Although vehicle rollovers represent a low percentage of all vehicle crashes, rollovers are disproportionably represented in terms of vehicle occupant injuries and fatalities. The National Highway Transport Safety Administration (NHTSA) has been consistently exploring the issues relating to rollover. NHTSA have implemented a star rating system to inform consumers of the rollover propensity of vehicles as part of the New Car Assessment Program. The NHTSA propensity rating is based on the Stability Factor. Recent work by Monash University Accident Research Centre (MUARC) for the Victorian Police Force resulted in a range of a vehicle functional performance criteria being developed and utilised. A specific criterion relating to vehicle handling and Stability Factor was proposed. There is sufficient evidence to support the contention that Stability Factor and the rate of “real world rollovers” are linked. It is the contention of the authors that the ‘apparent noise’ (scatter) within the Stability Factor data is due to vehicle handling. This paper proposes a methodology that allows the combination of the Stability Factor and handling characteristic to be measured for a specific vehicle and enables the probability of rollover per single vehicle crash to be estimated.