Seeing reversals in ambiguous images: to know or not to know?

The present study is an attempt to replicate the 1992 findings of Rock and Mitchener that knowing about the ambiguity in ambiguous images is necessary for seeing reversals. In contrast to Rock and Mitchener's study, the present study used a between-subjects design, in which 190 participants were randomly assigned to either an uninformed or an informed condition, and saw 12 images, 6 ambiguous and 6 unambiguous. The findings show that participants in the informed condition were significantly more likely to see reversals in the ambiguous images than were those in the uninformed condition. In contrast to Rock and Mitchener's findings, knowledge of ambiguity did not ensure that all informed participants saw all six reversals; there were spontaneous reversals in the uninformed condition for most of the images. These findings are discussed in terms of suggesting that seeing a reversal in an ambiguous image is a complex process that requires consideration of the interplay of visual perception and higher order cognition.