Brady Butterfield and Janet Metcalfe documented the effect in a 2001 Journal of Experimental Psychology paper. Contrary to the intuitive expectation that confident errors would be the stubbornest to correct, they proved the opposite. High-confidence corrections stick hardest.
The proposed mechanism: the surprise of being wrong while certain triggers amygdala activation, which modulates hippocampal encoding. The correction arrives with emotional weight that passive learning lacks.
Fluera’s confidence slider (rate 1–5 before you see the answer) is not UI polish. It primes the contrast. A confidence-5 wrong answer revealed by Ghost Map leaves a far deeper trace than any number of passive reviews. We ask for your confidence precisely because being wrong about it is the point.