SummaryDespite a large research effort, there is still some doubt whether life events can increase the rate of psychiatric morbidity. It is argued that this doubt arises from a confusion between measures of goodness of fit and measures of causal importance. The distinction between these types of measures is discussed.Epidemiological studies of life events and psychiatric cases are reviewed. It is argued that, in general, 32 per cent of psychiatric cases can be attributed to stressful life events. Within female samples, approximately 41 per cent of psychiatric cases can be attributed to life events. It is suggested that the effect of specific types of events on specific psychiatric disorders may be of even greater importance.