Published estimates of the prevalence of substance abuse in schizophrenia were correlated with several aspects of the studies in which they were obtained. Higher estimates of alcohol and stimulant abuse were found in studies published more recently. Rates of alcohol abuse were particularly high in one rural investigation. Rates of stimulant abuse were highest when patients were asked directly about use of stimulants. Increased prevalence estimates in more recent years could not be attributed to method of assessing substance abuse, method of defining substance abuse, or characteristics of the study sample.