An Erratum has been published for this article in Applied Cognitive Psychology 18 (3) 2004, 371.

Three experiments examined whether changes in stimuli features would affect the pattern of memory biases reported by Ste-Marie and colleagues (1991, 1996, 2001). Experiment 1 served as a replication of Ste-Marie and colleagues' findings with new stimuli. In Experiments 2 and 3, surface features of the gymnastic stimuli were altered from that presented in the study phase for both perceptual (implicit memory test) and recognition test (explicit memory test) phases. In Experi-ment 2, the bodysuit that was worn by the gymnast was changed between the study phase and test phase presentation, whereas it was the gymnast who performed the gymnastic element that was changed in Experiment 3. Memory biases were still evident, despite the stimulus feature change of the bodysuit. In contrast, memory biases were significantly reduced when there was a change in the gymnast performing the element. Discussion is focused on the activation of memory representations underlying these effects and the nature of task relevance for memory biases in gymnastic judging. In addition, specific recommendations for the gymnastic competition setting are included as the findings suggest that judges' objectivity can be compromised by these memory biases. Copyright © 2003 John Wiley & Sons, Ltd.