“…We selected the Sander parallelogram illusion for several reasons. First, the magnitude of illusion is relatively large (often 20-25%), is comparable to the magnitude of the Mueller-Lyer illusion (25-30%), and increases when the diagonals are removed (Cooper, Runyon, Tatz, & Heimer, 1972;Coren & Girgus, 1978). Second, the Sander illusion permitted a comparison of performance on the different tasks without modification of the parallelogram; that is, the informational context was the same across tasks.…”