The aim of this study was to investigate the effectiveness of pulsed short wave (PSW) in the relief of pain in osteoarthritis of the hip and knee. Ninety-two patients, mean age 63 years, (34 men and 58 women) were randomly allocated to one of three groups: (1) Active PSW, using the dosage found in a pilot study to be non-significantly most effective, (2) Placebo PSW, (3) No treatment control group. Nine sessions of treatment were provided over a 3-week period, each application lasting for 15 min. The machine was modified by the manufacturers so that the therapist was able to administer the treatment and carry out assessments without being aware of the treatment allocation. Outcome measures included sensory and affective pain diary reports averaged over days and weeks, self-reported benefit and the General Health Questionnaire. Analysis of variance with repeated measures over time was used to find out if the active treatment had a specific effect, incremental to the placebo effect. There were no significant differences between the active and placebo groups over time. According to the pain diary reports, both active and placebo groups tended to improve slightly during treatment, but worsened after its withdrawal. Patients who were given the placebo application tended to report more benefit than those who had the active treatment, although this did not quite reach statistical significance (P < 0.06). Patients who were not on a waiting list for surgery did significantly better over time than those who were (P < 0.03). There were no significant differences between the groups over time for the other outcome variables. Any treatment effect on this patient population appears to have been largely placebo-mediated. No evidence was found therefore for the specific effectiveness of PSW for treatment of osteoarthritic hip or knee pain.