Phobic dimensions: I. Reliability and generalizability across samples, gender and nations

Abstract

A brief overview is given of results of: (a) factor analytic studies with Fear Survey Schedules of relevance to the cross-sample and cross-national constancy of phobic dimensions; (b) reliability analyses with these dimensions. Employing data of somewhat more than 3000 Ss from 10 independent clinical and normal samples, a number of analyses with the FSS-III and the FQ showed the dimensions which have been found to underlie these measures (Arrindell, 1980; Marks and Mathews, 1979) to be replicable in similar samples and invariant across distinct ones, and to be sufficiently reliable for purposes of research, in some instances even for applied problems. In addition, the FSS-III and FQ dimensions were shown to be invariant across sex. Of the main conclusions which were drawn three were particularly outstanding: (1) the Agoraphobic cluster of fear is not peculiar to samples comprising phobics in which agoraphobic fears occur full-blown in a significant number of Ss. Such a factor can be interpreted in data of, for instance, (non)student normals as well provided that suitable techniques of multivariate analysis are employed, thus refuting the claim (cf. Hallam, 1978; Hallam and Hafner, 1978) that agoraphobia is a phenomenon that has “all or none” characteristics; (2) the use of total fear scores would appear to lead to invalid results; therefore, this general practice should be discarded; (3) small-to moderate-size decreases are likely to occur on phobic measures during relatively short and long non-treatment periods; such changes should be taken into consideration when drawing conclusions from results obtained in studies on pre- and post-therapy measurement.A differential use of the FSS-III and the FQ is stressed.