Test Item Content: Does It Matter?

Abstract

The purpose was to explore a possibly disregarded contribution of the content of items in inventories to responses given to items. The dependent variable was the frequency of extreme responses, positive or negative, given to the items, rather than "neutral," "undecided," responses that are minimally informative. The hypothesis was that items judged as relevant from the viewpoint of a central theme assigned by the subject to the questionnaire would get more extreme responses than items judged as irrelevant. The subjects were 224 males and females (mean age = 22.8). The were administered an attitude scale (Machiavellianism) and a personality scale (IPAT Anxiety). A comparison of the number of extreme responses given to items judged as relevant and as irrelevant fully supported the hypothesis. Judgment of relevance was shown to be the major factor determining extremeness of responses. The findings have important implications for the psychology and methodology of responding to questionnaires.