Reliability generalization meta-analysis of the internal consistency of the Big Five Inventory (BFI) by comparing BFI (44 items) and BFI-2 (60 items) versions controlling for age, sex, language factors.

Abstract

Introduction: The Big Five Inventory (BFI) is a popular measure that evaluates personality on the Big-Five model. Apart from its utilization across cultures, the literature did not reveal any meta-analysis for the reliability of the different versions of the BFI and its translations. The current study carried out a reliability generalization meta-analysis (REGEMA) to establish the reliability of the BFI across cultures and languages.

Methods: We searched 30 databases for the relevant studies from 1991 to mid-November 2024. The studies that we intended to include in our meta-analysis required to have utilized the BFI (44 items) and the BFI-2 (60 items) and have reported Cronbach's alpha or McDonald's omega reliability estimates. Our coded variables included BFI version, sample size, population type, age, gender, clinical state, and reliability. A total of 57 studies (datapoints) published in 34 research articles (involving 43,715 participants; 60.24% women; Mean age = 30.08) from various cultures and languages were finally included. These studies used BFI and BFI-2 in Arabic, Chinese, Croatian, Czech, Danish, Dutch, English, French, German, Indonesian, Italian, Japanese, Malay, Norwegian, Polish, Portuguese, Russian, Serbian, Spanish, Swahili, and Turkish. Data analysis was conducted using the metafor and meta packages in R. The average correlation was computed using a random-effects model and reliability coefficients indicated effect size. I² and Cochran's Q tests were used to examine heterogeneity, with prediction intervals suggesting genuine influences around the pooled estimate. Using funnel plots, regression-based tests (e.g., Egger's regression, rank correlation), and trim-and-fill imputation, publication bias was adjusted to estimate unbiased effects.

Results: We calculated the individual and combined reliability of the BFI and BFI-2 across languages and cultures. The results revealed the reliability of all five factors used in the BFI/BFI-2. The BFI estimates provide the following results: openness is estimated at 0.77 (95% CI: 0.75; 0.80); conscientiousness is estimated at 0.80 (95% CI: 0.78; 0.82); extraversion is also estimated at 0.80 (95% CI: 0.79; 0.82); agreeableness is estimated at 0.73 (95% CI: 0.71; 0.76); and neuroticism is estimated at 0.80 (95% CI: 0.79; 0.82). The BFI-2 estimates are as follows: openness is estimated at 0.83 (95% CI: 0.82; 0.84); conscientiousness is estimated at 0.86 (95% CI: 0.85; 0.87); extraversion is estimated at 0.85 (95% CI: 0.84; 0.86); agreeableness is also estimated at 0.80 (95% CI: 0.79; 81); and neuroticism is estimated at 0.89 (95% CI: 0.88; 0.89).

Conclusion: The current meta-analysis represents the first reliability analysis of the BFI and the first comparison between its two different versions, the BFI (44 items) and the BFI-2 (60 items). The generalized reliability of both the BFI and BFI-2 were established. The findings confirm that the BFI and BFI-2 have good reliability across all five factors.