Self-monitoring is a metacognitive activity that individuals control and alter their behavior based on the assessments of their own cognitive status, which involves both explicit and implicit dimensions. Confidence rating task is mostly used to measure explicit monitoring. Recently post-decision wagering (PDW) task has been developed to quantify implicit monitoring ability. Specifically, temporal wagering task estimates the decision confidence by measuring the waiting time subjects are willing to spend in waiting delayed reward, which is called wagering time (WT). The index has only been applied in animals experiment so far, but it values in deeply exploring the underlying mechanism of metacognitive monitoring. Thus in this experiment, WT is tested whether it fits on measuring human’s implicit monitoring ability. Results show that there are significantly positive correlation between WT and confidence. Moreover, there are several consistent effects showing metacognitive features in both WT and confidence—being able to measure metacognitive sensitivity and metacognitive efficiency; predicting the decision accuracy as well as serial dependence effect. All results have illustrated that WT indeed can be a good proxy of confidence and therefore provide an effective measure to study metacognitive monitoring on human. It can also help instruct the clinical treatments of psychiatric patients with deficits in self-monitoring.

We apply a within-subject design in which both temporal wagering task and confidence rating task are used. 3600 trials were colleccted from 30 subjects (average age = 22.1, sd = 0.84). Half of the trials are temporal wagering data. 

Contributors
Duan Wei (dw6805@126.com)

Stimulus
Pairs of images extracted from video narrative, being different only in resolution. 

Confidence scale
6-point scale:  "Extremely Low", 'Very Low", "Low", "High", "Very High", "Extremely High"

Response type
Correct and high confidence: response type 1; correct and low confidence: response type 2; incorrect and low confidence: response type 3; incorrect and high confience: response type 4. Although catch trials are correct and high confidence, but we denote it as response type 5 for specific analyses. 

Level
6 levels. Level 1 is the hardest one and 6 is the easiest one. 

Manipulations
Participants performed either temporal wagering task or confidence rating task in a block.

Block size
60. 

Feedback
Trial-by-trial feedback was provided.

Subject population
Young adults (20-24 years old).

Response device
Computer keyboard.

Experiment setting
online; behavioral

Training
No.

Experiment goal
Examine the similarity of patterns shown by WT and confidence.

Main result
WT showed similar pattern to confidence.

Experiment dates
Data were collected between January, 2020 and March, 2020. 

Location of data collection
Because of the quaranteen, subjects did the experiment at home after downloading Psychopy and experiment programme.

Language of data collection
Chinese.