Table 1 Measures of metacognition examined in the current paper
From: A comprehensive assessment of current methods for measuring metacognition
Measure | Calculation | Based on a process model |
|---|---|---|
meta-d' | d’ value that provides best fit to Type 2 ROC | No |
AUC2 | Area under the Type 2 ROC curve | No |
Gamma | Rank correlation between confidence and accuracy | No |
Phi | Pearson correlation between confidence and accuracy | No |
ΔConf | Difference between average confidence for correct and error trials | No |
M-Ratio | meta-d’ divided by d' | No |
AUC2-Ratio | AUC2 divided by expected AUC2 under SDT assumptions | No |
Gamma-Ratio | Gamma divided by expected Gamma under SDT assumptions | No |
Phi-Ratio | Phi divided by expected Phi under SDT assumptions | No |
ΔConf-Ratio | ΔConf divided by expected ΔConf under SDT assumptions | No |
M-Diff | meta-d’ minus d' | No |
AUC2-Diff | AUC2 minus expected AUC2 under SDT assumptions | No |
Gamma-Diff | Gamma minus expected Gamma under SDT assumptions | No |
Phi-Diff | Phi minus expected Phi under SDT assumptions | No |
ΔConf-Diff | ΔConf minus expected ΔConf under SDT assumptions | No |
meta-noise | Metacognitive noise computed using the lognormal meta noise model | Yes |
meta-uncertainty | Metacognitive uncertainty computed using the CASANDRE model | Yes |
