Research has shown that the processing of risk information is negatively related to engagement in patterns of risky behaviour. However, no evidence yet links risk taking with lowered risk perceptions. The aim of this study was to examine this relationship using estimates of personal risk inherent in exceeding limits in common speed zones, rather than traditional risk perception items that can be confounded by engagement in risk taking. As predicted, a prospective inverse relationship between prior engagement in risk taking, and the perception of inherent risk was identified. After controlling T1 risk perceptions, 2–4% of variance in T2 risk perception variables was explained by T1 speeding and T1 to T2 speeding changes. The unique negative correlation between speeding at T1 and risk perceptions at T2 provides the first longitudinal evidence linking engagement in patterns of risk-taking behaviour to subsequently reduced risk perceptions. This supports the view that involvement in risk-taking behaviour is related to lower subsequent risk perceptions, and is consistent with studies that show inhibited processing of risk information by risk takers. T1 to T2 speeding changes were also related to lower risk perceptions. This suggests that positive changes in risk perceptions might arise from decreases in risktaking behaviour. However, the latter finding is not definitive, as the study design does not allow us to establish that the speed changes preceded the development of the T2 risk perceptions. Thus, the causal sequence may easily be the reverse. The theoretical implications of these findings could be considerable. A negative correlation between estimates of inherent risk and exposure means that there will be, at best, a positive curvilinear relationship between risk taking and traditional risk perception scores. Assuming that the benefits of risk taking are linked to risk taking in a linear way, the differential between perceived benefits and perceived risk will progressively increase in favour of benefits as risk taking increases (Job, 1990). This provides a potential explanation for the common finding that, whilst they perceive greater personal risk than non-risk-takers, risk-takers tend to understate risk compared to objective indicators (Weinstein, 1998; Wild et al., 2001). The current study does not provide a critical test of this model, which remains a topic for future research. This study is descriptive, and does not lead to an understanding of why risk taking is related to lower risk perceptions. There are several possible explanations. At a simple level, unpunished performance of risk-taking behaviour might confer a lesser perception of danger through increases in confidence and perceptions of personal control (Job, 1990), particularly as many risk behaviours either carry a low or delayed chance of negative outcomes. An alternative view, self-perception theory, suggests that people may partly infer their beliefs about risk from an examination of their own behaviour patterns (Albarracin & Wyer, 2000). Other theoretical perspectives suggest that risk-takers might deliberately employ strategies designed to bias risk cognitions. Theories of maladaptive coping (Rippetoe & Rogers, 1987) suggest that perceptions of vulnerability are inherently aversive. When behavioural change is either difficult or undesirable, people may attempt to assuage perceptions of vulnerability by reducing their estimates of the risk that they face (Freeman et al., 2001; Klein, 1996; Liberman & Chaiken, 1992).