APPLICATION OF PARTIAL LEAST SQUARES STRUCTURAL EQUATION MODELING (PLS-SEM) IN PSYCHOLOGICAL RESEARCH: MEASUREMENT AND STRUCTURAL MODEL ASSESSMENT USING SMARTPLS
Mohammed Shamsudeen T. Akeem A. Kenku, PhD and Shafa A. Yunus, PhD
Volume 13 Issue 2
Abstract
Partial Least Squares Structural Equation Modeling (PLS-SEM) has emerged as a powerful and flexible multivariate analytical technique increasingly adopted in psychological research to examine complex theoretical models involving latent constructs. Unlike covariance-based SEM (CB-SEM), PLS-SEM operates through a component-based estimation approach that accommodates small sample sizes, non-normal data distributions, and formative measurement specifications, conditions frequently encountered in psychological investigations. This methodological paper provides a comprehensive guide to applying PLS-SEM using SmartPLS software, covering the philosophical foundations, model specification, and sequential assessment procedures for both measurement and structural models. The study delineates the two-stage assessment protocol: first evaluating the outer measurement model through reliability (Cronbach's Alpha, Composite Reliability, rho_A) and validity (convergent validity via Average Variance Extracted, and discriminant validity via Fornell-Larcker criterion, cross-loadings, and HTMT ratio) indices; and second, assessing the inner structural model through path coefficients, coefficient of determination (R²), effect sizes (f²), predictive relevance (Q²), and Standardized Root Mean Square Residual (SRMR). The paper also addresses advanced techniques, including mediation and moderation analysis, common method bias assessment, and reporting standards for publication. Practical recommendations are offered throughout to assist psychological researchers in making methodologically sound decisions within the PLS-SEM framework. Keywords: Partial Least Squares, Structural Equation Modeling, Psychological, Research, SMARTPLS
Download Paper