BACKGROUND: In clinical optometric practice, autorefractors are used as an objective measure of refractive error prior to subjective refraction. We evaluate the clinical efficacy of autorefractor measurements by determining whether spectacles can be prescribed from autorefractor results. METHODS: Forty-seven subjects were randomly allocated spectacles using prescriptions determined either by an optometrist or from an autorefractor in a double-blind protocol. Subjects wore each prescription for two weeks and completed an oral questionnaire following each period of wear. The questionnaire assessed both pairs of spectacle lenses in terms of visual performance and ocular comfort. RESULTS: The level of negative responses to all questions was higher for spectacles based on the autorefractor findings than for those based on the optometrist's prescriptions. Having reported initial discomfort, many subjects indicated that they had adapted to the spectacle lenses over a two-week period. CONCLUSIONS: Our results suggest that, in a clinical environment, subjective refraction produces a more accurate and acceptable spectacle prescription than an autorefractor.