The Questionnaire for Eudaimonic Well-being (QEWB) was proposed to have a unidimensional structure (Waterman et al., 2010), but the statistical assumptions on which the analysis was based was questioned by Schutte et al. (2013). Instead, a three- and four-factor structure was proposed. The present study aimed to explore the psychometric properties of the one-, three-, and four-factor structures of the Afrikaans (QEWB_A, n=478), English (QEWB_E, n=325), and Setswana (QEWB_S, n=258) versions of the scale in three South African student groups and to test for measurement invariance. Confirmatory factor analysis (CFA), bifactor-CFA, exploratory structural equation modelling (ESEM) and bifactor-ESEM were applied. The one-factor structure revealed poor fit for all groups. Bifactor-ESEM models fitted the data best. The three-factor bifactor-ESEM model was selected for parsimony and to test measurement invariance of the QEWB_A and QEWB_E. A reduced three-factor bifactor-ESEM model fitted the QEWB_S. Reliabilities and measurement invariance were explored. The items seem to be influenced by a global and some specific factors. Cross-loadings on nontarget factors suggest that the specific QEWB factors are interrelated. The scale and theory seems appropriate for all samples. The results provide insight into the dimensionality of eudaimonic well-being, and the scale shows potential for use in practice.