The present study tested the hypothesis that positive mood facilitates cognitive flexibility in categorization, i.e., positive mood enhances the ability to categorize flexibly at broad and narrow levels contingent on task requirement. The study was a conceptual replication of the 1990 work by Murray, Sujan, Hirt, and Sujan. Unlike the original study, it was carried out on the WWW with some modifications in the procedure. Participants were 29 women and 132 men whose average age was 29 yr. They were put into a positive, negative, or neutral mood using the Velten technique. Mood was measured with standard mood scales. Subjects then listed similarities differences between two entertainment TV shows. Contrary to the prediction, participants in a positive mood did not list a higher number, more diverse, or more unusual similarities and differences between the shows, and they were not faster than participants with neutral or negative moods. It is concluded that the influence of mood on cognitive flexibility needs to be reconsidered. In research, temporal aspects of categorization performance as well as a wider range of mood valences should be taken into account.