Formative measurements in operations management research: Using partial least squares

Abstract

The partial least squares (PLS) approach to structural equation modeling (SEM) appears across a wide array of business research publications, including those in operations management (OM). However, the authors’ summary of PLS use in the OM literature suggests some concerns and issues. First, the debate on the use of PLS-SEM is intensifying instead of being mediated despite the increasing use of PLS-SEM. Second, a lack of clarity exists among OM researchers about the use of reflective and formative measurements for constructs. Third, the validation of formative measurement is not routinely conducted in studies, which supports the need to summarize and illustrate the validation procedure of formative measurement. Without addressing these questions, the rigor involved in selecting reflective versus formative measures, especially in the OM field, is compromised. This research summarizes the procedures for choosing and validating formative measurement. The authors provide an illustrative OM example to demonstrate how the specific steps are applied. Through proactive selection and judicious operationalization of the measurement model and appropriate comparisons of the overall research model effectiveness based on criteria such as the R2 of the dependent variable OM, researchers provide a tool to help them extend existing theoretical frameworks and explore new theories.