Table 2.

Linear mixed effects model table of model evidence displaying the effects of face-looking on α values with the Social condition as the reference event.

	Model 1	Model 2	Model 3	Model 4	Model 5
Fixed effects
Between-person
Intercept	0.834*** (0.006)	0.835*** (0.006)	0.839*** (0.005)	0.850*** (0.003)	0.848*** (0.003)
Avg Age			− 0.001* (0.001)	− 0.001* (0.001)	− 0.001* (0.001)
Face-looking%, person Avg					0.117*** (0.035)
Within-person, between-visits
Age		0.003*** (0.000)	0.001** (0.000)	0.001** (0.000)	0.001** (0.000)
Face-looking (%), visit-Avg					0.087* (0.037)
Within-visit, between-movies
Pixelated (dummy)				− 0.022*** (0.004)	− 0.017*** (0.004)
Face-looking (%), movie Avg					0.029* (0.012)
Within-movie, between-segments
Face-looking (%), segment-Avg					− 0.027* (0.013)
Random effects
Between-person
Intercept	0.001	0.001	0.000	0.000	0.000
Age		0.000	0.000	0.000	0.000
Within-person, between-visits
Intercept	0.002	0.002	0.001	0.001	0.001
Pixelated				0.001	0.001
Within-visit, between-movie
Intercept	0.000	0.000	0.000	0.000	0.000
Residual	0.007	0.007	0.007	0.007	0.007
Model fit indices
AIC	− 13,944.2	− 13,998.9	− 14,197.6	− 14,229.7	− 14,248.3
BIC	− 13,910.0	− 13,944.1	− 14,115.4	− 14,126.9	− 14,118.1
Log likelihood	6977.1	7007.5	7110.8	7129.9	7143.2

P < 0.001.

Coefficients are presented with standard errors in parentheses, with significant effects bolded for emphasis. Model 5, which includes a Face-looking effect is the final best fitting model.