Construct Validity Challenges in Psychiatry Using Human and Animal Models

Kenneth Schaffner, PhD, MD
Distinguished University Professor Emeritus
Professor of Philosophy, Psychology, and Psychiatry

Abstract: The concept of ‘construct validity’ in psychology and psychiatry had its origin in the extraordinarily influential 1955 article by Lee Cronbach and Paul Meehl (also reprinted in the first volume of Minnesota Studies in Philosophy of Science in 1956). Initially formulated in the context of latter-day logical empiricism, the notion has evolved through various versions through to today, where it continues to influence not only its original disciplines, but has also been used by the prominent behavioral geneticists Caspi and Moffitt as an alternative to genome-wide associations studies (GWAS). However, two alternative approaches to validity in these areas have developed almost independently of the Cronbach and Meehl tradition. One of these, most influential in psychiatry, arose from the work of Robins and Guze on “diagnostic validity” in 1970, and was refined and extended by Kendler in the 1980s. This approach in its most recent version was used by the DSM-5 work groups for modifications of disorders presented in that 2013 publication. A third related but distinct approach involving the use of animal models to advance genetic and neuroscience approaches to psychiatric disorders comes from the work of Willner in the 1980s as well as Nestler and Hyman in 2010. More recently, this line of investigation has been analyzed using the Research Domain Criteria (RDoC) framework of the National Institutes of Health. This talk summarized the major features of these three approaches to construct validity and proposed an integrated model of the concept by drawing on these three traditions.

Sponsored by the Center for Philosophy of Science