When investigating a vehicle crash, the issue of seatbelt usage is frequently part of the information needed to perform an occupant kinematics or injury analysis. A physical inspection of the vehicle is the preferred method to investigate seatbelt usage. However, if the vehicle is no longer available, or the condition has changed since the time of the crash, preventing analysis of seatbelt usage by an occupant, the investigators must rely on other available evidence to assess occupant seatbelt usage. This would typically include a review of the police report, scene or early photographs of the vehicle, physical marks on the occupant in medical records and statements from witnesses. More recently, event data recorders (EDR) can provide data regarding seatbelt status for front seat occupants, and occasionally, rear seat occupants. However, the EDR data must have been previously recovered or the vehicle must be available.In cases where the available data is limited or includes only subjective data such as a police report or statements of occupants, some investigators have used the post-crash seatbelt position to determine seatbelt usage at the time of the impact.The theory is if the seatbelt is retracted or stowed post-crash, it was not in use at the time of the collision. The validity of this theory was investigated using EDR data from the NHTSA Crash Investigation Sampling System (CISS) as well as EDR files collected during in-house crash investigations. Photographic documentation of the post-crash seatbelt position was compared to EDR reported seatbelt use to determine if post-crash seatbelt position is reliable in determining seatbelt usage at the time of impact. Additionally, EDR seatbelt usage was compared to police reported seatbelt usage.The analysis of the data in this study found that in a third of the cases where the EDR data indicated the occupant was seatbelted, the seatbelt was found in the stowed or retracted position. Therefore, finding a stowed or retracted seatbelt following a crash is not a reliable means of determining seatbelt use by the occupant at the time of the crash. Additionally, a comparison of EDR data to police reported seatbelt usage revealed that 13 to 25 percent of the occupants reported by the police as seatbelted did not have the seatbelt fastened based on EDR data.