Page 123 - Emotions through the eyes of our closest living relatives- Exploring attentional and behavioral mechanisms

that contagious scratch rates did not differ between tense and relaxed context (bootstrapped paired samples t test: p = .795), suggesting that the observed effect of context is not just a by-product of increased self-scratching due to increased stress levels during tension.
Table 1. Type III tests for fixed effects on the occurrence of scratch contagion
Self-scratch contagion observed in orangutans
Intercept
Context (tense)
Relationship quality (low)
Seen/unseen (seen)
Context * relationship quality (tense * low)
Context * seen/unseen (tense * seen)
Relationship quality * seen/unseen (low * seen)
Context * relationship quality * seen/unseen (tense * low * seen)
Estimate SE c21
−1.897 0.38 24.864 0.088 0.457 0.038 −0.228 0.428 0.283 0.24 0.418 0.33 −0.576 0.725 0.631 −0.653 0.956 0.466 0.384 0.675 0.324 2.869 1.334 4.627
p
<.001
0.846 0.595 0.566 0.427 0.495 0.569 0.032
Note: GLMMs were used with a binomial distribution and logit link function. Effects with p < .05 are depicted in 5 italics. Abbreviations: GLMMs, generalized linear mixed models; SE, standard error.
Figure 2. Predicted probability of scratch contagion (±SEM) based on the three-way interaction between seeing the triggering scratch, context and relationship quality. SEM, standard error of mean. *p < .05; ***p < .001
121