This paper reports on a series of studies of the validity of children's self-reported exposure to traffic. The studies were conducted within a larger case-control study carried out in the metropolitan area of Perth, Western Australia. For these validity studies, subjects were randomly selected from the original pool of case (n = 100) and control (n = 400) subjects. Three techniques were used to assess the validity of the self-reported 'habitual exposure', namely, the 'moving observer' technique, pedestrian diaries, and a test of construct validity. The child's regular walking activities during the course of a typical week was considered the child's 'habitual exposure'. The 'moving observer' technique involved either the researcher or research assistant following a random sample of children on their walking trips. A further sample of children maintained a 'diary' of their weekly walking trips, and a number of constructs, in this case variables which related, indirectly, to the child's level of exposure to the road environment, were included in the interview schedule. As two researchers were involved in all aspects of this study, intra- and inter-rater reliability were assessed and tape-recordings of the interviews were used to determine the reliability of the coded data. No significant differences were found between the child's reported exposure to the road environment and either the observed exposure or exposure recorded in pedestrian diaries. For some exposure variables, namely, the number and duration of walking trips, children tended to underestimate their exposure (when compared to the observations of the researchers). For case subjects, the number of roads crossed was also underestimated. The test of construct validity indicated that time spent in 'out-of-home activities' (activities other than going to or from school, which involved time spent in the road environment) does not correlate strongly with 'habitual exposure'. Intra- and inter-rater reliability and the reliability of transcribed data from taped interviews were all found to be very sound. The studies indicate that children's self-reported 'habitual exposure' data is a valid measure of his or her actual exposure in the road environment.