Measurement scales for scoring or ranking sets of interrelated items

Abstract

Surveys concerned with human values, economic utilities, organisational features, customer or citizen satisfaction, or with preferences or choices among a set of items may aim at estimating either a ranking or a scoring of the choice set. In this paper we discuss the statistical and practical properties of five different techniques for data collection of a set of interrelated items; namely the ranking of items, the technique of picking the best/worst item, the partitioning of a fixed total among the items, the rating of each item and the paired comparison of all distinct pairs of items. Then, we discuss the feasibility of the use of each technique if a computer-assisted data-collection mode (e.g. CATI (telephone), CAPI (face-to-face), CAWI (web) or CASI (self-administered)) is adopted. The paper concludes with suggestions for the use of either technique in real survey contexts.