Parametric statistics were performed, and an α-value of 0.05 was set as the significance threshold, unless corrected by the Bonferroni method. Our main hypothesis was that selection for differences in alcohol preference changed temporal discounting. Because each experiment was performed with a high-alcohol drinking line and a low-alcohol drinking line, we assessed the effect of Alcohol Preference in each study in an overall Experiment × Sex × Alcohol Preference ANOVA, but collapsed across Experiment and Sex in the absence of significant interactions between alcohol preference and these other 2 independent variables. The ns for experiment 1 were HAP2 10(m), 6(f); LAP2 10(m), 7(f), and for experiment 2 they were HAP1 8(m), 10(f); HS/Ibg 10(m), 7(f). In the presence of significant interactions, data were stratified by factor and analyzed separately for simple main effects and/or group differences. Huyn-Feldt corrections were used where appropriate.
