The purpose of this paper is to investigate occupant injuries which may be sustained during a single-event crash with known roll mechanism.The data was obtained from the weighted National Automotive Sampling System/ Crashworthiness Data System (NASS-CDS) for calendar years 1992 to 1996. The effect of number of rollover turns, roll direction, ejection and belt usage on driver injury responses was analyzed in single-event trip-overs. Trip-overs were chosen for the analysis because they account for over 50% of rollover crashes. The number of rollovers was divided in 3 categories: ¼ to ½ turn, ¾ to 1 turn and above 1 turn. Roll direction was either roll-left or a rollright along the longitudinal axis of the vehicle. Roll-left represents a roll with the driver side leading, while a roll right is with the right front passenger side leading.In the database used in this study, there were three times more belted drivers than unbelted. Ejection was more frequent when the driver was unbelted than belted. Of the 50,504 unbelted drivers, 27% were completely/ partially ejected, while, of the 150,426 belted drivers, less than 1% were ejected. The rate to be seriously injured was higher when the driver was completely or partially ejected. Drivers in a roll-left trip-over were most commonly ejected through the left front area (for example, left door) of the vehicle, while drivers in a rollright were ejected from the left front area and the roof. Of the non-ejected drivers, the rate to be seriously injured seemed higher for unbelted drivers than belted, at 2.2% and 1.1% respectively. For belted drivers in a roll-left, injuries were most frequent in the head, lower extremity (LX), thorax and upper extremities (UX), while, in a rollright, injuries were most often to the spine, head, and thorax. Spinal injuries resulted when the vehicles rolled right, independent of the number of rolls. Unbelted drivers who rolled left, sustained serious injuries to the head, LX, thorax and UX, while those who rolled-right had serious injuries to the spine, head and thorax.Field data can be useful in the development of safetycountermeasures for rollovers as it provides insights on the significance of various parameters. The results of this study suggest the need to first prevent ejection by keeping the occupant in the vehicle. This could be accomplished by changing driver behaviour through increased seat belt use and through technology by helping to obstruct ejection paths. Occupant/vehicle contacts should also be reduced to minimize the potential for injuries. In addition, a better understanding of the head, thorax, spine and lower extremity injury mechanisms is essential for the development of future safety-countermeasures.