Abstract
Test-retest data were obtained binocularly, monocularly and monocularly in the presence of glare from 20 healthy older patients using the Pelli-Robson and Regan (96% and 11% contrast) charts. Reliability was assessed by means of two indices: intra-class correlation coefficient and coefficient of repeatability which determines the 95% confidence limits for change. Binocular viewing showed the highest intra-class correlation and the smallest confidence interval with all the charts. Monocular viewing in the presence of glare showed the lowest correlation and the largest confidence interval. Results suggest that when assessing the significance of change, the confidence interval should be determined for identical viewing conditions.