Confidence-accuracy gap tracking

After each test, compute and record the gap between your predicted and actual scores.

Why it works

The feeling of knowing is a continuous signal, not a binary; it can be calibrated. By repeatedly comparing predicted performance (pre-test confidence) to actual performance (post-test score), learners receive direct feedback on how much their FOK inflates reality. The gap, tracked over sessions, reveals whether calibration is improving and which topics generate the largest FOK illusions.

How to do it

1. Before each quiz or retrieval session, predict your score as a percentage.
2. Complete the test, score it honestly.
3. Record predicted score, actual score, and the gap in a learning log.
4. After three to five sessions, review: which topics show the largest systematic over-prediction?

Evidence

Calibration research in educational psychology consistently shows that learners are overconfident on average, and that explicit calibration feedback reduces overconfidence and improves study-time allocation. (observational)

Calibration improvement from feedback is real but gradual; it does not eliminate overconfidence, and the benefits for actual performance (vs. calibration accuracy) are more variable.

Sources

• Dunlosky & Rawson (2012), "Overconfidence produces underachievement," Perspectives on Psychological Science

Common mistake

Predicting performance after seeing the material rather than before testing, which contaminates the pre-test confidence with cue exposure and prevents honest FOK calibration.