We surveyed assumptions with regard to the relation between age and outcome of head injury (a.k.a. the "Kennard Principle"). Two groups (N = 120) of practicing clinicians were sent background information about a mild to moderate head injury in which only the age of the patient varied and were asked to predict likely outcome (degree of impairment). Predicted outcome differed significantly, with an adolescent case judged much more likely to have serious impairment than a child case. Implications of the (mis)application of the Kennard Principle are discussed.