Locating a Truly Invariant Reference Variable in Testing Factorial Invariance: Using the Smallest Modification Index

• In Event: 65.079 - Roundtable Session 23
  In Roundtable Session: 65.079-17 - Measurement Invariance: Practical Applications

Abstract

When testing factorial invariance, identification of a multi-group confirmatory factor analysis (MCFA) model is usually done by imposing invariance constraints on parameters of chosen reference variables (RV). If the chosen RVs were not actually invariant, one could draw invalid conclusions regarding the source of noninvariance. However, no method is yet available. This study proposed a promising method for selecting a truly invariant RV. Under various simulated partial factorial invariance scenarios this study evaluated the accuracy of the smallest modification index (Min-Mod) in identifying a truly invariant RV. Results indicated that the Min-Mod accurately identified a truly invariant set of factor loading or intercept with very high accuracy rates. Limitations, and implications of the study are also discussed.

Authors

• Eunju Jung, Indiana University - Bloomington

• Myeongsun Yoon, Texas A&M University