Despite reports of falling first-admission rates for schizophrenia in the UK and other Western countries, it would be rash to conclude that the incidence of schizophrenia is falling. An attempt was made to tackle the many methodological problems and sources of bias influencing the relationship between admission rates and incidence in an analysis of inception rates for schizophrenia and other psychoses in Edinburgh between 1971 and 1989. However it was calculated, the inception rate for schizophrenia fell significantly, but because there was evidence that diagnostic criteria for schizophrenia had narrowed between 1971 and 1989, and because a substantial and changing proportion of recorded first admissions were not true first admissions, it was impossible to conclude that the incidence of schizophrenia had fallen. Changes in the incidence of psychiatric syndromes are difficult to establish, particularly in retrospect, and future studies must pay more attention to the many possible confounding influences.