A review of statistical methods for road traffic accident mass databases

This review considers statistical techniques that can be applied to data from mass road traffic databases, concentrating on approaches which do not require ancillary information to the database. Two general problem types are considered: estimation of the degree to which factors affect crash propensity and the evaluation of the effectiveness of factors in reducing injury outcome once the crash has occurred. In both problem types, a number of techniques are identified which promise to provide significant advantages over methods currently in use.