Confirmatory factor analysis of the Fear Questionnaire in panic disorder with agoraphobia.

The Fear Questionnaire responses of 390 patients with panic disorder with agoraphobia were used in a confirmatory factor analysis. The results provide strong support for the 3-factor model of this scale (agoraphobia, social phobia, blood/injury phobia) and the multidimensional model of fears proposed by Arrindell (1980). The presence of fear clusters other than agoraphobia existing in panic disorder is also discussed. The Fear Questionnaire (FQ; Marks & Mathews, 1979) is a short (15-item), standardized self-report measure of three phobic dimensions that has become popular for both clinical and research purposes. A recent meta-analysis found that this scale was the most frequently used outcome measure of agoraphobia in the treatment literature (Ogles, Lambert, Weight, & Payne, 1990). Accordingly, the FQ has received increased attention from several psychometric investigators who used mostly European and Australian samples of mixed anxiety diagnoses. One of the FQ's greatest assets is its demonstrated ability to differentiate patients with agoraphobia from patients with social phobia (Cox, Swinson, & Shaw, 1991; Oei, Moylan, & Evans, 1991; van Zuuren, 1988). These studies have found the FQ to have good reliability and validity. The FQ contains three subscales that purportedly assess separate phobic domains: agoraphobia, social phobia, and blood/ injury phobia. A related scale, the Fear Survey Schedule (FSSIII; Wolpe & Lang, 1964), contains similar items but is considerably longer (76 items) and includes more simple phobias. Arrindell, Emmelkamp, and van der Ende (1984) have suggested that the FSS-III may be more suitable for assessing fears in the general population, whereas the FQ's greatest utility is with agoraphobic and social phobic samples. Although the FQ presents a clearly testable model, there have been few factor-analytic studies of the three-factor model. Most factor-analytic studies of phobic domains have included only some FQ items in a larger pool of phobia items (usually from the FSS-III), relied on mixed samples of phobic patients rather than a specific anxiety disorder, or both. Using an Australian sample of mixed anxiety-disorder out