Cross-sample invariance of the structure of self-reported distress and difficulty in assertiveness: Experiences with the scale for interpersonal behaviour

Recent surveys of psychometric studies with respect to the construct and/or factorial validity of self-report assertion measures (e.g. Beck and Heimberg, 1983) point to the paucity of systematic research addressing such issues. To date no attempts have been made to examine the cross-sample generalisability of dimensions of assertion by means of objective techniques of factorial invariance (i.e. hypothesis-testing variations of principal factor or principal components analysis). Employing such a technique in the present study, distress and performance factors identified previously in a phobic population (N = 703) with a Dutch measure of assertion (Scale for Interpersonal Behaviour) — the factors being: (I) DISPLAY OF NEGATIVE FEELINGS, (II) EXPRESSION of and DEALING with PERSONAL LIMITATIONS, (III) INITIATING ASSERTIVENESS and (IV) POSITIVE ASSERTION — were shown to be replicable in a new phobic sample (N = 119) and invariant across three distinct samples, namely psychiatric outpatients (N = 253), (non-patient) students (N = 412) and (non-student) normals (N = 250). Additional reliability and validity data supported the utility of a fifth overall measure within the Scale, which was termed GENERAL ASSERTIVENESS. These data also underline the contention of Futch et al. (1982) that it would be unwise to take any demonstrated multidimensionality of an assertion construct as evidence in the context of the trait-state issue (e.g. Eisler et al., 1975). However, the exclusive use of an overall assertion scale of a purportedly multidimensional measure is likely to invalidate research findings (including clinical research) since it masks relevant component information. To counteract this problem it is proposed that either the subscales should be used on their own or that the subscales and the overall scale should be used together. It is further concluded that (1) in developing and using an assertion measure a multidimensional approach should be taken, there being at least four distress and four performance factors to be taken account of; (2) of these four factors, that of Positive Assertion in particular should be considered more thoroughly in adequately defining and operationalising the assertion construct.