Accuracy of confidence judgments: stability and generality in the learning process and predictive validity for...

Area of Science:

• Cognitive Psychology
• Educational Psychology

Background:

• Metacognition involves monitoring memory and comprehension.
• Retrospective confidence judgments (RCJs) are a key measure of metacognitive monitoring.
• Understanding the stability and predictive power of RCJ accuracy measures is crucial for educational assessment.

Purpose of the Study:

• To assess the stability of different RCJ accuracy measures over learning time.
• To determine if RCJ accuracy generalizes across different tests.
• To investigate the predictive validity of RCJ accuracy for learning outcomes.

Main Methods:

• University students (N=113) learned operant conditioning concepts for 30 minutes.
• Knowledge tests with post-item confidence judgments were administered at three time points: pre-learning, 10 minutes into learning, and post-learning.
• Absolute measures (bias, absolute bias) and relative measures (gamma, Pearson's r, d(a)) of RCJ accuracy were calculated.

Main Results:

• Absolute measures of RCJ accuracy demonstrated stability over the learning period.
• Relative measures of RCJ accuracy were not stable over time.
• Absolute bias, gamma, and Pearson's r at the 10-minute mark predicted the final learning outcome.

Conclusions:

• Absolute measures of confidence judgment accuracy are reliable indicators of metacognitive monitoring during learning.
• Certain RCJ accuracy measures, particularly when assessed early in the learning process, can predict subsequent learning success.
• Findings contribute to understanding the dynamics of metacognitive monitoring and its relationship to learning performance.