self-report-vs-performance

The gap between what a person says they can do and what they actually do when measured — a gap so consistent it is almost a law: the mouth reads warmer than the hand.

The finding recurs across domains. In metacomprehension, the self-read is the weakest instrument (kappa .08 for confusion self-assessment — see fog-meter). In emotion regulation, the DERS questionnaire gain may not predict the instructed-reappraisal performance gain (see ders-vs-performance). In therapy, EFT measures regulation by self-report, never by task (see eft-regulation-performance). In learning, reappraisal's grip on frustration fades as the impasse drags — felt regulation cools while performed regulation stalls (see frustration-as-fascination). The pattern: the feeling of competence or regulation runs ahead of the thing itself, and the gap is invisible until a performance measure holds the self-report against the light.

The law is not absolute — sometimes the self-report and the performance converge — but it is consistent enough that any study measuring only the questionnaire carries an untested assumption: that the felt change tracks the real one. The performance task is the missing half that would close the gap, and the cheapest design is always the same: add a 20-minute task alongside the questionnaire, and check whether the one predicts the other.