Abstract
We develop a seven-step empirical procedure for comparing formative and reflective measures: 1. Checking for common method bias, 2. Estimating and evaluating the formative model, 3. Examination of the stability of the formative measurement model, 4. Examining external consistency of the formative model, 5. Estimating and evaluating an assumed reflective measurement model to the data, 6. A model fit comparison between the formative and reflective approaches and, 7. A comparison of the predictive power of both measurement approaches using data-splitting and validation. We validate the procedure on a formative measure of perceived customer equity, offering useful research and managerial implications.