Abstract
This study examined vehicle rollovers in terms of site-specific exposure and speeds of vehicles of varying stability. Fifty-one rollover sites in two states were visited at the same time of day and day of week as the rollover. A sample of vehicles moving in the same direction as the rollover were observed, and vehicle-specific data were obtained from identification numbers. Low stability, exacerbated by the addition of passengers, increased the risk of rollover. Speed was not correlated with stability and is not a confounder. Rollovers could be substantially reduced if motor vehicles were manufactured with a static stability of 1.2 or greater. (A)