Table 1.

Reliability, Falconer’s estimates, and variance component estimates for individual differences in trustworthiness, dominance, and attractiveness facial impressions (as measured by facial impression scores; see main text)

	Trustworthiness	Dominance	Attractiveness
Reliability
Internal (split-half)	0.75	0.82	0.80
Test–retest	0.73	0.58	0.52
Falconer’s formula: ACE estimates
A	0.11	0.10	0.14
C	0.20	0.08	0.10
E	0.70	0.82	0.77
Model fit: −2 log-likelihood; AIC (P value for difference in −2 log-likelihood)
ACE	3,016.32; 732.32	2,984.34; 718.34	2,781.98, 503.98
AE	3,016.71; 730.71, P = 0.530	2,984.34; 716.34, P > 0.999	2,781.98; 501.98, P > 0.999
CE	3,017.16; 731.16, P = 0.359	2,985.73; 717.73, P = 0.238	2,784.66; 504.66, P = 0.102
E	3,051.54; 763.54, P < 0.001	2,997.67; 727.67, P = 0.001	2,813.3; 531.3, P < 0.001
Full model: ACE estimates (95% CI)
A	0.17 (0; 0.37)	0.18 (0; 0.27)	0.30 (0; 0.39)
C	0.11 (0; 0.33)	0 (0; 0.22)	0 (0; 0.30)
E	0.72 (0.63; 0.81)	0.82 (0.73; 0.92)	0.70 (0.61; 0.81)
Best-fit model: AE estimates (95% CI)
A	0.29 (0.20; 0.38)	0.18 (0.08; 0.27)	0.30 (0.20; 0.39)
E	0.71 (0.62; 0.80)	0.82 (0.73; 0.92)	0.70 (0.61; 0.80)

A, additive genetic influences; C, shared environmental influences; E, both personal environmental influences and measurement error. Lower and upper bound 95% CIs are shown in parentheses for model estimates so that each estimate that does not touch zero is significant at alpha = 0.05 (two-sided). ACE, df = 1,142, 1,133, 1,139 for trustworthiness, dominance, attractiveness. difference in df ACE-AE = 1; difference in df ACE-CE = 1; difference in df ACE-E = 2.