Mental Disorder Is Disability: In Support of Our Design

Although generally supportive of our overall position, both Zachar and Gala and Laughon raise questions about our research design. Herein, we respond to these questions by presenting counterarguments that support the soundness of this design.

Reply to Zachar

Subsequent to stating some broad agreement with our central thesis that mental disorder is best viewed as a disability concept and not a behavioral one, Dr. Zachar expresses a number of reservations about our work. We focus on the following discussion on what seem to be the most significant of these.

Dr. Zachar states the following (2017, p. 46): “In 8 of their 12 conditions, the dependent variable is evaluation of problematic behavior and the independent variable is disability, operationalized as loss of control. To test the hypothesis that disability is the distinguishing factor in differentiating mental disorder from ‘non-disordered’ behavior, the only difference between the conditions should be the presence or absence of disability. Unfortunately, all of these vignettes are confounded by multiple differences.” Dr. Zachar then goes on to recommend a methodological alternative in which we use a series of matched cases in which everything is held constant between the two except the presence or absence of disability.

We agree, with some reservations, that this is a sound methodological idea. Indeed, it is one we ourselves mention in our Discussion as an idea for future research. However, we disagree that this invalidates the logical soundness of the present design. In our work, we presented judges with a panoply of cases that contained characteristics associated with a spectrum of mental disorders (e.g., antisocial, eating, sexual, and anxiety disorders). We asked them to read each case and render a judgment regarding the likelihood the person described had a ‘mental disorder.’ With great regularity, and perhaps without even knowing what feature they were using to discriminate (our pretest judges did not), they judged that those who were described as disabled were likely mentally disordered, whereas those who were not disabled were likely not disordered. By way of analogy, consider a hypothetical experiment in which we ask judges to pick out all the instances of ‘bachelor’ in a set of descriptions. We then present them with 12 descriptions of persons varying in many ways such as gender, age, race, socioeconomic status, religion, marital status, and attractiveness. We find that judges overwhelmingly use a combination of only three criteria, male, unmarried, and adult, in making their decision as to who is a bachelor and who is not. Logically, in this hypothetical and in our actual study, the fact that the described persons have all of these other quite disparate characteristics does not undermine the logical soundness of the design.

We have two reservations about Dr. Zachar’s suggested alternative design. First, showing participants cases that matched in every respect except presence versus absence of disability could well have made our hypothesis too transparent to participants. Second, our design, with its just discussed variety of other characteristics (gender, age, etc.) renders it highly unlikely that any other third or confounding variable varies systematically in exactly the same way as the variable we intended to systematically manipulate, that is, presence versus absence of disability.

Dr. Zachar’s second reservation about the present work, and the one to which he devotes the most attention, is encapsulated in the following quotes:

My second reservation pertains to the semantic theory that the authors adopted as the a priori framework of their research program, namely the classical view of concepts and categories. According to this view, concepts have correct meanings defined by necessary and sufficient features...an important alternative to the classical view was offered by Wittgenstein. Using the concept of a game, he observed that there is no feature or fixed set of features that every game shares. Instead, games resemble each other in a variety of ways analogous to members of a family.... At the very least, the scientific research of Rosch (1976)... supports a more Wittgensteinian view in contrast to...