On the Dependability of a Popular Procedure for Studying Measurement Invariance: A Cause for Concern?. Issue 4 (3rd July 2020)

• Record Type: Journal Article
• Authors: Raykov, Tenko; Marcoulides, George A.; Harrison, Michael; Zhang, Mingcai
• Is Part Of: Structural equation modeling. Volume 27:Issue 4(2020)
• Page Start: 649
• Page End: 656
• Publication Date: 2020-07-03
• Subjects: Cross-group equality -- factor loading -- intercept -- measurement invariance -- parameterization; Multivariate analysis -- Periodicals; Social sciences -- Statistical methods -- Periodicals; 519.535
• DOI: 10.1080/10705511.2019.1610409
• Languages: English
• ISSNs: 1070-5511

Abstract

Abstract : A theoretically and empirically important multiple-population setting is discussed where a highly popular procedure for studying measurement invariance (MI) can fail. A numerical counter-example is presented where this widely used, traditional method misses a sizeable violation of MI. It is argued that the reason for the failure is in part inherent in the parameterization approach underlying that popular approach to examining MI. Implications for empirical behavioral, social, educational, business, and marketing research are discussed in light of these findings.