Reducing Uniform Response Bias With Ipsative Measurement in Multiple-Group Confirmatory Factor Analysis

Response bias, defined by Paulhus (1991) as "a systematic tendency to respond to a range of questionnaire items on some basis other than the specific item content," has been observed in various disciplines, especially in cross-cultural research. In this study, a mathematical model of uniform response bias (URB) is defined. Ipsative measures (i.e., individual scores subject to a constant sum constraint) are proposed to minimize the effect of URB in multigroup confirmatory factor analysis (CFA) to study the measurement invariance properties across different cultural groups. The method based on Chan and Bentler (1993, 1996) for analyzing ipsative data is extended here for analyzing multigroup data potentially contaminated by URB. A real data set based on the Chinese Personality Assessment Inventory (CPAI) is used to demonstrate how the proposed procedure can be applied in real-life situations.