Table 2.
Results of the linear mixed effects analysis for emotion recognition performance testing Model 5 against Model 1.
| Predictors | Estimates | p | df |
|---|---|---|---|
| (Intercept) | 93.20 *** | <0.001 | 3,514.00 |
| Neutral | Reference | ||
| Anger | −4.76 | 0.094 | 3,514.00 |
| Disgust | −1.70 | 0.549 | 3,514.00 |
| Fear | −1.70 | 0.549 | 3,514.00 |
| Happiness | 6.46 * | 0.023 | 3,514.00 |
| Sadness | −22.79 *** | <0.001 | 3,514.00 |
| exprEmo_anger:Mask | −17.35 *** | <0.001 | 3,514.00 |
| exprEmo_disgust:Mask | −52.72 *** | <0.001 | 3,514.00 |
| exprEmo_fear:Mask | −0.00 | 1.000 | 3,514.00 |
| exprEmo_happiness:Mask | −26.53 *** | <0.001 | 3,514.00 |
| exprEmo_sadness:Mask | −8.40 * | 0.037 | 3,514.00 |
| No mask | Reference | ||
| Mask | 1.70 | 0.549 | 3,514.00 |
| ICC | 0.05 | ||
| N depictPers | 6 | ||
| N CaseID | 49 | ||
| Observations | 3,529 | ||
| Marginal R2/Conditional R2 | 0.179/0.224 | ||
| AIC | 3,5084.229 | ||
| Log-likelihood | −17,527.114 | ||
>The table shows the statistics for all involved fixed effects in the linear mixed effects analysis for Model 5, regarding the tested dependent variable % correct (percentage of correct emotion classifications). Abbreviated notations for the terms were used to save space: exprEmo_XY = facial emotion, e.g., anger; Mask = face with face mask. Significant values of p are in bold. *p < 0.05 and ***p < 0.001.