Consistency of the Structural Properties of the BFI-10 Across 16 Samples From Eight Large-Scale Surveys in Germany

Abstract

The assessment of the Big Five personality domains is standard practice in most large-scale social surveys nowadays. The instrument most widely used for this purpose is the BFI-10, an ultra-short measure assessing each Big Five domain with two items. Recent studies have identified issues with the structural properties of the BFI-10, especially its factorial validity. To investigate whether these issues arise from the instrument itself or biases due to translation or sampling, we examined the extent to which the structural properties of the BFI-10 in terms of descriptive statistics, intercorrelations, reliability, and factorial validity vary when keeping the target population and language constant. Results revealed that, across 16 independent samples (total N ~ 60,000) from eight large-scale surveys representative of the adult population in Germany, the structural properties of the BFI-10 were (a) largely consistent and (b) mostly adequate. Most importantly, in nearly all samples, patterns of loading were congruent with an idealized Big Five structure, thereby supporting factorial validity. These results demonstrate that the structural properties of the BFI-10 are highly stable and replicable in large-scale samples. Especially given its brevity, the BFI-10 can thus be regarded as adequate for use in large-scale survey settings.

Differences between lifelong singles and ever-partnered individuals in Big Five personality traits and life satisfaction

“…The indices for each scale were already part of the SHARE dataset. Further, given that this brief measure did, expectedly, not lead to high internal consistency coefficients, but was seen as acceptable except for agreeableness, the SHARE team performed a validity check (see Rammstedt et al, 2023 for more supportive evidence of construct validity). More precisely, a principal component analysis using varimax rotation confirmed the Big Five factor structure when controlling for acquiescent responding, but only for the original BFI-10 items, not when adding the third agreeableness item (Levinsky et al, 2019).…”

“…Limitations of the current study include the cross-sectional design, which constrains investigations of dynamics or disentangling selection from socialization effects. Further, we had to rely on the specific assessments available in the panel study, including a rather unreliable measure of the Big Five personality traits (but see Rammstedt et al, 2023) that did not include finer-grained personality facets. A combination of self-and other-reports might have provided a more accurate estimate of people's traits.…”

Correction to Rammstedt et al. (2024)

“…An error occurred in the data preparation for the paper “Consistency of the structural properties of the BFI-10 across 16 samples from eight large-scale surveys in Germany” by B. Rammstedt, L. Roemer, & C. M. Lechner (2024; European Journal of Psychological Assessment , 40 , 204–215; https://doi.org/10.1027/1015-5759/a000765), resulting in the inadvertent duplication of participants from two datasets.…”