Abstract
Previous research has indicated that in the first few years drivers are relatively poor at detecting hazards. The present report explores developments in research on hazard perception. In two experiments hazard perception training is developed and evaluated revealing that significant improvements can be achieved. A third experiment examines the prevalent assumption that driving is automatic. It is shown that hazard perception is not automatic and does suffer interference from other tasks. The fourth experiment demonstrates how the general methodology can be used to examine other important dimensions of driver behaviour such as speed choice. (A)