Title

A taxometric investigation of the latent structure of anxiety sensitivity

Date of Award

2005

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Psychology

First Advisor

Mitchell E. Berman

Advisor Department

Psychology

Abstract

Anxiety sensitivity is the fear of anxiety and anxiety-related sensations arising from the belief that these sensations can have or signal harmful somatic, social, or psychological consequences. A common assumption among anxiety sensitivity researchers is that the latent structure of anxiety sensitivity is dimensional rather than categorical. However, it is possible that a clear distinction exists between normal and pathological levels of anxiety sensitivity. The present study investigated the latent structure of anxiety sensitivity and each of its subcomponents using the coherent cut kinetics method. Participants consisted of 744 undergraduates who completed three self-report anxiety sensitivity measures---Anxiety Sensitivity Index-Revised, Body Sensations Questionnaire, and Agoraphobic Cognitions Questionnaire---and a structured clinical interview. Three mathematically independent taxometric procedures---MAXEIG, MAMBAC, and L-mode factor analysis---were used to analyze the latent structure of anxiety sensitivity and its subcomponents. A case-removal consistency test and multiple plot raters were implemented as additional consistency tests. Results from these analyses provided consistent evidence for the dimensionality of anxiety sensitivity and each of its subcomponents. The vast majority of plots were rated as dimensional by both examiners, and base-rate estimates were generally discrepant both within and across groups. The case-removal consistency procedure provided additional evidence for an anxiety sensitivity dimension. These findings indicate that: the general assumption that anxiety sensitivity represents a latent dimension is correct, investigators should not dichotomize anxiety sensitivity in their research designs, appropriate statistical procedures such as regression analysis should be used to analyze the full spectrum of anxiety sensitivity, measures of anxiety sensitivity should not attempt to group individuals into pathological versus non-pathological categories, and anxiety sensitivity is characterized by an additive etiology.