Abstract
An analysis was made of accident data from a three-county area in North Carolina to identify a set of unsafe driving actions (UDA's) and determine their relative frequencies in accidents. Through field observations at accident locations, frequencies of occurrence were estimated and subsequently used to calculate relative risk factors for a selected group of UDA's (speeding, following too closely, driving left of centre, running a traffic control, turning in front of oncoming traffic, and pulling in front of oncoming traffic). The frequencies of traffic citations were estimated for comparison with accident frequency and relative risks to determine any differences in accident causation and police enforcement priorities.