The debate concerning the relation between anxiety sensitivity (AS) and trait anxiety has been constructive for the field and has suggested a number of important directions for future research. Reiss’ (1997) commentary on AS and trait anxiety in this journal, however, contains several serious factual misstatements and logical errors that confuse, rather than clarify, many of the central issues in this debate. These misunderstandings are corrected and the implications of the issues raised by Reiss are addressed here. The authors suggest that future research on AS (a) embed this construct within the context of broader temperamental and personality variables and (b) explicitly recognize the bidirectionality of emotional and cognitive influences.
The debate concerning the relation between anxiety sensitivity (AS) and trait anxiety has been constructive for the field and has suggested a number of important directions for future research. Reiss’ (1997) commentary on AS and trait anxiety in this journal, however, contains several serious factual misstatements and logical errors that confuse, rather than clarify, many of the central issues in this debate. These misunderstandings are corrected and the implications of the issues raised by Reiss are addressed here. The authors suggest that future research on AS (a) embed this construct within the context of broader temperamental and personality variables and (b) explicitly recognize the bidirectionality of emotional and cognitive influences.
As a number of philosophers of science (e.g., Bartley 1984 and Popper 1965) have noted, science at its best involves the maximization of criticism. In other words, scientific progress occurs when new concepts are subjected to searching and often intense scrutiny, and either emerge intact or require abandonment or modification. Thus, debate and controversy, so long as they are ad rem rather than ad hominem, are healthy for science and should be actively encouraged (see Kendrick & Funder, 1988).
In this context, the debate regarding the relation between trait anxiety and anxiety sensitivity (AS), a cognitive construct reflecting individual differences in beliefs concerning the harmful consequences of anxiety, has been constructive for the field (e.g., Lilienfeld et al. 1993, McNally 1996 and Taylor 1995). In this debate, we raised the question of whether AS and trait anxiety are entirely distinct or independent constructs, and contended that the incremental validity of AS above and beyond trait anxiety required closer examination Lilienfeld et al. 1989 and Lilienfeld et al. 1993. In addition, we argued that a number of early findings relating AS to anxiety disorders were potentially attributable to the effects of trait anxiety and similar unmeasured variables (see Lilienfeld et al., 1989).
This debate now appears to have run its course. There is a consensus that AS, as operationalized by the Anxiety Sensitivity Index (ASI; Reiss, Peterson, Gursky, & McNally, 1986), correlates moderately with trait anxiety, although it contributes to the prediction of certain phenomena (e.g., panic attacks; anxiety responses to challenge procedures; McNally 1989 and McNally 1996; cf. Brown & Cash, 1990) above and beyond trait anxiety. Like many controversies, this debate has clarified several important issues and suggested novel avenues for future research. For example, this debate has pushed researchers to examine the extent to which AS possesses incremental validity above and beyond trait anxiety in the prediction of anxiety disorders, including panic disorder (see McNally, 1996, for a review), and has spawned the development of new models of the relation between AS and trait anxiety. For example, Lilienfeld et al. (1993) (see also Taylor 1995 and Zinbarg, Barlow, and Brown 1997) posited a hierarchical model in which AS is conceptualized as a lower-order facet of a higher-order trait anxiety construct. In addition, researchers are beginning to examine the extent to which AS is associated with personality dimensions related to trait anxiety, such as self-consciousness and other facets of neuroticism, as well to dimensions that are largely independent of trait anxiety, such as introversion and absorption (Borger, Cox, Feuntes, & Ross, 1996; Lilienfeld, 1997).
I
