Table 2.

Results from binomial generalized linear mixed effects regression model predicting voice choice.

	$\widehat{\beta }$	|Z|	p
Intercept	−5.69	6.61	< 0.001	***
Agent humanness (=non-human)	0.19	0.24	0.81
Patient humanness (=non-human)	−7.92	5.88	< 0.001	***
Agent humanness × patient humanness	−6.04	3.57	< 0.001	***
First-fixated character (=patient)	0.73	1.54	0.12

^***p < 0.001.