Abstract
Magicians use misdirection to prevent their audiences from detecting their secrets. We designed a magic trick that could be used to investigate the way in which attention can be manipulated to prevent people from perceiving a visually salient event. By digitally manipulating this trick we demonstrate that participants' verbal reports reflect what they have seen rather than inferences about how they thought the trick was done. Moreover, eye movement records revealed that for a subset of participants, the detection of the event was not related to where they were looking at the time of the event. These results demonstrate a striking real world example of how overt and covert attention can be spatially dissociated. However, detection of the event resulted in rapid shifts of eye movements towards the detected event, thus indicating a strong temporal link between overt and covert attention. However, rather surprisingly, these results suggest that covert attention can be allocated at least 2 or 3 saccade targets ahead of where people are fixating.
Wolfson Research Institute, University of Durham.