Ng sad emotions (better than the older participants), while they were Avermectin B1a site performing worse for disgust and surprise. Bonferroni corrected (15 comparisons, critical p-value = 0.00333) post-hoc analyses on HIV-1 integrase inhibitor 2 manufacturer subsets of the data (e.g., only happy and sad stimuli) using the same Model 2 for the analysis, showed significant differences between Anger and Disgust, Anger and Surprise, Happy and Disgust, Happy and Surprise, Fear and Disgust, Fear and Surprise, Sad and Disgust, and Sad and Surprised (all p’s <0.001). The main effect of emotion for the Fear-Disgust comparison was accompanied by a significant interaction with age of the viewer (p = 0.0029), reflecting the particular high accuracy in children in the Sad condition and low accuracy for Disgust stimuli. Comparison of younger and older adults showed a significant effect of emotion (p < 0.001) and an interaction between emotion and age-group (p < 0.001). Bonferroni corrected comparisons between younger and older adults for each of the emotions showed that older viewers were significantly worse for Fear and Sad stimuli (p < 0.001). A comparison between young adults and children showed an effect of emotion, age-group, but no emotion by age-group interaction, reflecting the overall worse performance for children, and differences in the emotions overall. A comparison between older adults and children showed a significant emotion by age-group interaction, but no main effect of emotion. Interestingly, this interaction appears to be driven by differences for the Fear emotion alone (p = 0.0017), when gender of the viewer, and the interaction between age of the viewer and age of the actor are taken into account. Contact Potential associations between contact and accuracy in body expression categorization were first explored with correlations between the estimated number of hours per month (HPM) spent with children, younger adults and older adults and the average percentage correct responses (collapsed over expression) to body expressions posed by children, young adults, and older adults. These analyses revealed only one borderline significant positive linear relationship: for child viewers only, the amount of contact with older adults correlated positively with accuracy of children for clips of older actors (Pearson's r = 0.23, p = 0.05). All remaining correlations in all three age groups were not significant. Moderated LinearPollux et al. (2016), PeerJ, DOI 10.7717/peerj.11/Figure 2 Percentage correct responses for viewers (Young viewers, Older viewers, Child Viewers) as a function of body expression (Anger, Happy, Fear, Sad, Disgust and Surprise).Regression (Hayes, 2013) was subsequently used to explore if the age-group of the viewer is a significant moderating factor for the relationship between contact with older adults and percentage correct categorizations of body expressions posed by older adults. Age-group of the viewer was coded as a binary variable, contrasting child viewers with adult viewers. Age-group Viewer and children's contact with older adults (estimated average hours of contact per month (HPM-older adults)) were entered as predictors for percentage correct responses to expression posed by older adults in the first step of the hierarchical model (Model 1). After centring the means to reduce collinearity, the interaction term (Age-group Viewer ?HPM-older adults) was added in the second step (Model 2). Model 1 accounted for a significant amount of variance in children's ability to recog.Ng sad emotions (better than the older participants), while they were performing worse for disgust and surprise. Bonferroni corrected (15 comparisons, critical p-value = 0.00333) post-hoc analyses on subsets of the data (e.g., only happy and sad stimuli) using the same Model 2 for the analysis, showed significant differences between Anger and Disgust, Anger and Surprise, Happy and Disgust, Happy and Surprise, Fear and Disgust, Fear and Surprise, Sad and Disgust, and Sad and Surprised (all p's <0.001). The main effect of emotion for the Fear-Disgust comparison was accompanied by a significant interaction with age of the viewer (p = 0.0029), reflecting the particular high accuracy in children in the Sad condition and low accuracy for Disgust stimuli. Comparison of younger and older adults showed a significant effect of emotion (p < 0.001) and an interaction between emotion and age-group (p < 0.001). Bonferroni corrected comparisons between younger and older adults for each of the emotions showed that older viewers were significantly worse for Fear and Sad stimuli (p < 0.001). A comparison between young adults and children showed an effect of emotion, age-group, but no emotion by age-group interaction, reflecting the overall worse performance for children, and differences in the emotions overall. A comparison between older adults and children showed a significant emotion by age-group interaction, but no main effect of emotion. Interestingly, this interaction appears to be driven by differences for the Fear emotion alone (p = 0.0017), when gender of the viewer, and the interaction between age of the viewer and age of the actor are taken into account. Contact Potential associations between contact and accuracy in body expression categorization were first explored with correlations between the estimated number of hours per month (HPM) spent with children, younger adults and older adults and the average percentage correct responses (collapsed over expression) to body expressions posed by children, young adults, and older adults. These analyses revealed only one borderline significant positive linear relationship: for child viewers only, the amount of contact with older adults correlated positively with accuracy of children for clips of older actors (Pearson's r = 0.23, p = 0.05). All remaining correlations in all three age groups were not significant. Moderated LinearPollux et al. (2016), PeerJ, DOI 10.7717/peerj.11/Figure 2 Percentage correct responses for viewers (Young viewers, Older viewers, Child Viewers) as a function of body expression (Anger, Happy, Fear, Sad, Disgust and Surprise).Regression (Hayes, 2013) was subsequently used to explore if the age-group of the viewer is a significant moderating factor for the relationship between contact with older adults and percentage correct categorizations of body expressions posed by older adults. Age-group of the viewer was coded as a binary variable, contrasting child viewers with adult viewers. Age-group Viewer and children's contact with older adults (estimated average hours of contact per month (HPM-older adults)) were entered as predictors for percentage correct responses to expression posed by older adults in the first step of the hierarchical model (Model 1). After centring the means to reduce collinearity, the interaction term (Age-group Viewer ?HPM-older adults) was added in the second step (Model 2). Model 1 accounted for a significant amount of variance in children's ability to recog.