The Potential for Differential Findings Among Invariance Testing Strategies for Multisample Measured Variable Path Models  

Oleh:

Mann, Heather M. ; Rutstein, Daisy W. ; Hancock, Gregory R.

Jenis: Article from Journal - ilmiah internasional
Dalam koleksi: Educational and Psychological Measurement vol. 69 no. 04 (Aug. 2009), page 603-612.
Topik: invariance testing; multisample; path analysis
Fulltext: 603.pdf (145.67KB)

Isi artikel

Multisample measured variable path analysis is used to test whether causal/structural relations among measured variables differ across populations. Several invariance testing approaches are available for assessing cross-group equality of such relations, but the associated test statistics may vary considerably across methods. This study is a population analysis, examining five different strategies for invariance testing using an illustrative measured variable path model. The results demonstrate how inferences about parameters across populations can depend greatly upon the invariance testing approach used, thereby potentially leading to improper inference regarding the true invariance status of relevant parameters.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali