Despite the effectiveness of safety restraints in reducing the risk of death or injury in motor vehicle accidents, many children still travel unrestrained. Two methodological issues related to child restraint studies which had not been adequately addressed were identified: firstly, the accuracy of parental self-reports of restraint use with their children; and secondly, the consistency of observational data over time. These issues were investigated in the present research. The data indicate that parents' report is inaccurate, suggesting that if self-report measures are used, there is a need to compare them with direct observational measures to obtain information on reporting bias. A single observation of a child's restraint use appears to provide an accurate reflection of restraint use on other occasions in the same setting.