Methods for Estimating Sample Size in Structural Equation Modeling

Background. Sample size estimation in Structural Equation Modeling (SEM) remains a debated methodological issue. Although minimum sample size rules are widely used, they often lack integration of theoretical, statistical, and practical considerations, particularly in complex research designs. Objectives. This study aims to provide a synthesis and comparative analysis of different approaches to sample size estimation in SEM, integrating theoretical foundations, statistical criteria, methodological limitations, and applied research contexts. Materials and Methods. A multi-method approach was employed, consisting of three components: a narrative review of sample size estimation methods, consideration of simulated scenarios reflecting realistic research conditions, and a comparative evaluation of method performance across these scenarios. Results. Findings show that rule-of-thumb approaches are useful for preliminary estimates but may misrepresent required sample sizes in complex models. Power analysis provides more precise and rigorous estimates, particularly for confirmatory research. Monte Carlo simulations offer robust guidance for complex, longitudinal, and multigroup models. Estimator-based recommendations highlight the importance of aligning sample size with data characteristics and estimation techniques. Model complexity emerged as a key determinant, with advanced SEM models often requiring 400–700+ participants. Conclusions. A systematic and context-sensitive approach to sample size estimation is essential for ensuring robustness and methodological rigor in SEM studies across disciplines such as management and education.