Page 119 - Demo

117Can third-party observers detect attraction?6tuitive. Therefore, we report multiple estimates to illustrate the robustnessand uncertainty of an effect (e.g., see Martin et al., 2020). The median estimate coefficient is reported together with the 95% Highest-Density CredibleIntervals (HDI), which summarize the posterior parameter values with thehighest probability density (Kruschke, 2018). Furthermore, we report theprobability of direction (pd), the proportion of the probability in supportof a hypothesized positive or negative effect (Makowski et al., 2019). Toexamine the robustness of interactions, we performed model comparisons tocalculate Bayes Factors (BF). BFs are interpreted following the scheme ofJeffreys (1961), who suggested BF values of 0-3 to be considered anecdotalevidence, 3-10 moderate, and greater than ten strong evidence in favour ofeither the alternative (BF10) or null (BF01) hypothesis.Model convergence was examined using the guidelines detailed in theWAMBS checklist (Depaoli & van de Schoot, 2017). Specifically, for everymodel, we examined the Gelman-Rubin diagnostic values (a value close to 1indicates convergence), as well as trace, autocorrelation plots, and densityhistograms for all posterior distributions. Analyses were conducted in R(version 3.6.1; R Core Team, 2019) using the brms package (Bürkner, 2017,2018).Experiment 1MethodsParticipantsSixty-one adults (n = 61; age range = 18-54; M age: 26.13, SD = 6.40; 42female) and 60 children (2 excluded for inattentiveness, final n = 58, agerange: 8-14; M age: 10.00, SD = 1.63; 25 female) were recruited during ascience festival (Rotterdam, the Netherlands). The sample size was determined by the number of people that wanted to participate during this eventand is comparable to the studies by Place et al. (2009). All participants provided informed consent and were informed that they could withdraw theirparticipation with no adverse consequences as according to the Declarationof Helsinki. For children younger than 12 year old, consent was providedby their parents, whereas for children older than 12 years old, consent wasprovided by both the parents and the children. The study was approvedby the Leiden University Psychology Ethics Committee (CEP19-0424/290).Participants were not remunerated for their participation.StimuliStimuli consisted of videos of the couple members during the first impression(FI) presented side by side on the display for 3 s (see Figure 3). The 3 s videoswere selected as in the original study, participants could report within 3 sIliana Samara 17x24.indd 117 08-04-2024 16:36