Methods:
For the presentation of the stimuli, we used Eye Lines (Beagley, 2001). The stimulus was made up of one horizontal and one vertical line bisecting each other. See Figure 1. For this stimulus, either the vertical or horizontal line was adjustable, this varied randomly but occured in even quantities. In the horizontal condition, a horizontal frame was put on the monitor, so that the stimulus is in the center and the vertical lines are closer to the edge. In the vertical condition, a vertical frame is used, again with the stimulus in the center. The third condition is control, where a square frame is used. The adjustable line randomly switches between horizontal and vertical for different trials. Subjects did 3 trials each of horizontal, vertical, and control, and in each trial there were two stimuli that were horizontally adjusted and two stimuli that were vertically adjusted, varying randomly. The room is set up so that there are black sheets hanging around the monitor, so that the only frame the subject will have is the one on the monitor.
Results:
16 subjects were used. We compared the effect of the illusion for adjusting the horizontal line with the that for adjusting the vertical line. With the vertical frame in place, the subjects averaged 5 monitor units under adjusted when adjusting the horizontal line and 18 monitor units under adjusted for vertical line adjustment (Figure 1). Each monitor unit is equal to .03mm. When the horizontal frame was in place, the subjects averaged 20 monitor units over adjusted for the horizontal line adjustment and 22 monitor units over adjusted for the vertical line adjustment (Figure 2). With the square, control frame, the subjects averaged 12 monitor units over adjusted in the horizontal line adjustment and 14 monitor units over adjusted for vertical line adjustment (Figure 3).
Figure 1. Average Adjustments on Vertical Frame
Figure 2. Average Adjustments on Horizontal Frame
Figure 3. Average Adjustments on Square (Control) Frame
Discussion:
It does seem possible that the frame hypothesis is correct. In the vertical frame condition, the subjects under adjusted the horizontal line and over adjusted the vertical line. This means that, when the vertical frame is on, the stimulus appears to have a longer horizontal line. This goes along with our hypothesis because the horizontal line is closer to the edges and thus may be the reason that it appears to be longer. However, the same effect did not happen in the horizontal frame condition. The subjects overadjusted both lines to an equal degree, just as what happened in the control. Thus, it appears as though the frame hypothesis does not hold true for the horizontal condition. However, there are several reasons why this may have happened. Firstly, people are used to computer screens having a horizontal frame, since that is how monitors are shaped, so the vertical frame appeared to be much more drastic than the horizontal frame. Secondly, the horizontal-vertical illusion may have played into our frame as well, causing the frame to appear longer vertically. These problems seem difficult to overcome, for the first has to do with how computers are built and what people are used to, and the second has to do with an illusion that naturally occurs. One solution may be to use a different kind of monitor and setup, so that frames would not be required, perhaps by having subjects look through a scope at a screen with frames set up on it. Because the vertical frame worked for the frame hypothesis, it is certainly worth further research to see if it can be proved that the horizontal-vertical illusion is caused by frames.