Measurement equivalence of a concise customer engagement metric across country, language, and customer types

Abstract

The establishment of measurement equivalence/invariance (ME/I) across groups is a precondition to conducting any cross-group comparisons; otherwise, one could not know for sure whether the observed between-group difference is due to real differences among the constructs of interest or to differences in how they are measured. However, many multinational corporations (MNCs) actually measure and compare results of customer surveys across groups defined by countries, languages, or between Business-to-Business (B2B) and Business-to-Consumer (B2C) customers without first testing ME/I. In this article, we illustrate how to apply multiple statistical techniques to investigate ME/I, plus we evaluate both construct equivalence and differential item functioning (DIF). We hope that illustrating how such analyses can be done with relative ease may encourage more practitioners to perform ME/I as a routine practice rather than an exception.