Within the item response theory (IRT) framework, this study compared cross-culturally different approaches to the assessment of differential item functioning (DIF) in two personality tests of the Comprehensive Personality Scale (Wang, 2013). A dominance IRT model (SGRM) and an ideal point model (the GGUM) were applied within the NHST paradigm, due to the debate over which is the more appropriate model for personality research. Nye’s (2011) DIF effect size measure was also used in the current study to overcome the oversensitivity of NHST to large sample size. Participants from the U.S. (n = 861) and China (n = 1023) responded to two personality scales from the CPS: the Well-being scale, and the Curiosity scale. Results indicated that SGR was applicable for DIF assessment, but the NHST paradigm was so sensitive to large samples that even trivial DIF could be significant. GGUM failed to work in the DIF analyses due to ill-conditioned matrices. The DIF effect size measure compensated for the NHST method by providing the magnitude of DIF. Implications for future research and practice are discussed.
【 预 览 】
附件列表
Files
Size
Format
View
Measurement equivalence of the Comprehensive Personality Scales across cultures: an item response theory approach