Much health promotion research is directed towards increasing our understanding of the process of health-related behavioural change. However, little is known about the validity of self-reported changes in behaviour. This study aimed to assess the validity of self-reported changes in five behaviours—diet, cigarette smoking, alcohol consumption, weight control and physical activity—by comparing reports obtained from the same people through postal questionnaires and face-to-face interviews. It was found that overall there was a consistency between self-reports with interviewers confirming that in most cases change had taken place. However, in a large number of cases changes were reported that had occurred before the period specified in the questionnaire. The nature of the change, whether discrete or incremental, as well as the behaviour in question, were found to be related to the accuracy of self-reporting. The possible reasons for the observed discrepancies are discussed.