We conducted a systematic review of randomized control trials examining the efficacy of self-hypnosis as a clinical treatment. Searching for 'self-hypnosis', 'selfhypnotic', 'autosuggestion', and 'autohypnosis' returned 576 studies, of which 22 met the definition of being an RCT. Self-hypnosis has been reported to be efficacious in studies of pain, childbirth, paediatric applications, stress and anxiety. Methodological differences among studies are discussed. Self-hypnosis is most likely to be effective when taught as an independent self-directed skill and when it involves at least three practice sessions before participation in the trial. Experience of heterohypnosis does not seem to be essential in producing an effect for self-hypnosis. Studies reporting no effect typically involved participants listening to audio recordings of hetero-hypnosis only. Meta-analysis revealed a medium-to-large effect size for self-hypnosis in clinical treatment.