Differential heritability and consistency: a reanalysis of the National Merit Scholarship Qualifying Test (NMSQT) California Psychological Inventory (CPI) data.

Abstract

Several biometrical genetic models were fitted to the National Merit Scholarship Qualifying Test (NMSQT) California Psychological Inventory (CPI) data. These models include parameters for an additive genetic effect (D R), an among-families environmental effect (E 2), and a withinfamilies environmental effect (E 1). It is shown that the CPI scales do not fit the same model; the heritabilities, calculated from the parameter estimates of the model fit for each scale, are not the same. Most scales fit a model with parameters for the additive genetic and within-families environmental effects, although the CS and IE scales requireE 2 to achieve fit, and the RE, AI, and FE scales require separate-sexE 1 andE 2 parameters. The SO and CM scales fit none of the models tested. No model could be fit to the SO scale; CM did not fit these models because of apparent gene-environment interaction effects. Similar parameter estimates were found for the models fit to the full sample and the two random subsamples, and the heritabilities calculated from these estimates were quite consistent.