All data were examined for outlying data points, distribution normality, and missing values. To assure missing responses were not systematically biased by SOB, missing versus non-missing data for each phenotype scale was analyzed by SOB in 2×2 contingency tables. Examination of bias in missing data by DRD2 TaqI A and DRD4 VNTR genotypes has been conducted previously and are not reported on here [53], [54]. Missing values were not imputed, but excluded from analysis. Fits to HW equilibria were tested with the HWE program [90], with which the DRD2 HW equilibrium was tested with Fisher's Exact and DRD4 with the Markov Chain algorithm.
