A meta-analysis of psychological and pharmacological treatments for social phobia was conducted to evaluate whether the various treatments differ in their efficacy for treating social phobia, whether they are more effective than wait-list and placebo controls, whether rates of attrition differ, and whether treatment gains are maintained at follow-up. A total of 108 treatment-outcome trials for social phobia met inclusion/exclusion criteria for the meta-analysis. Eleven treatment conditions were compared: wait-list control, pill placebo, benzodiazepines (BDZs), selective serotonin reuptake inhibitors (SSRIs), monoamine oxidase inhibitors, attention placebo, exposure (EXP), cognitive restructuring (CR), EXP plus CR, social skills training, and applied relaxation. The most consistently effective treatments for social phobia were pharmacotherapies. BDZs and SSRIs were equally effective and more effective than control conditions. Dropout rates were similar among all the active treatment conditions. Assessment of the durability of treatment gains for pharmacotherapies was not possible because an insufficient number of drug studies included follow-up data. The treatment gains of psychological therapies, although moderate, continued during the follow-up period. BDZs and SSRIs seem to be effective treatments for social phobia, at least in the short term. Recommendations for future research include assessing the long-term outcome for pharmacotherapies and evaluating the inclusion of a cognitive-behavioral treatment during the drug tapering period.