Anxiety Sensitivity and Information-Processing Biases for Threat
Richard J. McNally Harvard University
In his classic article on the two disciplines of psychology, Cronbach ( 1957) emphasized the complementary strengths of the correlational (e.g., psychometric) and experimental traditions in our field. Most chapters in this volume concern studies conducted in the first tradition. This chapter addresses those conducted in the second -- namely, experiments on threat-related cognitive biases in people characterized by high anxiety sensitivity (AS). Most studies on this topic concern panic disorder; patients with this syndrome commonly score high on the Anxiety Sensitivity Index (ASI; McNally & Lorenz, 1987; Reiss, Peterson, Gursky, & McNally, 1986; Taylor, Koch, & McNally, 1992). The chief question is whether people, identified psychometrically as at risk for panic disorder, exhibit the same biases as do those who have developed the disorder.
The theoretical importance of this issue arises because AS is a dispositional variable ( McNally, 1989; Reiss & McNally, 1985) whose predictive significance for panic attacks somehow should be rendered manifest in biases that occasion panic attacks. That is, the ASI taps beliefs (explicit and implicit) about the negative valence of certain sensations and these beliefs should influence online processing of these sensations. To illustrate, moment-to-moment interpretation of bodily perturbations may depend on relatively stable AS beliefs. AS beliefs may thereby undergird attentional, memory, and interoceptive biases