The value of facial attractiveness for encouraging fruit and vegetable consumption: analyses from a randomized controlled trial

Table 3 Standardized coefficients for all regression models investigating ratings of all facial attributes before and after 2 vs 5 vs 8 portions/day FV consumption over the past 4 weeks in all facial images*

	FV consumption – attractiveness¹
	β		95% CI		p.
FV consumption	0.06		−0.15, 0.26		0.57
	FV consumption – Mediating variables²			Mediating variables – Attractiveness³
	β	95% CI	p.	β	95% CI	p.
Skin yellowness	0.49	0.20, 0.78	< 0.01	0.07	0.02, 0.11	< 0.01
Skin yellowness squared	36.3	9.7, 62.9	0.01	< 0.01	< 0.01, < 0.01	0.04
Skin redness	− 0.02	− 0.28, 0.24	0.86	0.03	− 0.01, 0.08	0.12
Skin redness squared	8.08	−14.6, 30.8	0.48	< 0.01	- < 0.01, < 0.01	0.36
Skin healthiness	0.17	−0.04, 0.39	0.13	0.28	0.23, 0.33	< 0.01
Skin healthiness squared	3.7	18.9, 26.3	0.74	< 0.01	< 0.01, < 0.01	< 0.01
Skin clarity	0.02	−0.26, 0.30	0.90	0.16	0.12, 0.19	< 0.01
Skin clarity squared	−17.8	−45.9, 10.3	0.21	< 0.01	< 0.01, < 0.01	< 0.01
Facial symmetry	−0.08	−0.34, 0.18	0.52	0.18	0.13, 0.23	< 0.01
Facial symmetry squared	−18.4	−45.1, 8.3	0.17	< 0.01	< 0.01, < 0.01	< 0.01
	FV consumption – attractiveness including mediators⁴
	β		95% CI		p.
FV consumption	−0.01		− 0.22, 0.20		0.92
Skin yellowness	0.25		0.08, 0.42		< 0.01
Skin yellowness squared	- < 0.01		- < 0.01, − < 0.01		0.03
Baseline attractiveness	0.65		0.60, 0.70,		< 0.01

*Facial images were gained in Belfast, UK, 2011. Ratings of facial attributes were gained in Bournemouth, UK, 2013–2015
Significant relationships (p < 0.05) are emboldened
¹Primary models - ratings of perceived attractiveness were predicted using FV consumption over the past 4 weeks, when controlling for face ID and baseline ratings of attractiveness. Overall model performance R²=0.47, F(3,72) = 310.37, p < 0.01
²Secondary models - ratings of skin yellowness, skin redness, skin healthiness, skin clarity and facial symmetry were predicted using FV consumption over the past 4 weeks (when controlling for face ID and baseline variables)
³Secondary models - ratings of attractiveness were predicted using ratings of skin yellowness, skin redness, skin healthiness, skin clarity and facial symmetry (when controlling for face ID and baseline variables);
⁴Final models - ratings of perceived attractiveness were predicted using FV consumption over the past 4 weeks, when controlling for face ID, baseline ratings of attractiveness and significant mediators. Overall model performance R²=0.48, F(5,72) = 278.82, p < 0.01

ISSN: 1471-2458