Tainted Truth: Overcorrection for Misinformation Influence on Eyewitness Memory

Existing research shows that the postevent misinformation effect on eyewitness memory can be successfully reduced by retrospective warnings. Exploring potential costs of such warnings, we investigated whether warned eyewitnesses might overcorrect for misinformation influence from the postevent source. Across three experiments, warned participants recognized relatively fewer event items that were truthfully described in a postevent narrative (“tainted” event items) than did unwarned participants. This tainted truth effect was obtained for peripheral but not for central event items (Experiments 1–3); it persisted when the warning took the form of an explicit source–monitoring instruction (Experiment 2) and when the accessibility of the postevent information was increased (Experiment 3). Overall, the effect was stronger with a socially framed warning than with more direct, explicitly phrased warnings. The findings suggest that the tainted truth effect is due to more careful monitoring for information from a suspicious postevent source rather than mere memory impairment.