Piehl J
Percept Mot Skills. 1976 Aug;43(1):47-50. doi: 10.2466/pms.1976.43.1.47.
Rectangles haveing different ratios between the lengths of their sides were presented in three different ranges: with the golden section either (a) the next to most elongated rectangle, (b) the middle rectangle, or (c) the next to shortest rectangle. Each of 90 subjects had to pick the rectangles they liked best from all three ranges. This basic procedure was run under three different conditions: (a) subjects did not know the stimuli before, (b) subjects knew them from a foregoing psychophysical experiment, (c) the same, with reinforcing the golden section. The rectangles at both extremes of each range were chosen first most often under the first two conditions, whereas the golden section was chosen most often under the last condition. The conclusion is that preference for the golden section is an artifact of the stimulus range and of the demand characteristics of the experimental procedure rather than of any intrinsic aesthetic quality.
黄金分割矩形要么是（a）第二细长的矩形，（b）中间的矩形，要么是（c）第二短的矩形。90名受试者中的每一个都必须从所有三个范围中挑选出他们最喜欢的矩形。这个基本程序在三种不同条件下进行：（a）受试者之前不知道这些刺激物，（b）受试者从之前的心理物理学实验中了解它们，（c）同样的情况，但强化了黄金分割。在前两种条件下，每个范围两端的矩形最常被首先选中，而在最后一种条件下，黄金分割矩形最常被选中。结论是，对黄金分割的偏好是刺激范围和实验程序的需求特征造成的假象，而非任何内在审美品质的体现。
