We tested the hypothesis that alcoholics develop a disease-related attentional bias. Therefore, alcohol-related, but task-irrelevant, words should cause a specific perceptual-processing bias. We investigated this by using a special color-naming task. We subjected 40 male alcohol-dependent inpatients and 40 healthy male controls (matched according to age and verbal IQ) to a modified card version of the Stroop color-naming task that consisted of a neutral and an alcohol word condition ("Alcohol Stroop"). Alcoholic inpatients performed significantly poorer than the control group under the critical experimental condition (color-naming of disease-related words), as compared with the noncritical condition (color-naming of neutral words; p = 0.03). Concerning the possible neuropsychological impairment of the patients, no effects could be found on the reaction time of the "Standard Stroop" using only neutral words (i.e., color-naming of incongruent color words administered without time limitation). The information processing bias on the "Alcohol Stroop" thus qualifies as a cognitive process, which is independent from putative neuropsychological deficits of alcoholic patients and might represent an essential feature of alcoholic psychopathology. The "Alcohol Stroop" contributes to the experimental psychopathology of alcoholism.