Test stimuli are rated less "good" following very good context stimuli than when presented either alone or following neutral context stimuli. This diminution in rating is called hedonic contrast. In two experiments, the degree of hedonic contrast depended on how subjects were instructed to categorize context and test stimuli. Contrast was substantially attenuated if context and test stimuli were said to belong to different categories. The effect was demonstrated for beverages (Experiment 1) and birds (Experiment 2). Stimuli's hedonic ratings were far less affected by other stimuli declared to belong to a different category than by stimuli declared to belong to a common category.