Structural validity of the BFI-44: impact of using CFA vs. ESEM and reverse-worded items

The Big Five Inventory (BFI-44; John, Donahue, & Kentle, 1991) is a commonly used measure of individual differences in personality, but previous research has largely failed to show strong evidence of its structural validity. This study takes into consideration recommendations by Hopwood and Donnellan (2010) and Sellbom and Tellegen (2019) for how to better assess and interpret personality assessment structure: carefully choosing CFA estimation method, comparing CFA to ESEM results, and interpreting model fit in context with theory and external criterion validity. In addition, this study also examines whether reverse-worded items could explain previous evidence of poor model fit. Using BFI-44 data collected from university students, best model fit was found using ESEM when reverse-worded items were removed; however, removing the reverse-worded items did not result in improved prediction compared to the original measure. This study supports previous work suggesting CFA with default estimation methods may not be the best assessment of personality measure model fit and reverse-worded items may be impacting model fit, though more research into the impact of removing the reverse-worded BFI-44 items on criterion validity is needed.