Anger, hostility and aggression : the psychometric capabilities of a revised Buss-Durkee inventory

Abstract

It would appear that the emotion 'anger* together with hostile attitudes and aggressive behaviors is receiving more attention in the clinical literature. Within clinics, themselves, more programs are being developed to deal with anger and its sequelae. This has lead to the need for psychometric devices to assess anger, etc., more meaningfully. The Buss-Durkee Hostility Inventory was originally developed to assess certain aspects of anger but proved to, be psychometrically questionable. This device does, however, present an item pool that makes intuitive sense. The current project, then, did^y in essence: a) rewrite the items such that the behaviors may be ranked on a Likert-type scale (Research work has demonstrated the ambiguity and lack of power of true/false formats), and b) run a preliminary work-up on the revised test using participants from introductory psychology classes. This procedure was conducted in two stages, the first being an item analysis. This was done through a factor analysis; principal component with iteration followed by an oblique rotation. On this basis, factorially pure scales were generated. An analysis of the internal consistency of each new scale was then modified further to assure high homogeneity. Some preliminary validation was then assayed by correlating the scales with a related instrument developed by Novaco. It was expected that 'hostility-related* scales would correlate with this device. The second stage consisted of the revised scales being administered to a new group for cross-validation (that is, a corroborative factor analysis with oblique rotation was conducted). The internal consistency of the scales was analyzed further and an anova was run using the revised scales against gender and a measure developed by Crowne and Marlowe(1960) designed to detect the influence of item endorsement on the basis of social desirability.