Reliability of a standardized reading chart system: variance component analysis, test-retest and inter-chart reliability.

BACKGROUND

The purpose of this study was to analyze statistically the test-retest and inter-chart reliability of three standardized Radner Reading Charts with respect to reading acuity, maximum reading speed, critical print size and LogRAD/LogMAR ratio and to quantitatively identify the sources of variability.

METHODS

At two testing sessions (3 to 4 weeks apart), 36 volunteers read monocularly all three Radner Reading Charts in randomized order (orthogonal Latin square design). The subjects were grouped according to their distance LogMAR visual acuity (group 1, 0.1 or better; group 2, 0.12-0.4; group 3, 0.42-0.8). Based on these 216 reading tests, crude correlations were calculated between the two testing sessions and between the three charts, and "variance component analyses" were performed for all variables.

RESULTS

For all visual acuity groups and variables, the test-retest and the inter-chart reliability was high. The variance component analyses showed that the differences between individuals predominantly influenced the variance, whereas the effect of the test replication was low. The three charts caused minimal variability, indicating equality for clinical examinations.

CONCLUSION

The standardized Radner Reading Charts provide clinically reliable and reproducible results for individuals with normal eyesight and for patients with visual impairment. These findings indicate that reading test systems, which consider the current international standards for visual acuity measurements (EN ISO 8596, NAS-NRC) and the psychophysical requirements for controlling optical item interactions, can provide reliable measures for clinical and scientific analyses of reading performance.