A test of a new short Big-Five tool in large probabilistic samples from 19 countries. (1st December 2019)

• Record Type: Journal Article
• Authors: Minkov, Michael; van de Vijver, Fons J.R.; Schachner, Michael
• Is Part Of: Personality and individual differences. Volume 151(2019)
• Publication Date: 2019-12-01
• Subjects: Personality -- Five-Factor model -- Big Five -- Cross-cultural psychology; Personality -- Periodicals; Individuality -- Periodicals; Individuality -- Periodicals; Personality Development -- Periodicals; Personnalité -- Périodiques; Individualité -- Périodiques; 155.205
• DOI: 10.1016/j.paid.2019.109519
• Languages: English
• ISSNs: 0191-8869

Abstract

Abstract: Few short Big Five instruments have been tested for invariance across samples from a large number of countries. To fill this gap, we measured the Big Five personality traits in probability samples ( n > 1000 in each country) from Belgium, Brazil, China, France, Germany, India, Indonesia, Italy, Japan, Korea, Mexico, Nigeria, Philippines, Russia, Thailand, Turkey, UK, US, and Vietnam. We used a new, short Big Five questionnaire, based on the BFI and other short instruments after consultations and tests in diverse cultures. Across the US sample ( n = 2811, nationally representative), as well as in Japan and other countries, our 15-item questionnaire yielded higher loadings on intended factors and lower cross-loadings than what has been reported for the BFI, and the BFI-2-XS in the US and worldwide. Exploratory Structural Equation Modeling of our tool recovered the Big Five, and its structure was very stable across countries, yet without metric comparability. We conclude that the Big-Five model covers important cross-cultural commonalities in the structure of personality, and our study reveals its invariant core across many countries. Yet metric comparability is compromised for some items. Cross-cultural comparisons of scores on our measure, and most likely other Big Five measures, seem premature. Highlights: Big-Five studies across many countries at a time recover the model but the item loadings are unimpressive or unreported. We tested a new short Big-Five tool in probabilistic samples ( n > 1000 per nation) in 19 nations on all continents. Our tool recovers the Big Five in each nation and yields higher loadings than previous short tools. Despite the configural invariance, we did not obtain sufficient metric invariance. This outcome makes us skeptical about the validity of large-scale comparisons of nations on the Big Five.