Performance of Second-Order Latent Growth Model Under Partial Longitudinal Measurement Invariance: A Comparison of Two Scaling Approaches. Issue 2 (4th March 2021)

• Record Type: Journal Article
• Authors: Jeon, Min-Jeong; Kim, Su-Young
• Is Part Of: Structural equation modeling. Volume 28:Issue 2(2021)
• Page Start: 261
• Page End: 277
• Publication Date: 2021-03-04
• Subjects: Second-order latent growth models -- scaling methods -- marker variable -- effects coding -- longitudinal measurement invariance; Multivariate analysis -- Periodicals; Social sciences -- Statistical methods -- Periodicals; 519.535
• DOI: 10.1080/10705511.2020.1783270
• Languages: English
• ISSNs: 1070-5511

Abstract

ABSTRACT: Second-order latent growth models (SLGMs) have recently been highlighted over the traditional first-order latent growth model. Although SLGMs can show several intuitive strengths, the model has remained less understood due to the issue of scaling-related misspecification. As one source of model misspecification, scaling could influence the estimation of SLGM. Since the impact could differ depending on which scaling method is employed, selecting a scaling method becomes crucial for the practical use of SLGM. The present study investigated and compared the impact of two different scaling methods for the estimation of SLGM under various partial measurement invariance situations. The results of comprehensive Monte Carlo simulations do not support a single superior scaling method under all generated partial MI conditions. In this regard, the careful and strategic selection of a scaling method in the context of SLGM is required, as discussed in the final section of the study.