These PDP estimates were submitted to a 3 (Beverage) × 2 (Process; automatic, controlled) mixed factorial ANOVA. This analysis showed a main effect of Process, F(2, 64) = 7.30, p < .01, which was qualified by a significant Beverage × Process interaction, F(2, 64) = 3.46, p < .05. Follow-up simple effect tests showed that whereas the automatic estimate was unaffected by beverage (Ms = 0.55, 0.58, and 0.57 for the alcohol, control, and placebo groups, respectively), ts < 1, the controlled estimate was significantly reduced in the alcohol group (M = 0.59) compared with both the control group (M = 0.75), t(43) = 4.08, p < .001, and the placebo group (M = .80), t(42) = 4.96, p < .001, which did not differ reliably, t(22) = 1.21, p > .20. These findings suggest that alcohol increases expression of race bias by limiting the extent to which control can be brought to bear once automatic stereotypes are activated (see also Bartholow et al., 2006).
