Cross-cultural measurement invariance of the Eysenck Personality Questionnaire across 33 countries

Abstract

Cross-cultural invariance of personality measurement provides important information regarding the universality of personality traits. With the recent release of historical data from 33 countries on the Eysenck Personality Questionnaire (EPQ) the opportunity arose to test the invariance of the three personality dimensions measured by the EPQ, together with the response set scale. Although the factor structure of the EPQ has been much studied in previous decades, there was a need to validate the previously reported four-factor structure using modern factor analytic techniques. As anticipated, both exploratory and confirmatory factor analysis provided unsatisfactory models, for different reasons. Instead, exploratory structural equation modeling provided confirmation of the hypothesized oblique, four-factor structure. Results of measurement invariance comparisons across countries, separately by gender, showed striking evidence of the generalizability of the four-factor structure in pairwise comparisons with the English sample as the reference group. In addition, there was evidence of failure of item invariance, for a varying subset of items, in all of the between-country comparisons. However, the pattern of partial measurement invariance does not preclude effective use of the EPQ as a research tool in diverse cultures. The results provide strong theoretical convergence, with other published studies, on the universality of the four-factor structure.