Handbook of Cognition and Emotion

Richard J. McNally
Department of Psychology, Harvard University, Cambridge, MA, USA

Psychopathologists have used two primary approaches to study aberrant cognition in people with anxiety disorders. One approach requires patients to disclose their beliefs about fear-evoking objects on structured self-report instruments (e.g. McNally & Steketee, 1985; Thorpe & Salkovskis, 1995) and on questionnaires, such as the Anxiety Sensitivity Index (ASI; Reiss et al., 1986). Studies in this tradition rest on the assumption that people are capable of describing what they dread and why they dread it. Patients' beliefs, in turn, presumably figure in the maintenance and perhaps the etiology of their disorder.

The second approach to studying aberrant cognition entails the application of Experimental psychology methods (e.g. Williams et al., 1997; McNally, 1996). These laboratory-based studies emphasize behavioral measures (e.g. reaction time) as a means of making inferences about information-processing derangements that presumably generate the signs and symptoms of anxiety disorders. Unlike self-report methods that tap propositional abnormalities (e.g. “If my heart beats too fast, I may have a heart attack”), they instead capture on-line processing abnormalities in attention, memory and so forth.

These approaches have deep roots in our field's two great traditions of psychometric and Experimental psychology (Cronbach, 1957). But some psychopathologists doubt whether the first tradition is capable of providing a credible basis for a cognitive science of anxiety disorders. In a characteristically brilliant essay, MacLeod (1993) argued that self-report approaches to cognition signify a return to the discredited introspectionism that derailed psychology in the early twentieth century. Accordingly, he holds, contemporary attempts to elucidate cognitive abnormalities via self-report are doomed to repeat the mistakes of the past. MacLeod urged psychopathologists to remain rigorously Experimental and to use behavioral data as the chief basis for inferring cognitive dysfunction.

-479-