A Forced-Choice Adjective Preference Scale for Personality Assessment

Abstract

With the aim of investigating a forced-choice format for studying selected personality check-list variables as a means of overcoming response set variance encountered in other adjective check lists, 17 bipolar scales were defined, and items comprising a sample of the items thought to represent the item universe of each scale were accumulated. From these samples of items, 20 were selected to represent each of the 17 scales, 10 from each extreme. Opposing trait-names were paired randomly to form 17 scales each consisting of 10 forced-choice items. The Adjective Preference Scale was administered to two groups of Ss. Reliabilities and intercorrelations were computed. Intercorrelations were found to be generally low in relation to scale reliabilities, which were considered promising for 10-item scales. Because of the effects of the forced-choice format in enhancing content reliability and in eliminating the massive response set to check many or few items, it is concluded that combining items into scales and casting them into a paired-comparison context is the method of choice in constructing adjective check-lists for personality assessment.