作者: Gordon W. Cheung , Roger B. Rensvold
DOI: 10.1177/014920639902500101
关键词:
摘要: Many cross-cultural researchers are concerned with factorial invariance; that is, whether or not members of different cultures associate survey items, similar measures, constructs. Researchers usually test items for invariance using confirmatory factor analysis (CFA). CFA, however, poses certain problems must be dealt with. Primary among them is standardization, the process assigns units measurement to constructs (latent variables). Two standardization procedures and several minor variants have been reported in literature, but these when testing can lead inaccurate results. In this paper we review basic theory, propose an extension Byrne, Shavelson, Muthgn’s (1989) procedure identifying non-invariant items. The extended solves problem by performing a systematic comparison all pairs loadings across groups. A numerical example based upon large published...