When not remembering is memorable: Personal distinctiveness explains a strength-based misprediction in recognition with repeated tests.

We investigated how initial recognition confidence predicts future recognition success and certainty using a novel double test task in which all items from an initial test (Test 1) are retested during an immediate second test (Test 2). Equal variance signal detection and MINERVA 2 simulations predicted that Test 1 recognition strength, and hence confidence, should positively predict Test 2 recognition success and confidence. This occurred for Test 1 hits but not for Test 1 correct rejections across four experiments. For correct rejections, increasing confidence in novelty (viz., reduced strength) yielded increased Test 2 recognition and certainty, the opposite of the strength-based predictions. Experiment 3 tested the hypothesis that subjective memorability assessment was the cause of the strength-based misprediction by examining the language used to justify recognized items during Test 2 that were previously Test 1 correct rejections. For previously high- versus low-confidence correct rejections, subjects were more likely during Test 2 to report explicitly remembering the prior testing experience, and that the items were rejected, because, despite their subjectively distinctive nature, they failed to evoke remembrance of the initial study. Experiment 4 further examined the subjective memorability hypothesis with a novel version of the remember/know paradigm requiring subjects to rate subjective memorability following Test 1 new responses. Items rated as subjectively memorable were both more confidently rejected at Test 1 and later more easily recognized at Test 2 than those rated as not subjectively memorable, again contradicting strength-based predictions. We discuss the implications of these findings for models of recognition. (PsycInfo Database Record (c) 2026 APA, all rights reserved).