Rethinking Invariance: A Monte Carlo Evaluation of Critical Structural Equation Modeling

• In Event: AERA Roundtable Session Thursday 7:45 am - Gold 4
  In Roundtable Session: Measurement Complexity in SEM: Cross-Loading, Invariance, and Heterogeneity

Abstract

Critical Structural Equation Modeling (CritSEM) treats measurement non-invariance and structural heterogeneity not as errors, but as evidence of socially contingent construct meaning. This Monte Carlo simulation compares traditional SEM and CritSEM across four conditions: full invariance, measurement non-invariance, structural heterogeneity, and both. Using a moderated mediation model, we find traditional SEM yields biased estimates and masks group-specific patterns under inequivalence. CritSEM, by contrast, recovers true relationships, preserves construct validity, and reveals inequities. Rather than forcing uniformity, CritSEM honors contextual variation, offering a more valid and equity-aware modeling approach. Findings underscore the need to rethink invariance as a statistical necessity and instead view it as a site for critical insight into identity, context, and structural inequality.

Authors

• Zachary K. Collier, University of Connecticut

• Joshua Sukumar, University of Connecticut

• Kirsten Reyna, University of Connecticut

• Nicholas S. Bell, University of Connecticut

• Veronica Velez, Western Washington University