A commentary on “Magic number .95? Or was it .08? A refresher on SEM approximate fit indices thresholds for applied psychologists and management scholars”
Year: 2025
Type: editorial
Abstract: This article explores critical controversies in Structural Equation Modeling (SEM) and Confirmatory Factor Analysis (CFA), focusing on the misuse of approximate fit indices, criticisms of their application, and the overreliance on rigid rules. Drawing on foundational work by Hu and Bentler and recent advancements such as dynamic fit thresholds and resampling methods, the manuscript highlights the limitations of fixed thresholds and proposes alternative, context-sensitive approaches. Additionally, it examines the merits of the Partial Least Squares-Structural Equation Modeling (PLS-SEM) approach, including the consistent PLSc estimator, as a robust alternative to traditional covariance-based SEM (CB-SEM). By integrating these methodologies, the paper advocates for a multidimensional and flexible evaluation framework that enhances model accuracy and reliability. Practical recommendations include adopting dynamic thresholds, leveraging advanced methodologies like PLSpredict for predictive validation, and balancing reliability with construct validity. The manuscript underscores the importance of moving beyond rigid rules to embrace a nuanced understanding of model evaluation, ultimately contributing to more robust findings in social science research.
Source: Research in Statistics
Author: The Cuong Nguyen
Institutions: Vietnam National University, Hanoi, VNU International School
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4413928110