Communicating risk is a key public health strategy. The implicit assumptions are that the public interprets risk information in a logical fashion and adopts behavioural changes to reduce risk. We assessed risk behaviour, and knowledge and perception of voluntary and involuntary risks using an anonymous questionnaire completed by 472 students. Risk-taking behaviour was measured as the number of different risk behaviours undertaken in the previous 12 months. Knowledge and perception were measured by the extent to which subjects agreed with statements of risk-related information. These varied in complexity from simple statements linking a behaviour with a health risk to numerical statements describing the strength of such relationships. Risk-taking behaviour was highest amongst younger people, males, people whose parents were in non-manual occupations, and people who believed in God (risk-taking behaviour was not related to voting preference or birth order). Overall, knowledge was not significantly related to risk-taking behaviour ( P =0.889). However, risk-taking was positively related to more accurate responses to numerical risk questions ( P <0.001) and risk-takers were also more likely to perceive both voluntary and involuntary risks as less risky ( P <0.05). At least in this cohort, more information about risk is not related to lower risk behaviour. In fact, those individuals with a better understanding of the precise risk associated with certain behaviours were more likely to be higher risk-takers while those who consistently over-estimated risks were low risk-takers. Overall, knowledge and perception of risk explained relatively little of the variance in behaviour. Although these findings need further examination within the general population, public health measures should not assume that information campaigns will necessarily lead to a reduction in risk behaviour. Public Health (2001) 115, 54–61.