Police officers' injury ratings and inebriation judgements are key variables in studies based on traffic crash data. We undertook a retrospective analysis of 1281 cases which had both North Carolina Division of Motor Vehicles crash reports and medical evaluations at a level-one Trauma Center to compare police officers' assessments of injury severity and inebriation to those of Emergency Department physicians'. In approximately 49% of the crash victims, there was complete agreement between the investigating officer and physician assessment of injury severity. Assessment agreed best in the most severe and least severe injury categories. Police officers' judgment of the victims' alcohol impairment demonstrated 80% agreement with objective blood alcohol measurements or with Emergency Department physicians' judgment in the inebriated victims and in 96% of the unimpaired patients. There was no systematic bias towards officer overrating of injury in the impaired patients. These findings support the use of police officer assessments of injury severity and ethanol intoxication in studies of motor vehicle crashes.