Variations within and between observer-judges reduce the accuracy of clinical research. Judgement Analysis allows strategies to be developed and applied which reduce variation in judgement. The prediction that the removal of important sources of error variance by this means would reduce the likelihood of committing a Type 2 Error was supported by the application of Judgement Analysis to assessments by 15 psychiatrists of 92 patients in a clinical trial of 2 antidepressive treatments. The statistical significance of differences between the effect of the treatments on the severity of depression was increased, and significant differences appeared earlier. Ten stimulated patient profiles were also converted into narrative case histories, enacted by experienced psychiatrists or psychologists and videotaped. The participants' judgements of the overall severity of the depression were in good agreement with those they had made on the original cases. Videotapes so prepared help training to reduce variation in observation, just as Judgement Analysis can lead to reductions in the variation of judgement.