It is well established that alcohol (ethanol) is associated with increased probability of traumatic injury. This relationship has been attributed to alcohol's impairment of judgment and psychomotor performance, leading to increased probability of an injury-producing mishap. Once an accident occurs, it is widely believed that alcohol may protect against injury. However, controlled laboratory studies using animal models indicate that alcohol exacerbates the injurious effects of trauma. In this study, detailed analysis of data from more than 1 million drivers involved in motor vehicle crashes indicates that when the effects of injury-related variables such as safety belt use, vehicle deformation, vehicle speed, driver age, and vehicle weight are taken into account, the drinking driver is more likely to suffer serious injury or death compared with the nondrinking driver. These findings do not support the widespread belief that alcohol is protective against injury, but rather indicate that alcohol increases vulnerability to injury in any given crash.