. 2021 Sep 21;12:718004. doi: 10.3389/fpsyg.2021.718004

TABLE 4.

Relationship between perceiver attributes and rating.

Model	Formula	Tested parameter	χ²	df	p	AIC	BIC
LMM0	Rating_ij = β₀ × β₁ image_i × β₂ morph_j + υ_0id + υ_id:image + υ_id:morph + υ_id:ident + ε_ij					181743	181814
LMM1	Rating_ijk = β₀ × β₁ image_i × β₂ morph_j × β₃ sex_k + υ_0id + υ_id:image + υ_id:morph + υ_id:ident + ε_ijk	β₃/sex	2.64	4	0.621	181748	181851
LMM2	Rating_ijk = β₀ × β₁ image_i × β₂ morph_j × β₃ GERAS_k + υ_0id + υ_id:image + υ_id:morph + υ_id:ident + ε_ijk	β₃/GERAS	29.57	4	<0.001^*	181721	181824
LMM3	Rating_ijkl = β₀ × β₁ image_i × β₂ morph_j × β₃ GERAS_k × β₄ GIS_l + υ_0id + υ_id:image + υ_id:morph + υ_id:ident + ε_ijkl	β₄/GIS	57.14	8	<0.001^*	181680	181846

β₀, mean rating for superfemale images in female participants; β₁, estimate for the effect of sex of image (original male or female); β₂, estimate for the effect of morphing (0–140); β₃, estimate for the effect of sex (male or female) or gender (masculine or feminine) of participants; β₄, estimate for the effect of sexual attraction (attraction toward males or females); υ_0id, random effect of participants on intercept of rating; υ_id:image, random effect of participants and image on intercept of rating; υ_id:morph, random effect of participants and morphing on intercept of rating; υ_id:ident, random effect of participants and face identity on intercept of rating; ε_ij, residual variance; image_i, sex of image (original male or female); morph_j, morphing level (0–140); sex_k, sex of participants; GERAS_k, gender of participants (masculine or feminine); GIS_l, sexual attraction of participants (attraction toward males or females); AIC, Akaike information criterion; BIC, Bayesian information criterion; LMM0, basic model; LMM1, model with interaction effect of sex; LMM2, model with interaction effect of gender; and LMM3, model with interaction effect of gender and sexual attraction.

*p < 0.05.