Demographic data were analyzed with Pearson chi-square (categorical variables) or linear mixed models (continuous variables) (Table 1). The HR and LR offspring did not significantly differ in gender composition or length of follow-up. HR offspring had a significantly earlier age of study entry (F(1,207)=12.95, p<.001), though age at last follow-up was not significantly different. The HR group had statistically significantly lower SES (F(1,207)=22.08, p<.001) than LR controls, though this may not reflect meaningful differences because both groups fall within the same Hollingshead SES social strata (medium size business, minor professional, technical). The success of the selection strategy used to define the high and low risk groups can be seen in the differing number of first and second degree relatives with AD. HR offspring had an average of 4.61 first- and second-degree relatives with alcohol dependence, compared to an average of 0.64 affected relatives in LR controls.
