Factor structure and psychometric properties of the French versions of the Big Five Inventory-2 Short (BFI-2-S) and Extra-Short (BFI-2-XS) Forms.

Objectives: In many circumstances, researchers and clinicians need short personality questionnaires that can be administered quickly, either because a patient's attentional abilities preclude the use of longer, more comprehensive questionnaires, or because the assessment of personality, while desirable, is not the primary focus of their approach. The purpose of this study was to validate two short forms of one of the personality questionnaires most widely used in international research, the Big Five Inventory (BFI-2), which uses 60 items to measure the Big Five personality factors: Extraversion, Agreeableness, Conscientiousness, Negative Emotionality/Neuroticism, and Open-mindedness/Openness to Experience, plus 15 facets (three for each major factor). The 30-item short form (BFI-2-S) is intended to assess the five main factors and the 15 facets and the "extra-short" form (BFI-2-XS) in 15 items only allows the five big factors to be assessed.

Methods: We had 1003 students (540 women, 54%) from two universities in central-eastern France complete the full BFI-2. We analyzed the factorial structure of the two sets of items making up the BFI-2-S and the BFI-2-XS using an exploratory structural equation model and assessed the reliability of the scales measuring the big five factors or domains using conventional indices. The reliability of the scales measuring the BFI-2-S facets was estimated based on the significance and magnitude of the correlation between the two items of which they are composed.

Results: The complex five-factor structure in which the 15 facets are modeled as residual correlations between the two items of each facet provided an excellent fit to the BFI-2-S data. The reliability of the scales measuring the five main factors was very satisfactory and that of the scales measuring the facets was acceptable to good. The BFI-2-S domain scales were free from differential item functioning related to gender, and the latent means of males and females could be compared. The high correlations of the domain and facet scales with the corresponding scales of the full BFI-2 argues in favor of the validity of the BFI-2-S. With respect to the BFI-2-XS, the simple five-factor structure also exhibited a satisfactory fit, and the five scales also had high correlations with the full BFI-2 corresponding scales. However, the invariance of the items with respect to gender was not as well established as for the BFI-2-S.

Discussion: This study's main limitation was the composition of the sample which was made up of students. More research is needed to verify whether these results are generalizable to older and less educated populations.

Conclusion: The BFI-2-S and the BFI-2-XS provide a quick and reliable assessment of the Big Five personality factors. The BFI-2-S, which also provides an evaluation of six facets that seems reliable, appears to offer the best trade-off between parsimony and quality of evaluation.