Electrochromic rearview mirrors can provide continuous levels of reflectivity and unobtrusive, automatic control. The availability of this technology has increased the importance of understanding how to select the best level of reflectivity for a given set of lighting conditions. For night driving with glare from following headlights, the best reflectivity level will always depend on a tradeoff among several variables. This study was designed to help clarify what variables are important and how they should be quantified.Twenty subjects, 10 younger and 10 older, performed a number of visual tasks while viewing stimuli through an electrochromic rearview mirror. Subjects were seated in an automobile mockup in a laboratory, and the reflectivity level of the mirror was changed before each of a series of discrete trials. On each trial, subjects saw reflected in the mirror a visual-acuity stimulus and a glare source of varying intensity. They performed three tasks: (1) judgment of the location of a gap in the acuity stimulus, (2) rating of their subjective confidence in the gap judgment, and (3) rating of the discomfort they experienced from the glare source.Results showed that: (1) visual performance decreased with decreasing reflectivity approximately as a linear function of log reflectivity, (2) subjects' confidence consistently underestimated their actual performance but otherwise accurately reflected the effect of reflectivity on performance, and (3) the effect of reflectivity on discomfort glare is not fully predicted by its effect on illumination at the subject's eye point.Implications of these findings for mirror design include: (1) the continuous reflectivity control provided by electrochromic mirrors should be of value in achieving optimal tradeoffs in the presence of glare, and (2) at least for discrimination tasks, even though changes in reflectivity are automatic and subtle, subjects recognize the resulting changes in seeing ability.