Figure 1
Representation of Under and Overestimations.
Note. Top left panel: all the items of a domain known correctly (red dots) or incorrectly (black dots) by a person, distributed according to their level of knowledge. Top right panel: all the items sampled (green subset) through a questionnaire. Bottom left panel: the sample of items with their given confidence and their level of knowledge. Bottom right panel: the sample of items with, in yellow, those that are underestimated (confidence lower than level of knowledge) and, in blue, those that are overestimated (confidence higher the level of knowledge).
Table 1
Number of Correct and Incorrect Responses by Level of Confidence for a Respondent on a 60 2AFC Questionnaire.
| 50% | 60% | 70% | 80% | 90% | 100% | |
|---|---|---|---|---|---|---|
| Correct | 4 | 2 | 4 | 4 | 10 | 23 |
| Incorrect | 2 | 2 | 3 | 1 | 3 | 2 |
Figure 2
Two Different Representations of Observed and Estimated Data.
Note. The number of correct/incorrect responses as a function of confidence represented in 2 different ways. Bottom panels: the distribution of the LOK, superposed with the confidences given by the respondent on the left, and with the highlighting of under and overestimations on the right.
Figure 3
Percentages of Correct Responses as a Function of the Degree of Confidence with Standard Deviations.
Note. The red line shows what the percentage should be if the students were perfectly realistic.
Figure 4
Distribution of the Realism of the 41 Students.
Figure 5
Distribution of the Mean Amplitude of the Confidence Error in Absolute Value among the 41 Students.
Figure 6
Distribution (in Mean Number of Items per Student) of Confidence at the Second Trial for the Items with Confidence Degrees 50%, 60%, 70%, 80%, 90% and 100% at the First Trial.
Figure 7
Distribution (in Mean Number of Items per Student) of Confidence Given at the First Trial for the Items with Confidence Degree 50%, 60%, 70%, 80% at the Second Trial.
Table 2
Identification of the Best Model with the AIC.
| MODEL S | MODEL H | MODEL O | MODEL A | |
|---|---|---|---|---|
| Number of students | 16 | 12 | 4 | 9 |
Figure 8
Conditional Distribution of Confidences at One Trial Given Confidence Levels of 50%, 60%, 70%, and 80% at the Other Trial.
Note. Left panel: distribution (in mean number of items per student) of confidence at the second trial for the items with confidence degrees 50%, 60%, 70%, 80% at the first trial, for students significantly attracted by the confidence level 100% according to model A. Right panel: the same distributions when the two trials are permuted.