A computer-assisted subjective eye examination was administered to 78 patients as they volunteered. They were also examined in the conventional, manual way, some before and some after the computer-assisted examination. Results of the computer-assisted examination were compared to those obtained by conventional subjective methods. For the distance prescription, satisfactory agreement was obtained for 83% of the patients. The 17% unsatisfactory prescriptions were subdivided into about 6% avoidable errors (caused by errors in software, hardware, or optometric flow charts), about 6% errors of undetermined cause, and 4% fundamental errors (caused by patient's physical and mental impairment or confusion). A similar analysis is made for the near add. With further development of the system it is expected that between 90 and 96% of patients in an optometric clinic can be successfully examined by this computer-assisted method.