Structural validity and measurement invariance of the short version of the Big Five Inventory (BFI-10) in selected countries

Figures & data

Table 1. B5P-traits and their quantification

Items: I see myself as someone who …	B5P traits	Reverse coding
… is reserved	Extraversion	Yes
… is generally trusting	Agreeableness	No
… tends to be lazy	Conscientiousness	Yes
… is relaxed, handles stress well	Neuroticism	No
… has few artistic interests	Openness	Yes
… is outgoing, sociable	Extraversion	No
… tends to find fault with others	Agreeableness	Yes
… does a thorough job	Conscientiousness	No
… gets nervous easily	Neuroticism	Yes
… has an active imagination	Openness	No

Table 2. Model fit cut-off criteria

Measure	Terrible	Acceptable	Excellent
CMIN/DF	> 5	> 3	> 1
P-value	0.05<	>0.05
CFI	<0.90	<0.95	>0.95
RMSEA	>0.08	>0.06	<0.06
PClose	<0.01	<0.05	>0.05

Source: Hu and Bentler (1999).

Table 3. Demographic variables and WEIRD classification

		Germany	Nether-lands	Rwanda	South Africa
Sex	Men	48.9	46.5	49.6	48.3
	Women	51.1	53.5	50.4	51.7
Age	Mean	49.63	53.34	33.77	37.72
	Std. Dev.	17.94	16.44	11.23	15.67
W—Westernised ^a	White ^b	87.1	75.7	Low ^d	12.1
	Black ^c	.7	1.3	High ^d	76.5
E—Educated	Graduated ^e	11.7	11.5	7.7	4.2
I—Industrialised	PC use ^f	58.7	80.5	12.7	13.7
R—Rich	Savings ^g	57.0	50.5	30.6	19.6
D—Democratic	Perceived ^h	50.2	47.5	42.9	36.6

^aGermany and the Netherlands are countries in Europe, and can therefore be classified as Western, whilst Rwanda and South Africa are African countries. However, ethnicity is also used as an indicator of being defined as either Western or non-Western. This aspect is reported here as well, with “white” interpreted as Western, and “black” as non-Western.

^bListed as Caucasian white in Germany and the Netherlands, and as white in South Africa.

^cListed as African in Germany, black Negro in the Netherlands, and black in South Africa.

^dNo questions on ethnic groups were posed to the Rwandan sample. However, when considering V247, namely “What language do you speak at home”, 97.6% of respondents indicated Kinyarwanda and 0.8% Swahili, with 0.1% indicating English and 1.5% French. It would be a reasonable assumption that most respondents in the Rwandan sample would have been classified as African, and therefore as non-Western.

^eV248 was used to create the variable “Educated”. Presented here is the percentage of respondents who indicated that they had completed a university degree.

^fV225 was used to create the ”Industrialised” variable. The percentage of respondents who indicated that they use their personal computer frequently, rather than occasionally or seldom, is presented here.

^gV141 was used to create the “Rich” variable. The percentage of respondents who provided answers to 8, 9 and 10 is relevant in this case.

^hV237 was used to create the “Democratic” variable. Presented here is the percentage of respondents who stated that they had sufficient savings.

Table 4. Germany: EFA with components (a priori) and component (by eigenvalues greater than 1)

G	Five components (a priori)					Component with eigenvalues >1
	1	2	3	4	5	1	2	3	4	-
E(R)	−.769	.091	−.042	.041	.157	−.789	−.035	.039	−.017	-
A	−.003	−.085	.008	−.09	.878	−.123	.342	.416	.215	-
C(R)	−.09	−.014	−.004	.872	.025	−.167	.131	−.794	.100	-
N	−.073	−.713	.12	.233	.33	−.100	.801	−.088	.164	-
O(R)	.026	.165	−.815	.177	.126	−.032	−.081	−.126	−.709	-
E	.752	−.128	.185	−.083	.187	.728	.17	.132	.233	-
A(R)	.731	.176	−.011	.097	−.047	.707	−.164	−.125	.031	-
C	−.15	−.042	.138	−.603	.444	−.158	.098	.731	.166	-
N(R)	−.08	.853	.028	.164	.086	−.153	−.757	−.116	.137	-
O	.221	.107	.748	.093	.199	.186	−.038	−.007	.794	-

Note: G = Germany, E = Extraversion, A = Agreeableness, C = Conscientiousness, N = Neuroticism, O = Openness, R = Reverse coding. Also, in bold in the table are the two highest loadings per component, and “shadowed” are the a priori loadings, as per the B5P model, where the highest loading was used as a marker item.

Table 5. Netherlands: EFA with components (a priori) and component (by eigenvalues greater than 1)

N	Five components (a priori)					Component with eigenvalues >1
	1	2	3	4	5	1	2	3	4	5
E(R)	.109	.117	−.042	−.774	.296	.109	.117	−.042	−.774	.296
A	−.120	−.090	−.022	.008	.785	−.120	−.090	−.022	.008	.785
C(R)	.138	−.059	.825	−.134	.138	.138	−.059	.825	−.134	.138
N	−.781	.065	.045	.136	.297	−.781	.065	.045	.136	.297
O(R)	.050	.846	.028	.017	.074	.050	.846	.028	.017	.074
E	−.018	−.049	−.152	.736	.327	−.018	−.049	−.152	.736	.327
A(R)	.461	.241	.345	.297	.152	.461	.241	.345	.297	.152
C	.090	.080	−.673	−.043	.407	.090	.080	−.673	−.043	.407
N(R)	.853	−.050	.062	−.055	.047	.853	−.050	.062	−.055	.047
O	.113	−.701	.140	.172	.200	.113	−.701	.14	.172	.200

Note: N = Netherlands, E = Extraversion, A = Agreeableness, C = Conscientiousness, N = Neuroticism, O = Openness, R = Reverse coding. Also, in bold in the table are the two highest loadings per component, and “shadowed” are the a priori loadings, as per the B5P model, where the highest loading was used as a marker item.

Table 6. Rwanda: EFA with components (a priori) and component (by eigenvalues greater than 1)

R	Five components (a priori)					Component with eigenvalues >1
	1	2	3	4	5	1	2	3	4	-
E(R)	.261	.819	.132	.054	−.100	.323	.251	.753	.001	-
A	−.058	.869	.039	.009	.218	−.036	.038	.883	.129	-
C(R)	−.097	.119	.834	.015	.135	−.237	.718	.174	.097
N	.187	.211	.034	.272	.808	.047	−.169	.352	.683	-
O(R)	.010	−.008	.100	.927	.098	.071	.259	−.132	.825	-
E	.462	.202	−.376	.482	.207	.523	−.273	.148	.526	-
A(R)	.421	.322	.320	.326	−.534	.528	.593	.136	−.008	-
C	.824	.107	.008	.072	−.062	.816	.083	.073	.054	-
N(R)	.315	.067	.665	.075	−.277	.272	.732	.008	−.071	-
O	.843	.032	.093	.006	.119	.770	.075	.058	.102	-

Note: R = Rwanda, E = Extraversion, A = Agreeableness, C = Conscientiousness, N = Neuroticism, O = Openness, R = Reverse coding. Also, in bold in the table are the two highest loadings per component, and “shadowed” are the a priori loadings, as per the B5P model, where the highest loading was used as a marker item.

Table 7. South Africa: EFA with components (a priori) and component (by eigenvalues greater than 1)

p	Five components (a priori)					Component with eigenvalues >1
	1	2	3	4	5	1	2	-	-	-
E(R)	.183	.142	.127	.892	.033	.632	.128	-	-	-
A	−.081	.306	.28	.683	.237	.778	−.024	-	-	-
C(R)	.846	−.111	.145	.018	.052	−.062	.824	-	-	-
N	.104	.28	.846	.159	.048	.683	.188	-	-	-
O(R)	.196	.13	.723	.221	.326	.639	.34	-	-	-
E	.096	.261	.25	.133	.825	.581	.364	-	-	-
A(R)	.627	.135	.044	.114	.564	.256	.764	-	-	-
C	−.004	.836	.189	.189	.171	.733	.073	-	-	-
N(R)	.699	.479	.123	.123	.063	.345	.693	-	-	-
O	.164	.626	.295	.271	.245	.715	.253	-	-	-

Note: S = South Africa, E = Extraversion, A = Agreeableness, C = Conscientiousness, N = Neuroticism, O = Openness, R = Reverse coding. Also, in bold in the table are the two highest loadings per component, and “shadowed” are the a priori loadings, as per the B5P model, where the highest loading was used as a marker item.

Table 8. CFA model fit measures for four countries

Measure	Threshold	South Africa	Rwanda	Nether-lands	Germany
Chi-square, degrees of freedom, p-value	<0.05	Chi-square = 974.554, df = 34, p-value = 0.000	Chi-square = 947.158, df = 25, p-value = 0.000	Chi-square = 284.189, df = 25, p-value = 0.000	Chi-square = 474.078, df = 26, p-value = 0.000
CMIN/DF	1< and >3	∞	∞	NaN	NaN
CFI	>.95	.817	.684	.831	.789
RMSEA	<.06	.122	.155	.074	.092
PClose	>.05	0	0	0	0

Note: ∞ = infinity; NaN = Not a Number

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

 Web of Science ®Google Scholar

Data availability statement

The data that support the findings of this study are openly available on the following website: http://www.worldvaluessurvey.org/WVSDocumentationWV6.jsp