Placebo has been shown to be a powerful analgesic with corresponding reduction in the activation of the pain matrix in the brain. However it is not clear whether the placebo response is reproducible within individuals and what role personality traits might play in predicting it. We induced placebo analgesia by conditioning subjects to expect pain reduction following a sham-treatment in the guise of a local anaesthetic cream applied to one arm. Pain ratings were assessed before, during and after treatment. The procedure was repeated in a second session to assess the degree of reproducibility of the response. A high degree of correlation was found between the two sessions for the sham-treatment group (R(2) = 0.55; p < 0.001). Personality questionnaires were given during both experimental sessions to assess key traits such as optimism and state and trait anxiety. A regression model was used to statistically define a placebo responder in terms of personality scores. High dispositional optimism and low state anxiety were found to be significant predictors of placebo response. We suggest that repeated placebo responders are high in dispositional optimism and having a placebo response in the first session causes a drop in state anxiety at the beginning of the repeat session.