This paper reports on the usefulness of five brief tests of cognitive function for identifying older drivers who may be at increased risk of crash involvement; it also examines the broader issue of whether impaired cognitive function is associated with increased crash risk in the older driver population. Data for the study were collected from 3238 drivers aged 65 and older applying for renewal of their North Carolina driver's license. The specific cognitive assessments examined include the Trail Making Test parts A and B, the Short Blessed Orientation-Memory-Concentration test of cognitive impairment, a modification of the American Association of Retired Persons 'Reaction Time' test, and a timed Traffic Sign Recognition test. Information on crash involvements during the 3-year period prior to testing was obtained by linkage with the North Carolina driver history file. Although the individual tests were not found to be particularly effective screening tools for identifying subsets of high risk drivers, cognitive test performance remained significantly associated with crash risk even after controlling for driver age, race and measures of driving exposure. Drivers who scored in the lowest 10% on the cognitive tests were approx. 1.5 times more likely to be in crashes than were drivers who scored in the highest 10%. Implications for the counseling and licensing of older drivers are discussed, along with recommendations for future research.