Page 101 - Demo

99Underlying factors of the sexual overperception bias 5were exponentiated to obtain the odds ratio (OR). Effects were interpretedusing the OR 95% Highest Density Intervals (HDI), which summarize 95%of the posterior parameter distribution (Kruschke, 2018). If the 95% HDIspanned over 1, then the effect was not considered robust; given that thiswould suggest that accuracy spanned over 0.5 (i.e., chance level accuracy).To examine the reliability of interactions, we performed model comparisonsto calculate Bayes Factors (BF). As recommended, more than the default1000 iterations per chain (1500) were set to allow for the efficient calculationof BFs. To test differences in interactions, we calculated a BF using theSavage-Dickey method (see Wagenmakers, Lodewyckx, Kuriyal, & Grasman,2010).To further examine the direction of the errors associated with detectingattraction, we calculated a parameter estimation by subtracting the participants’ decision from their partners’ decision (see also Perilloux et al., 2012).This led to a parameter that took the values of 0 if the participants wereaccurate, 1 if they overestimated attraction, and -1 if they underestimatedattraction. We then modeled this variable as a function of Sex and OwnInterest (i.e., whether the participant was interested in going on anotherdate with his or her partner) and their interaction in an ordinal model. Weopted for adjacent category models (ACM) with category-specific effects,which allowed us to detect differences between each category level (e.g.,man vs. woman) for each of the potential outcomes. We set a prior ofStudent’s t with 7 degrees of freedom, scaled at 0 and with an SD of 2.For all models, we followed the procedure outlined in the WAMBS checklist (Depaoli & van de Schoot, 2017). Trace and autocorrelation plots, aswell as posterior density histograms were examined. All analyses were conducted in R Studio (version 3.6.2; R Core Team, 2019) using the brmspackage (Bürkner, 2017, 2018; Bürkner & Vuorre, 2019).ResultsA Bayesian chi-square test showed that men indicated more often thanwomen that they were interested in going on another date with their dating partner (BF10 > 10; see Table 1), consistent with previous findings.Bayesian independent t-tests showed that there was no difference betweenmen and women in sexual desire (women: M = 50.71, SD = 12.19; men: M= 56.48, SD = 15.76; BF10 = 0.83) and self-rated attractiveness (women:M = 4.68, SD = 0.73; men: M = 5.03, SD = 0.68; BF10 = 1.46), contraryto previous findings (A. J. Lee et al., 2020; Perilloux et al., 2012).AccuracyIn the first model, we examined whether Sex and Own Interest influencedattraction detection accuracy (Table 2; Model 1). The results showed thatIliana Samara 17x24.indd 99 08-04-2024 16:36