A New Look at Our Old Attitude Problem

Abstract

Across the disciplines, attitudes are commonly measured through the use of unidimensional scaling (e.g., Likert scaling). Given that the Likert measurement scale (multiple equivalent items regarding an attitudinal target from which a summated score is produced) and the Likert response format (positive to negative judgements) were developed over 80 years ago, this article questions whether Likert scaling is still optimal for attitude investigation. In this article, it is argued that the lack of resolution of the debate around the dimensionality of attitudes has encouraged a simplistic view of attitudes (assumed unidimensionality of both attitudes and attitudes responses) to predominate, without empirical justification for this view.