Marker Variables in SmartPLS: A Practical Guide to Common Method Variance

Abstract

Common method variance (CMV) remains a persistent concern in survey-based structural equation modeling (SEM), yet practical guidance for implementing marker-variable procedures in partial least squares structural equation modeling (PLS-SEM) remains fragmented. Existing studies differ in how marker variables are selected, specified, and interpreted, creating inconsistent applications in SmartPLS, particularly for mediation, moderated mediation, and higher-order construct models. This paper consolidates existing methodological recommendations into a practical workflow covering marker selection, model specification, implementation, interpretation, and reporting. The proposed guidance clarifies the distinction between dependent-variable-only and broader measurement-level specifications, provides practical criteria for selecting and specifying marker variables, and demonstrates their implementation across direct-effect, mediation, moderated mediation, and hierarchical component models estimated using the repeated-indicators approach. Across the worked models, substantive relationships remained stable after marker inclusion, while marker paths were consistently unsupported. Particular attention is given to hierarchical component models, where marker paths are specified at the lower-order construct level at which the indicators are represented. The paper further provides practical decision rules for interpreting marker-variable results based on substantive path stability and marker-path behavior while emphasizing appropriate reporting practices and the limitations of marker-variable diagnostics. Rather than introducing a new statistical method, the paper provides an applied SmartPLS guide intended to improve the consistency, transparency, and rigor of marker-variable applications in PLS-SEM research.