A Cross-Sectional Optimal Scaling Analysis of the Index of Consumer Sentiment

Abstract

This article reports empirical investigations of the measurement assumptions implicit in the Index of Consumer Sentiment (ICS). The ICS has been used in much of the past research that attempts to incorporate concepts of consumer psychology within frameworks of economic theory. The research employs a new numerical analysis model, known as alternating least-squares optimal scaling, to generate insights into this historically important indicator of consumer economic attitudes. The results of the present analysis, based on two national consumer surveys, raise new concerns about the procedures commonly used for the measurement and reporting of consumer sentiment. The procedures used provide insight concerning construction of composite variables from a set of categorical (or mixed-metric) questionnaire items, as well as a way of analyzing “don't know” responses to questionnaire items.