A sample of 991 five‐year‐old children was divided into three groups on the basis of accidental injuries they had received in their first five years of life. Group one (N = 505) had had no accidents, group two (N = 326) had had one accident and group three (N = 160) had had two or more accidents. The groups were then compared on ninety background, developmental behavioural, and medical measures. No consistent differences between the three groups were identified. It is argued that attempts to identify those more liable to accidents lack utility and have an inhibitory effect on accident prevention In view of this, it is advocated that future accident research should concentrate on identifying unsafe aspects of the environment.