The validity of reported ocular disease was investigated in a population-based epidemiologic study of persons aged 43-86 years residing in Beaver Dam, Wisconsin. In a telephone survey conducted from September 1987 through May 1988, histories of cataract and age-related macular degeneration were obtained from the subject for 2,155 cases and from a surrogate for 1,433 cases. Within 2 years, these persons underwent a complete ocular examination. At that time, an "in-person" self-reported history of eye disease was obtained and disease presence was determined based on ocular photographs. The reporting methods, telephone versus in-person and surrogate versus subject, were compared and the validity of each assessed. Reporting methods were in agreement in better than 90% of all cases. Reporting of cataract showed a sensitivity of 20.4 for surrogate by telephone, 30.2 for self-report by telephone, and 37.8 for self-report at the examination. Sensitivity of reported age-related macular degeneration was poorer, with the highest rate of 17.9 for the "in-person" self-report. Specificity was better than 90.0 for all reporting methods for both cataract and age-related macular degeneration. These data suggest that estimates of prevalence of ocular disease should not be based solely on reported histories, and that clinical determinations are necessary.