Heredity-environment analyses of Jencks's IQ correlations.

Three different published heredity-environment analyses of Jencks's summary correlations for IQ have yielded strikingly different results. It is shown empirically that differences in selection of data and in computational procedures and logical inconsistencies in specifying equations are not responsible for the differences in results. Rather, the differences trace to the underlying assumptions made by the various authors. The analyses suggest that the assumptions concerning genetic dominance, assortative mating, and special twin environments were especially critical, while those regarding selective placement and different modes of environmental transmission were not.