The relationship of selected health beliefs to self-reported crash experiences, cycling at night and the use of tail-lights at night was examined for a sample of 730 13 years olds. The questions relating to cycling health beliefs were similar to those asked by Weinstein in his study of unrealistic optimism about susceptibility to a range of health problems. They included: vulnerability, worry, controllability, chance, barriers, seriousness, apprehension and safety. The majority of 13 year olds interviewed did not exhibit a strong optimistic bias regarding their vulnerability and skills as safe cyclists. In general the majority considered a crash involving a car would most likely be serious and only a minority said they were not worried about being involved in such a crash. In addition, the majority did not consider that crashes are a matter of chance or that there are significant barriers to adopting safety measures. They tended to consider the chances of being apprehended by a traffic officer for failure to use a tail-light at night as being low. Overall the univariate analyses failed to show a consistent pattern of relationships between health beliefs and the two cycling behaviours. The same was true for crash experiences. The joint effects of beliefs and crash experience on the use of a light at night were examined using logistic regression. The results suggest that the health beliefs work in an additive way as far as boys are concerned, but that they have very little influence on the behaviour of girls. The model for boys suggests that the variables have a relatively small influence on behaviour.