Abstract
In several aspects of nighttime driving our estimates of the visual abilities of ourselves and of others is important. Our decision to cross a street at night, for example, may be influenced by our estimate of the visual abilities of the approaching driver. Existing data reveal that pedestrians dramatically overestimate drivers' visual abilities. This study explored the relationship between observers' estimated and actual visual abilities in a lab setting. Young observers (N=34) briefly wore one of four different goggles containing ND filters (ND 3, 2, 1, or 0). They then removed the goggles and estimated their high contrast visual acuity (VA), low contrast VA, and contrast sensitivity (CS) at the previously seen luminance. To estimate their VA, observers sat 1 m from a monitor and matched the size of a high or low contrast E to their estimated threshold if they were viewing the E at 6 m. Similarly, observers used a Pelli-Robson chart to estimate their CS at the different luminances. Surprisingly, observers significantly underestimated their actual abilities. The magnitude of the underestimation in low contrast VA grew as luminance increased but was consistent (∼.25 logMAR) for high contrast VA. The opposite pattern was observed for CS, where the magnitude of the underestimation decreased as luminance increased. Using photographs, observers also estimated the maximum distance at which they could detect a pedestrian walking in the middle of their lane when they drive along a dark rural road at night. These visibility estimates were made for low and high beams and for different clothing conditions. Observers consistently underestimated the importance of both luminance (beam setting) and contrast. Estimated pedestrian visibility distances were best predicted by high luminance estimates of CS (r=.54) and by high luminance estimates of low contrast VA (r= −.47) but not by measures of actual visual ability. Implications for nighttime traffic safety will be discussed.