Page 125 - Demo

123Can third-party observers detect attraction?6Figure 5. (a) Accuracy as a function of Group (Children vs. Adults).The graph depicts that both children and adults performed at chancel level(0.5); (b) Accuracy as a function of Attraction to Partner (Attracted vs.Not attracted). The graph depicts that participants performed above chancelevel (0.5) when the person depicted was attracted to their partner comparedto when they were not. The red line denotes the chance level, and all errorbars reflect 95% Credible Intervals (CrI).The model showed that accuracy did not substantially differ as a function ofGroup (β = -0.05; 95% HDI [-0.14, 0.04]; p− = 85.70%). Next, we modeledparticipants’ accuracy by including the fixed effect of Attraction to Partnerand its interaction with Age Group. As in Experiment 1, the model showedthat participants were more accurate when the person in the video indeedwas attracted to their partner than not (β = 0.25; 95% HDI [0.16, 0.34];p+ = 100%; see Figure 5; Table 2, Model 3). Accuracy did not differ as afunction of Age Group (β = -0.05; 95% HDI [-0.14, 0.04]; p− = 86.25%).The interaction between Age Group and Attraction was not reliable (β =0.09; 95% HDI [0.00, 0.18]; p+ = 97.09%; BF01 = 3.45). For that reason,the interaction is not interpreted.Discussion Experiment 2The goal of Experiment 2 was to assess whether the low accuracy observedin Experiment 1 was the result of the simultaneous video stream used inExperiment 1. Our results are straightforward. First, we found no differencebetween children’s and adults’ accuracy. Further, we replicate the findingIliana Samara 17x24.indd 123 08-04-2024 16:36