PSYCHIATRY’S SECOND VALIDITY CRISIS: THE PROBLEM OF DISPARATE VALIDATION

Abstract

In response to the crisis of confidence in the validity of the DSM’s diagnostic categories, psychiatry has seen a proliferation of alternative research frameworks for studying and classifying psychiatric disorders in new ways. The big three alternative approaches—the Hierarchical Taxonomy of Psychopathology (HiTOP), the Network Approach to Psychopathology, and the Research Domain Criteria (RDoC)—have been characterized as healthy responses to the DSM’s crisis of validity. A yet unexplored aspect of psychiatry’s validity crisis is related to disagreements regarding the standards of validity. Each of the approaches have multiple distinct senses of validity, which point to a thornier methodological problem for psychiatry that I term the problem of “disparate validation.” This two-part problem can be summarized as follows: scientific psychiatry aims at achieving empirically informed classifications that demonstrate validity in the sense that they correspond to real attributes of psychopathology. To achieve this, alternative research frameworks are now approaching the conceptualization, testing, organizing, and validation of different features of psychopathology by their own standards in the hopes of one day informing more valid systems of psychiatric classification. The first problem is given a system of classification, by whose standard of validity should such a system be validated? Is there a single validation procedure by which validation should proceed, or some other combination thereof? Second, when we attempt to validate classifications informed by differing standards of validity, will any such validation be capable of assessing a unified fundamental sense of validity that exists across the various frameworks, or will they only be valid in their own narrow sense? In this dissertation, I offer an assessment of the problem of disparate validation through faithful reconstructions of the Holy Quadrinity of distinct senses of validity in psychiatry: starting with diagnostic validity (DSM) and proceeding with psychometric validity (HiTOP), network psychometric validity (the Network Approach), and etio-pathophysiological validity (RDoC). I introduce commonalities across frameworks that have not been previously addressed, including how each framework employs expert curation, being the selection and justification of certain elements into their model based on compromises, and how the goal of each framework eventually becomes a return to the original validators of Robins and Guze to evaluate prognosis, biomarkers, and etiology of psychiatric classifications. By evaluating psychiatry’s distinct senses of validity, I argue that despite an appearance of a shared goal of informing more valid classifications, the existence of multiple frameworks in which each employs their own standards of validity and validation is a detrimental methodology to achieve any kind of unified validation work. At its core, fundamental disagreements concerning 1) the underlying phenomenon that researchers are attempting to make inferences about; 2) the sources of validating evidence; and 3) the very nature of validity and validation, move each framework further and further toward a state of unrecognized plurality, in which these frameworks are really not at all talking about the same thing and are in fact engaged in different projects with different aims. I conclude with a positive program that suggests in what ways such different frameworks with distinct validation procedures can achieve validity in their own specific sense while also coming to inform one another through a kind of complementary pluralism.