The post-identification feedback effect. The lineup used in this exercise was a target-absent parade: the actual perpetrator was not present. This means that whatever position you chose, the person you selected was innocent. Yet if you received confirming feedback (“Well done — you selected the perpetrator”), your confidence after that feedback was likely higher than it was immediately after making your choice.

This is the post-identification feedback effect, first demonstrated by @wells1998. Confirming feedback inflates not only confidence but retrospective reports of how good a view the witness had, how much attention they paid, how certain they felt at the moment of identification, and how easy they found the task. None of these inflated reports reflects any real difference in witnessing conditions.

The mechanism is not deliberate fabrication. Witnesses are reconstructing their past experience, and feedback provides information that appears relevant to that reconstruction. When the original memory is weak, the influence of external cues is strongest. This is precisely when identifications are least likely to be accurate. The result is that confidence expressed after investigative contact — months later at trial — can be substantially higher than confidence expressed at the moment of identification, without any genuine improvement in memory. This is why researchers and reform advocates treat the first immediate confidence statement as the only reliable evidential record of a witness’s certainty.