Table 2. Results of the pre-study (Model 1).

	Coefficient	z	SE	p
Fixed effects
Level 1
Intercept	2.92	20.51	0.14	<.001
I 1it	-0.01	-0.07	0.11	.941
I 2it	0.10	0.68	0.15	.496
C it	2.59	14.72	0.18	<.001
I 1it C it	-0.05	-0.38	0.13	.701
I 2it C it	-0.17	-0.94	0.18	.350
Random effects
s 2(ε it)	1.49	37.30	0.04	<.001
s 2(υ 0t)	0.41	3.02	0.13	.003
s 2(υ 1t)	0.16	2.32	0.07	.020
s 2(υ 2t)	0.20	2.45	0.08	.014
s 2(υ 3t)	0.45	3.05	0.15	.002
s(υ 0t, υ 1t)	0.06	0.81	0.07	.418
s(υ 0t, υ 2t)	-0.07	-0.95	0.08	.343
s(υ 0t, υ 3t)	-0.22	-1.96	0.11	.050
s(υ 1t, υ 2t)	0.09	1.57	0.06	.116
s(υ 1t, υ 3t)	-0.14	-1.86	0.08	.063
s(υ 2t, υ 3t)	-0.08	-0.97	0.08	.332

I _1it and I _2it are Indicator or Dummy variables indicating the stimulus category (I _1it: 1 = African infants, 0 = Caucasian infants; I _2it: 1 = dog puppies, 0 = Caucasian infants). C _it represents cuteness category (C _it: 0 = less cute, 1 = cute). Interactions are I _1it C _it and I _2it C _it. Robust estimators were used for statistical inference with respect to fixed effects and variance components to account for possible violations of model assumptions, such as normality of Level-2 residuals. Degrees of freedom were computed based on the Satterthwaite’s Approximation to account for the moderate sample size at Level 2 [46]. Therefore, the degrees of freedom were not necessarily integers and could vary across tests independent of the number of parameters.