| Anxiety sensitivity is defined by Reiss and McNally (1985) as a set of negative beliefs about the consequences of anxiety. Anxiety sensitivity is proposed to be a risk factor for the development of anxiety disorders and fears. While anxiety sensitivity is theoretically a cognitive construct, the construct has been criticized for combining both affective and cognitive dimensions of anxiety (Lilienfeld, Jacob, & Turner, 1989). A review of both the psychometric and construct validity evidence for anxiety sensitivity indicates contradictory psychometric evidence regarding the underlying structure of the measure of anxiety sensitivity (the Anxiety Sensitivity Index: ASI). There is little empirical evidence to support the theoretical position that anxiety sensitivity is a cognitive construct. Stronger evidence exists for anxiety sensitivity as an affective construct, however many of the relevant studies have methodological and theoretical flaws. The purpose of the present study was to investigate the structure of the ASI, in terms of psychometric adequacy, and representation of the theoretical construct of anxiety sensitivity. A comparative exploratory factor analysis, item analysis, and exploration of specific a priori dimensions were conducted with a sample of 69 panic disorder patients. Item analysis revealed no items which detracted substantially from the consistency of the ASI. Exploration of the a priori dimensions indicated that the cognitive dimension displayed insufficient homogeneity, whereas the affective dimension was as homogeneous as the ASI as a whole. Factor analyses revealed a solution containing four relatively unrelated factors as the most adequate, accounting for the most variance while containing no poorly-defined factors. Exploration of the item loadings within each of the factors indicated that each distinctly represented either affect or cognition. This interpretation of the factor content was supported through correlation of the factor scores with external measures of affect and cognition. These results are discussed in terms of providing explanation for previous findings. However, these results also indicate little support for the theoretical description of anxiety sensitivity as a unitary cognitive construct. The necessity for clarification of the theory, and the possibility of changes to the ASI are discussed. |
