Traditional models of perspective change problems (i.e., judgment of egocentric target directions from an imagined perspective) assume that performance reflects one’s ability to imagine the new perspective. Three experiments investigated whether advanced cuing of the imagination direction improves performance in an imagined self-rotation task. RT performance did not improve when extended time was given to complete the imagination process, or after participants reported the completion of the imagination. Moreover, when pointing to multiple targets after a single imagination, later responses did not show improvement. These results cast doubt on the traditional imagination hypothesis and suggest re-interpretation of the angular disparity effects in perspective changes tasks.