The relationship between self-reported depression and a clinical diagnosis of depression was investigated. Within 2 weeks of completing the Center for Epidemiologic Studies Depression Scale (CES-D), a stratified sample of 425 primary medical care patients received the structured interview for the DSM-III-R. In the weighted data set, the CES-D was significantly related to a diagnosis of depression but also to other Axis I disorders. Most distressed subjects were not depressed, a fifth of the patients with major depressive disorder (MDD) had low distress, and the CES-D performed as well in detecting anxiety as in detecting depression. MDD, other depression diagnoses, and anxiety and substance use disorders were all significant predictors of CES-D score. Differences in demographic variables, treatment history, and impairment highlight the nonequivalence of the self-report scale and diagnosable depression. The use of a self-report in place of an interview-based diagnostic measure in the study of depression, as well as the use of such a report as a screening device, is discussed.