The slope of the psychometric function for Bailey-Lovie letter charts: defocus effects and implications for modeling letter-by-letter scores.

BACKGROUND Visual acuity measurement often results in an imprecise endpoint because subjects correctly identify some but not all of the letters on one or more size levels on a letter chart. The extent of this transition zone from seeing to nonseeing can be described by probit size, which is calculated by performing Probit Analysis on letter chart data. There has been no previous research into the effects of optical defocus on letter chart probit size. METHODS We tested 18 young visually normal subjects monocularly during three different defocus conditions: best spectacle correction (zero defocus) and +1.00 D and +2.00 D additions. Stimuli were Bailey-Lovie-style logarithm of the minimum angle of resolution (log MAR) letter charts constructed with a 0.05 logMAR size progression between size levels. Frequency of seeing data from these charts were used to calculate probit size. RESULTS There were statistically significant effects of optical defocus on mean probit size. After Monte Carlo correction for bias, we believe that true mean values for probit size are about 0.07 logMAR for well-corrected subjects and up to 0.12 logMAR with optical defocus. CONCLUSION The smaller probit size for well-corrected subjects should correspond to a sharper logMAR visual acuity endpoint and less intrasubject variability in logMAR acuity than for subjects with a larger probit size (optical defocus). Our modeling shows that these different probit sizes can also significantly affect letter-by-letter visual acuity scoring.