Page 112 - Demo

Chapter 4110and did not vary by emotion category (i.e., no interaction). Thus, independent of the displayed expression, confidence judgments were significantly lower with higher social anxiety trait levels (for a detailed description of the model fit, see Fig. 3A and Table S3 in the Supplemental Material).Autistic traits. In the LMM that included autistic traits instead of social anxiety traits to predict confidence in emotion recognition, the effect of emotion category was significant, F(5, 3285) = 118.164, p < .001, as was the interaction between emotion category and autistic traits, F(5, 3285) = 9.531, p < .001. While for neutral and happy facial expressions confidence was significantly reduced with higher autistic traits compared to the average effect of autistic traits on confidence ratings, %u03b2 = -0.148, z = -4.663, p <.001, and %u03b2 = -0.104, t(3285) = -3.271, p = .001 respectively, this effect was reversed for displays of fear and sadness. For those two categories, autistic traits were associated with higher confidence ratings compared to the average effect of autistic traits on confidence, fear: %u03b2 = 0.102, t(3285) = 3.229, p = .001; sadness: %u03b2 = 0.118, t(3285) = 3.715, p < .001 (see Fig. 3B and Table S4 in the Supplemental Material). Link between Confidence and Emotion RecognitionSocial anxiety and autistic traits. When exploring the link between confidence and accuracy in emotion recognition in relation to the clinical trait dimensions on a trial level and by emotion category, neither of the trait dimension had a substantial impact on this link (see Table S5 and S6 in the Supplemental Material for the entire model fits).Metacognitive SensitivityAccording to Mahalanobis distance measures, two participants had to be excluded from the correlation analysis between social anxiety traits and AUROC2, and three participants had to be excluded from the correlation analysis between autistic traits and AUROC2. After excluding these bivariate outliers, all distributions did not majorly deviate from normality. The two correlational analyses between the clinical trait dimension and metacognitive sensitivity (AUROC2) revealed that autistic traits, but not social anxiety traits, were significantly related to metacognitive sensitivity (see Fig. 4). As expected, metacognitive sensitivity was reduced with higher autistic traits, r = - .489, t(51) = -4.008, p < .001 and rs = - .476, p < .001. and no significant relation was found for social anxiety, r = - .222, t(53) = -1.66, p = .103 and rs = - .251, p = .064.