Table 1.

Effect sizes found in studies that used integral facial images to predict personality traits.

	Nphoto	Nrater	Effect size, r
			E	A	C	N	O
Artificial feature manipulation
Walker & Vetter, 2016, Study 3	30	160	0.23	0.27	0.12	0.36	0.24
Composite facial images
Little & Perrett, 2007	20	40	0.19	0.06	0.17	0.11	0.10
Penton-Voak et al., Study 2	20	42	0.56	0.24	0.63	0.37	0.47
Kramer & Ward, 2010	20	131	0.78	0.37	0.06	0.28	−0.11
Realistic photographs
Penton-Voak et al. 2006, Study 1	294	10	0.24	−0.01	0.06	0.17	0.18
Qui et al. 2015, Selfies	123	8	0.02	0.06	0.10	0.07	0.21
Borkenau et al., 2009	149	24	0.34	0.04	0.16	0.07	0.11
Naumann et al., 2009, Standard	123	6	0.39	−0.11	−0.03	0.17	0.17
Naumann et al., 2009, Spontaneous	123	6	0.42	0.20	0.12	0.18	0.35
Machine learning models
Hu et al., 2017, Male faces	405	n/a	0.16	0.11	0.21	0.15	0.14
Hu et al., 2017, Female faces	429	n/a	0.11	0.02	−0.06	−0.06	0.00

Note: N_photo – number of stimulus faces, N_rater – number of human raters per photograph, E – extraversion, A – agreeableness, C – conscientiousness, N – neuroticism, O – openness. Effect sizes are given as reported by the authors where available or calculated using meta-analytic formulae²⁶.