The current study examined electrophysiological measures of cognitive efficiency in alcoholics and controls using a negative priming paradigm derived from DeSchepper and Treisman (J. Exp. Psychol. Learn. Mem. Cog. 22 (1996) 27). In this task, sets of novel shapes were presented: Two overlapping green and red shapes on the left and a single white shape on the right. Participants were instructed to ignore the red shape, but to determine whether the green shape was the same as or different from the white shape. On primed trials, previously red (to be ignored; i.e., irrelevant) shapes became green (relevant) shapes in a second component of the task. Participants who were capable of ignoring irrelevant stimuli were expected to experience more difficulty in the primed condition. Therefore, we predicted that, if alcohol-related impairment in cognitive efficiency is due to inability to respond "appropriately" (i.e., ignore irrelevant stimuli), alcoholics would experience less negative priming than normal controls. Both amplitude and latency of the P300 component of the event-related potential were measured in response to each trial condition. Using a 2 x 2 (group x condition) ANOVA with repeated conditions, a significant pattern of group x condition interactions was observed at right frontal, F4 (p=.05) and central parietal, Pz (p=.03) electrode sites on measures of P300 amplitude. A significant group x condition latency interaction was observed at the central parietal electrode, Pz (p=.006). Overall, controls exhibited increased P300 amplitude and latency in response to negatively primed trials. As predicted, alcoholics did not demonstrate this pattern, a finding consistent with alcohol-related cognitive inefficiency.