Objective : To describe and compare performance-based and self-reported measures of visual function and to evaluate how each varied with age and how the performance-based measures were associated with the responses to self-reported questions about visual function. Design : Population-based epidemiologic study. Participants : Adults participating in the first follow-up of the Beaver Dam Eye Study (n=3,722). Main outcome measures : Performance-based measures of vision including current binocular, best-corrected and near visual acuities, sensitivity to light as measured by an automated perimeter, contrast sensitivity, and self-reported measures from a standardized interview. Results : All performance-based and self-reported visual functions decreased with increasing age (p<0.0001 for each). On average, women did more poorly than men on all the performance-based measures (p<0.0001 for each), and on some of the self-reported measures. For both genders, correlations among the performance-based measures ranged from 0.38 to 0.67. In general, individuals reporting good visual function or little difficulty with visual tasks scored better on performance-based measures compared to individuals whose questionnaire responses reflected limited visual function. Overall, about 39% of drivers reported that their vision caused them to limit their night driving. Correlations of performance-based with self-reported measures of visual function were greater for best-corrected, binocular and near visual acuity compared to correlations of visual sensitivity to light and contrast sensitivity. There were 1,285 persons whose best-corrected acuity was better than their current binocular acuity and there were 97 people with visual impairment considering only current binocular acuity whose vision improved to better than 20/40 after refraction. Conclusions : Performance-based and self-reported measures of visual function were consistently and inversely related to age. Performance-based measures were, in general, more highly correlated between themselves than were self-reported measures. Self-reported visual functions were significantly positively correlated with all performance-based measures, but correlations were moderate. While high contrast measures are usually used in clinical settings and in many studies, the data suggest that in assessing function it is appropriate to measure several different aspects of vision, both performance-based and self-assessed. Our data give some indication of the impact these visual functions have on daily life as reflected by night driving and on near and distant visual acuity. New refractive corrections would likely improve the visual acuity in many older adults.