This article discusses the way how change in depressed patients included in clinical trials was both conceptualized and measured in the 1970s to decide on the efficacy of the first candidate drugs for the treatment of depression. Understanding how this issue was resolved is of major interest as the protocol designed to distinguish the diagnosis of the depressive syndrome from the measurement of its evolution over time built the contours of the methodological device to which the whole of standardized evaluation in psychiatry has since referred. This paper draws on the debates in nascent psychopharmacology in both the United States and Europe and examines the protocols and findings of the research programs on depression the National Institute of Mental Health (NIMH) has developed in the United States at the same time, as well as on the writings of the main authors of the standardized evaluation. This investigation suggests that measurement of patients’ improvement could not be characterized on the basis of psychiatric inventories that were not well-suited to quantification, nor on the basis of psychometric tests which were intended for the evaluation of psychological invariants. This is why the answer given to the characterization of change in depression under therapy rather reveals a syncretism both conceptual and technical, mixing classificatory and psychometric traditions, a syncretism that the distinction endorsed by the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition,between syndromes and personality disorders no longer allows us to perceive.