Abstract

Psychiatric research is in crisis because of repeated failures to discover new drugs for mental disorders. Lack of measurement validity could partly account for these failures. If researchers do not actually measure the effects of drugs on the disorders they aim to investigate, one should expect suboptimal treatment outcomes. I argue that this is the case, focusing on depression, and fear & anxiety disorders. In doing so, I show how psychiatric research illustrates a more general phenomenon that I call “validity drift.” A validity drift occurs when, in the course of developing new measurement procedures, researchers end up studying something different from what they originally aimed to study. I analyze the different ways in which scientists attempt to validate procedures for measuring mental disorders for the purpose of testing drugs in animal models. I show how those validation efforts might fail, thereby leading to validity drifts. Overall, this analysis highlights the complex interplay between the development of new measurement procedures, their calibration, and scientific theories of the target phenomena.