The outcome and process of treating subclinical anxiety with self‐hypnosis and relaxation were compared. Twenty individuals who presented for treatment for ‘stress, anxiety, and worry’ were assessed (for anxiety and self‐hypnotizability), exposed to a 28‐day treatment programme (which involved daily measures of outcome and process variables), and re‐assessed (for anxiety). It was found that both self‐hypnosis and relaxation alleviated anxiety pre‐ to post‐treatment. Although there was no difference in the outcome data, throughout treatment self‐hypnosis rather than relaxation was associated with a greater sense of treatment efficacy and expectation and with a greater sense of cognitive and physical change. The findings are discussed in terms of the expectational and experiential aspects of self‐hypnosis, and their potential role in the perception, progress and impact of using self‐hypnosis in therapy. Copyright © 1999 British Society of Experimental and Clinical Hypnosis