Page 72 - Emotions through the eyes of our closest living relatives- Exploring attentional and behavioral mechanisms

Chapter 3
age bin were excluded. The reason for this procedure is that very young children and older people tend to be a lot slower on this task than those with an average age. After applying these criteria, 10.14% of the data was excluded. The data were modelled using a Gamma distribution (log link function), given the typical skewness in RT data.
The statistical procedure was similar to the analysis of the rating scales, unless stated otherwise. In the model for the dot-probe task, two additional factors were included, a) Congruency (probe position replacing the emotional stimulus or not) and b) Distance of dominant hand to the screen (Short/Long). This latter factor is added as a control to account for the shorter distance for a left-handed person from the left hand to the left side of the screen and for a right-handed person using the right hand to the right of the screen compared to left to right and vice versa.
Results
Perceived valence of human and bonobo scenes
In a first analysis where the emotion categories Sex and Aggression were excluded, participants (adults and children), showed a main effect of Emotion Category (F(1, 11.36) = 124.86, p < .001). Play was most positively rated out of all categories, followed by groom, neutral, yawn, distress and self-scratch (the latter being least positively rated of all). All emotion categories differed from the neutral scenes in the expected direction (ps ≤ .001). A main effect of Species showed that in general, bonobo scenes were perceived more positively than human scenes (F(1, 11.36) = 11.83, p = .001).
An interaction between Emotion Category and Species (F(5, 11.36) = 87.76, p < .001) showed that human scenes received more extreme ratings than bonobo scenes. That is, positive emotions received more positive ratings and negative ones more negative ratings when expressed by humans rather than bonobos. This was significant for the categories Play (t(12.54) = 6.37, p < .001) and Distress (t(11.36) = 20.63, p < .001), with a trend toward significance in the same direction for self-scratching (p = .031). In contrast, another trend was observed for yawning behavior, which was rated somewhat more negatively in the bonobo compared to human scenes, possibly because of the visibility of the canines (p = .024). There were no differences for neutral (p = .922) or grooming scenes (p = .551).
Further, an interaction between Emotion Category and Age (F(5, 11.36) = 7.81, p < .001) showed that compared to adults, children differentiated the categories less based on valence. Specifically, they gave less negative evaluations of the negative
70