| dc.description.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. | |
