Page 44 - Demo

42Chapter 2Figure 7. By-Subject Mean-Centered RTs per Level of Gaze Congruency andAttractiveness Category. Dots indicate the median reaction times (RT), whileerror bars represent the 89% credible interval. On both congruent andincongruent trials, we found no evidence for attractiveness resulting in a strongergaze cuing effect.between Sex, Attractiveness Category and Gaze Congruency: the simplemodel performed slightly better than the complex model (∆elpdLOO = 3.5[2.2]). Altogether, these findings s uggest t hat a dding A ge o r S ex t o thesimple model did not increase the simple model’s predictive accuracy.Table 3. Model output for the simple model of Experiment 2. All categorical independent variables were sum-to-zero coded.Parameter Median estimate SD 89% CIsIntercept 0.16 1.42 -2.06, 2.48Attractiveness Category [attractive] -1.09 0.95 -2.58, 0.46Attractiveness Category [intermediate] 0.06 0.95 -1.48, 1.57Gaze Congruency [Incongruent] 16.08 0.67 15.00, 17.16Attractiveness Category [attractive]:Gaze Congruency [Incongruent] -1.10 0.95 -2.59, 0.44Attractiveness Category [intermediate]:Gaze Congruency [incongruent] -0.58 0.95 -2.07, 0.96Random Effectssd [intercept] Trial order 9.63 1.18 7.90, 11.67sd [intercept] Subject 0.47 0.42 0.05, 1.33Nobs = 8425Nsubj = 150Note: All categorical independent variables were sum-to-zero coded.DiscussionAttractiveness is a salient social signal that affects not only our judgements,but also biases our attention and perception of other social information. Inthe current study, we investigated how facial attractiveness and symmetrymodulated attention. Moreover, we investigated whether facial attractiveness modulated gaze cueing. The results show, first, that participants hadan attentional bias towards attractive faces, but not towards unattractivefaces. Second, attention was not differentially modulated by facial symmetry. Third, gaze cueing was not affected by the attractiveness of the face.Fourth, we found no evidence for differences in attractiveness bias betweenmen and women, or between younger and older participants. These resultswill be discussed in more detail in the sections below.Our first key result, that people had an attentional bias towards attractive faces, is in line with previous research (Y. Ma et al., 2015, 2019; Maner,Gailliot, Rouby, & Miller, 2007). Using a similar dot-probe study as in thecurrent study, (Y. Ma et al., 2015, 2019) showed that Chinese undergraduate students (n = 108 females: Ma et al., 2015a; n = 109 males: Ma et al.,2019) had difficulties disengaging from attractive faces. While they foundno overall attentional bias towards attractiveness faces, only participantswho were single and primed with romantic words showed this effect. Thecurrent study builds on this work and extends it in several ways. First, wedid not only include the comparison between attractive and intermediatelyattractive faces, but also included the comparison between unattractive andintermediately attractive faces. Consequently, we can conclude that participants selectively attended to attractive, but not unattractive faces. Thisfinding suggests that the attentional bias towards attractive faces is notmerely the result of attractive faces deviating from the average face, as thisis the case for unattractive faces as well. Second, using a large communitysample with a wide age range, we were able to show that attractiveness alsoinfluences attention in Western people, regardless of their age or gender.Third, we limited the stimulus presentation duration to 300 ms to make itunlikely that participants shifted gaze once their attention had been captured by one of the two presented images (Petrova et al., 2013). Longerpresentation durations allow such oculomotor shifts to occur; however, theyare not recorded and thus yield noisier data (van Rooijen et al., 2017).Therefore, our results are likely to represent an attentional capture effect,while the previous studies mainly found disengagement effects. Thus, with afew methodological adjustments and a more heterogeneous sample, we wereable to show that attention to attractive faces is likely a more general effectthan previously assumed.Our second key result, namely that facial symmetry does not affect implicit attention, was against our expectations. If facial symmetry would bean important signal reflecting mate quality, one would expect symmetricalFigure 7. By-Subject Mean-Centered RTs per Level of Gaze Congruencyand Attractiveness Category. Dots indicate the median reaction times (RT),while error bars represent the 89% Credible Interval. On both congruentand incongruent trials, we found no evidence for attractiveness resulting ina stronger gaze cuing effect.between Sex, Attractiveness Category and Gaze Congruency: the simplemodel performed slightly better than the complex model (∆elpdLOO = 3.5[2.2]). Altogether, these findings suggest that adding Age or Sex to thesimple model did not increase the simple model’s predictive accuracy.Table 3. Model output for the simple model of Experiment 2.Parameter Median estimate SD 89% CIsIntercept 0.16 1.42 -2.06, 2.48Attractiveness Category [attractive] -1.09 0.95 -2.58, 0.46Attractiveness Category [intermediate] 0.06 0.95 -1.48, 1.57Gaze Congruency [Incongruent] 16.08 0.67 15.00, 17.16Attractiveness Category [attractive]:Gaze Congruency [Incongruent] -1.10 0.95 -2.59, 0.44Attractiveness Category [intermediate]:Gaze Congruency [incongruent] -0.58 0.95 -2.07, 0.96Random Effectssd [intercept] Trial order 9.63 1.18 7.90, 11.67sd [intercept] Subject 0.47 0.42 0.05, 1.33Nobs = 8425Nsubj = 150Note: All categorical independent variables were sum-to-zero coded.Figure 7. By-Subject Mean-Centered RTs per Level of Gaze Congruencyand Attractiveness Category. Dots indicate the median reaction times (RT),while error bars represent the 89% Credible Interval. On both congruentand incongruent trials, we found no evidence for attractiveness resulting ina stronger gaze cuing effect.between Sex, Attractiveness Category and Gaze Congruency: the simplemodel performed slightly better than the complex model (∆elpdLOO = 3.5[2.2]). Altogether, these findings suggest that adding Age or Sex to thesimple model did not increase the simple model’s predictive accuracy.Table 3. Model output for the simple model of Experiment 2.Parameter Median estimate SD 89% CIsIntercept 0.16 1.42 -2.06, 2.48Attractiveness Category [attractive] -1.09 0.95 -2.58, 0.46Attractiveness Category [intermediate] 0.06 0.95 -1.48, 1.57Gaze Congruency [Incongruent] 16.08 0.67 15.00, 17.16Attractiveness Category [attractive]:Gaze Congruency [Incongruent] -1.10 0.95 -2.59, 0.44Attractiveness Category [intermediate]:Gaze Congruency [incongruent] -0.58 0.95 -2.07, 0.96Random Effectssd [intercept] Trial order 9.63 1.18 7.90, 11.67sd [intercept] Subject 0.47 0.42 0.05, 1.33Nobs = 8425Nsubj = 150Note: All categorical independent variables were sum-to-zero coded.Iliana Samara 17x24.indd 42 08-04-2024 16:35