Using data from i) a self-administered survey of 673 older female drivers, and ii) a case-control study of 48 crash-involved and 44 non crash-involved older female drivers, some factors that may predict crash involvement were identified. Survey data gathered selfreported information on demographic characteristics, health status, travel patterns and driving practices. A battery of functional performance assessments were administered by trained experimenters in the case-control study, and participants completed a second detailed self-administered questionnaire. Factors found to predict crash involvement included driving characteristics such as being the main driver in the household, not highly confident of being a safe driver, experiencing difficulty driving in unfamiliar areas and having problems with the driving style of other drivers. In addition, low attentional, cognitive and motor skills and presence of multiple medical conditions were significant predictors of crash involvement.