SEM Method Choice in Resource-Constrained Settings: CB-SEM, PLS-SEM, or PLSc-SEM?

Resource-constrained contexts present high-variance data conditions, theoretical immaturity, and institutional instability, which challenge existing SEM guidelines. Empirical evidence comparing the predictive performance of covariance-based SEM (CB-SEM), partial least squares SEM (PLS-SEM), and consistent PLS-SEM (PLSc-SEM) under these compound contextual factors is lacking in SEM literature․ We test the methods on a sample of private-sector employees in Iraq. We find a systematic trade-off between PLS-SEM (substantially better measurement properties but only marginal global fit), PLSc-SEM (excellently fitting but poor convergent validity), and CB-SEM (requires re-specification to attain acceptable fit and scores in between the other two methods on measurement properties). PLS-SEM shows a better level of sensitivity for exploratory research and theory development. On the other hand, PLSc-SEM and CB-SEM performed differently across criteria for confirmatory research․ This evidence-based roadmap for SEM method selection provided by this study calls for a context-sensitive approach and methodological recommendations for research in resource-constrained contexts.