Comparison of Component-Based Structural Equation Modeling Methods in Testing Component Interaction Effects

Generalized structured component analysis (GSCA) and partial least squares path modeling (PLSPM) are statistical methods for component-based structural equation modeling, where constructs are represented as components or weighted sums of observed variables. While both methods employ different approaches to testing an interaction effect between components, their comparative performance in this context has not been systematically investigated. To address this gap, we conducted a simulation study to evaluate their relative performance across various experimental conditions. The results indicate that GSCA’s approach consistently outperforms PLSPM’s in testing a component interaction effect under all examined conditions.