The placebo response forms a growing problem in randomized, placebo-controlled clinical trials in psychiatry. Research into the placebo response is on the increase, but remains very limited in relation to social phobia. Together with the dropout rate, the placebo effect is an important factor limiting the discriminative properties of any study. In this study, we reviewed 15 placebo-controlled studies in social phobia, focussing on patients and study characteristics. In social phobia, the placebo effect has turned out to be moderately large and has shown no increase over the past decade. Placebo response was highest in large, multicentred trials and was independent of study duration. No validation for a placebo run-in was found. Taking into account both response to placebo and active drug, as well as dropout rate, the most discriminative results are probably to be expected in a sample of patients who are moderately to severely impaired. More research in the field of the placebo response is needed.