Affect change scores were calculated by subtracting baseline reports of NA and PA from second reports, measured 15 minutes after completion of beverages (alcohol Group BrAC = .087). Using this change score approach controls for individual differences in preexisting (baseline) levels of PA and NA, which are not relevant for testing current hypotheses, and focuses analyses on effects associated with the beverage group manipulation. Mean levels of PA and NA at baseline and post-drinking are reported in Table 1. Affect change scores were subjected to a 3 (Beverage) × 2 (Scale; NA, PA) mixed factorial ANOVA, which showed a significant Beverage × Scale interaction, F(2, 64) = 13.63, p < .001. This interaction was probed using separate one-way ANOVAs for NA and PA. The ANOVA on NA change scores showed a significant Beverage effect, F(2, 64) = 5.20, p < .01. Dependent t tests showed that alcohol group participants reported feeling less negative affect post-drinking compared with baseline (M = −1.52), t(22) = −2.82, p = .01, whereas placebo participants reported feeling somewhat (though not significantly) more negative affect post-drinking
