ASQ subscales (stimulant/ascending limb, sedative/descending limb) were examined separately, given a theoretical basis for predicting differential effects for stimulation/sedation as a function of OPRM1 (Ray and Hutchison, 2004) and laboratory self-administration behavior (de Wit et al., 1987). This decision was supported by evidence of moderate correlation between subscales (r = 0.54, p < .001) and somewhat higher internal consistencies for the separate scales (stimulant: α = 0.91, sedative: α = 0.91) compared to the total scale scores (α = 0.83). Participants were grouped into high- or low-sensitivity on each subscale by computing median splits (note that higher scores denote lower sensitivity to reported alcohol effects). Given the tendency for men to endorse a higher number of drinks than women, ASQ group assignments occurred separately for men and women (Bartholow et al., 2007).
