In the present study, we investigated the structural invariance of the Five-Factor Personality Inventory (FFPI) across a variety of cultures. Self-report data sets from ten European and three non-European countries were available, representing the Germanic (Belgium, England, Germany, the Netherlands, USA), Romance (Italy, Spain), and Slavic branches (Croatia, Czech Republic, Slovakia) of the Indo-European languages, as well as the Semito-Hamitic (Israel) and Altaic (Hungary, Japan) language families. Each data set was subjected to principal component analysis, followed by varimax rotation and orthogonal Procrustes rotation to optimal agreement with (i) the Dutch normative structure and (ii) an American large-sample structure. Three criteria (scree test, internal consistency reliabilities of the varimax-rotated components, and parallel analysis) were used to establish the number of factors to be retained for rotation. Clear five-factor structures were found in all samples except in the smallest one (USA, N = 97). Internal consistency reliabilities of the five components were generally good and high congruence was found between each sample structure and both reference structures. More than 80% of the items were equally stable within each country. Based on the results, an international FFPI reference structure is proposed. This reference structure can facilitate standardized communications about Big Five scores across research programmes. Copyright © 2003 John Wiley & Sons, Ltd.