Journal of the Association for Information Systems


This paper examines the concept of the nomological validity of second- and/or higher-order measurement models. It also proposes a new approach that consists of measuring two validity indices — predictive and mediating efficiencies — to compare the efficacies of a research model with and without a higher-order abstraction. To illustrate this concept, we test a second-order measurement model of trust and study how it behaves in a nomological network of the consumer’s prior experience as an antecedent and willingness to buy as a consequent variable.