« The Rebirth of Hebrew | Gene Expression Front Page | other psychometrics »
June 28, 2005

honest disagreement

I would like to suggest that many of the people who argue against large genetic influences on behavior are actually doing so in good faith. They assert that environment influences behavior more than genes do because it is obvious that this is the case. It is, for instance, obvious that the difference between the behavior of the African American students and the European students in a typical urban school is trivial compared to the difference between the Franks and their genetic decendents the French, or even compared to the difference between Athens 700BC, Athens 200 BC and Athens 300 AD or between Rome in 0AD, Rome in 1000AD and Rome in 2000AD, or Scandinavia in 1600 and 1900. In a given century, the difference in violence between Columbia and Costa Rica (both Hispanic, 9-fold), between Japan (both East Asian 18-fold), or between Russia and Sweden (both European, 25-fold). By comparison, Linda Gottfredson asserts that intelligence accounts for a 7-fold difference in incarceration rate. If you have any familiarity with anthropology or with history then the evidence that culture matters immensely more than genetics is simply obvious.

The shocking thing which most people don't realize is the underreported lack of measurable statistical consequences of parenting to the development of broad and predictively important measures of general ability or temprament. Obviously, parents can do much to encourage the development of expertise, from provision of a multi-lingual background to raising the Polgars or the Williamses, and expertise is far more impressive than sheer IQ in terms of the magnitude of the difference in ability it creates, and this also makes it seem that parents can obviously make their children smarter. They can do so, but a) in the vast majority of cases they don't, and b) psychometrics are intentionally defined in such a manner as to be resilliant to cultural biases, training, etc. Importantly, defining psychometrics in such a manner as to define traits which are resistant to change does not leave us with predictively unimportant traits. Instead, these traits, which seem to almost entirely reflect genes and random developmental patterns are on a statistical level an extremely good predictor of life-outcomes.

Posted by michaelv at 04:17 PM