Transfer of feedback-produced decrement in the Horizontal-Vertical Illusion

To determine the extent to which feedback would improve performance on Horizontal Vertical Illusion (HVI) and to transfer different forms of the illusion, subjects adjusted the extended vertical line in inverted-T figures with long (100 mm), medium (75 mm) and short (50 mm) horizontal lines in an effort to make the vertical line equal in length to the horizontal line. After 4 trials on each size, experimental subjects viewed a correctly-adjusted medium size inverted-T figure and were told that the horizontal and vertical lines in that figure were equal in length. After trial 24 they were tested on three alternate forms of the illusion; (1) a production task (draw a one inch horizontal line and a 1 inch vertical line), (2) an adjustment task using L figures and, (3) a task requiring them to choose which of several sailboat drawings had the mast height equal to the length of the hull. Prior to the manipulation, both groups made the vertical lines shorter than the horizontal lines on each of the three figure sizes. After feedback on the medium size figure, experimental subjects were more accurate than the controls on that figure and on the small and large figures. However, there was no evidence that the improved accuracy of experimental subjects transferred to the L figure or production task: Neither group showed the HVI on the L-figure and both experimental and control subjects showed the illusion equally in the production task. Experimental subjects were, however, more accurate than controls on the structural basis of illusion decrement. The possible involvement of changes in eye scanning patterns in illusion decrement is discussed.