The objective of the study was to examine the efficacy of fluoxetine in social phobia. Sixty subjects were randomly assigned to 14 weeks of double-blind therapy with either fluoxetine or placebo. Dose was fixed at 20 mg for fluoxetine during the first 8 weeks of double-blind treatment; during the final 6 weeks, the dose could be increased every two weeks by 20 mg to a maximum of 60 mg/day. An intentto-treat analysis was used. A significant change from baseline to endpoint was found for both fluoxetine and placebo on the Liebowitz Social Anxiety Scale. However, no significant difference was found between fluoxetine and placebo. The change for fluoxetine was somewhat lower than that found with other selective serotonin reuptake inhibitors, whereas the placebo response was greater. Fluoxetine failed to separate from placebo in this trial. It is unknown whether a larger dose for longer duration would have yielded separation from placebo. A higher than usual placebo response rate was found.