward revision since Lynn's last look at this question.
• The mean IQ of British Jews, based on several studies that are admitted by the authors to be unsatisfactory, is estimated to be 111. This estimate does not include the results of the studies described in the next two bullet points.
  
• 3,374 babies were born in Britain during the first week of March 1946. Twenty-six years later, 22 of these cohort members identified themselves as Jews. At age 8 all cohort members had been given tests of pictorial reasoning, reading comprehension, vocabulary, and "word reading." Lynn and Longley summed the scores on these four tests and converted the aggregate to an IQ scale. The Jews outscored their gentile peers by 8.5 IQ points (p = 0.03).
  
• 36 Jews and 11,104 gentiles were born in Britain during the first week of March 1958. All available cohort members were given mental ability tests at ages 7, 11, and 16. Attrition led to only 29 Jews being assayed at age 16. The authors claim that the Jewish children significantly outperformed their gentile peers in every test-by-age category at the 0.001 level, but my inspection of their table and some quick plug-and-chug on my calculator suggests to me that the p-values here may represent yet more data-reporting errors on the part of Lynn. Nevertheless, the overall trend is unambiguous. For example, at age 16 Jews outscore gentiles by 8 IQ-scale points on a reading comprehension test and by 12 points on a math test (both highly significant).
  
• Between 1901 and 1940, Jews were overrepresented by a factor of 3.7 among Fellows of the Royal Society (according to the Society's website, Fellowship is an intellectual distinction for scientists "second only to the Nobel Prize"). Between 1950 and 2003, they are overrepresented by a factor of 7.6.
• Four of Britain's 77 Nobel Prize winners have been Jews ("Bernard Katz, medicine, 1970; Brian Josephson, physics, 1973; Herbert C. Brown, chemistry, 1979; Harold Pinter, literature, 2005"). This is an overrepresentation by a factor of 8.0. I have calculated the significance of this result by the Poisson approximation to the binomial. The probabiliy of observing this many Jews among Britain's Nobelists given no gentile-Jew disparity in "Nobel juice" is ~0.002.
• The authors conclude their paper with an endorsement of the Cochran-Hardy-Harpending hypothesis.

Tangent: Another paper in press at Intelligence reports that individuals with less fluctuating asymmetry (FA) tend to have higher IQs. This line of research comes out of Geoffrey Miller's theory that g is subservient to a "super-g" of general fitness that has been under sexual selection in humans. With respect to FA, the idea is that individuals with "bad genes" do not develop stably and thus end up with asymmetric facial features, lower-quality brains, and so on. I withhold judgment on this theory for now. I will say, however, that it must be determined whether the correlation between FA and IQ holds within families. That is, we must see whether the sibling with the more symmetrical features also tends to have the higher IQ. If we do not find this, then it must almost certainly be the case that the correlation arises because of cross-assortative mating: status/wealth (determined in large part by IQ) and health/attractiveness are both highly valued traits, and thus the genes for both tend to be sorted into the same families. In fact, I find cross-assortative mating to be the most likely explanation for the FA-IQ relationship at present. And wouldn't any evolutionary theory of racial differences in IQ based on sexual selection have to be pretty weird?