The Effects of Splitting Attributes on Weights in Multiattribute Utility Measurement
研究发现，在价值树中，当目标被拆分为更详细的属性时，这些属性会被赋予更高的权重，这种过度加权偏差在多种权重方法中均存在，但整体判断法受影响较小。
This study examined how weights in multiattribute utility measurement change when objectives are split into more detailed levels. Subjects were asked to weight attributes in value trees containing three objectives which were specified by either three, four, five, or six attributes. The robust finding was that the more detailed parts of the value tree were weighted significantly higher than the less detailed ones. This overweighting bias was found for several weighting techniques, but the techniques that used holistic judgments to derive weights were affected somewhat less than techniques that used decomposed attribute weights. This bias is interpreted in terms of the increased salience and availability of attributes that are spelled out in more detail.