The present experiments focus on whether the post-identification feedback effect can be reduced by providing participants with warnings. Participants viewed a crime on video and identified a suspect from a target-absent lineup (Experiment 1) or target-present lineup (Experiment 2). Participants then received positive feedback, negative feedback or no feedback. Half of the participants received a warning saying their feedback was randomly generated by the computer, and the other half received no warning. Robust post-identification feedback effects were observed in both experiments in the no warning condition. These effects were largely eliminated when participants received a warning. In Experiments 3 and 4, we failed to find an ameliorative effect of a forensically realistic warning. These results indicate that warnings can reduce the effect of post-identification feedback in principle, but the application of warnings in practice may be more difficult. PUBLICATION ABSTRACT