The purpose of this study was to use latent variable analysis to model the relationships between the personality characteristics (personal and social adaptations), traffic safety attitudes, risky driving behaviour of drivers and involvement in traffic accidents. A structural equation modelling analysis was performed in order to investigate whether the hypothesised effect of personality upon traffic accidents was mediated through risky driving behaviour and safety attitudes. The results showed that there is a considerable effect of personal and social adaptation scales on traffic safety attitudes and risky driving behaviour. Latent variable analysis indicated that the latent variables of personal adaptation and social adaptation predicted two other latent variables (risky driving behaviour and traffic safety attitudes) with relatively high path coefficients. In addition to this, the analyses revealed that the attitudes to traffic safety in particular had an impact on road accidents.