North Carolina crash data for 1972, with a follow-up for 1981, were examined to identify driver violations as reported by the investigating officer. This information was linked to the driver history file to determine whether there was a conviction, and, if so, for how serious a violation. Categorical analyses of the data showed systematic biases in who is convicted once a violation occurs. Nonwhites, drivers aged 55 and over, and drivers of older vehicles had a higher likelihood of being convicted of a same or higher level violation. Because all these variables may be considered rough handles on socioeconomic status, it appears that those persons least able to defend themselves are most likely to experience convictions and hence fines, increased insurance costs, and increased probability of loss of license. Whether for research or administrative purposes, to the extent that driver records are used as an indicator of driver performance, it is important to be aware of such biases in the system.