Title: Behavioral and Procedural Consequences of Structural Variation in Value Trees
Authors: Mari Pöyhönen, Hans C. J. Vrolijk, and Raimo P. Hämäläinen, Helsinki University of Technology
Date: November 1997
Status: Systems Analysis Laboratory Research Reports A69

We present experimental results showing that the division of attributes in value trees either increases or decreases the weight of an attribute for an individual decision maker. The structural variation of value trees may also change the rank of attributes. We propose that our new findings related to the splitting bias, some other phenomena appearing with attribute weighting, and the number-of-attribute-levels effect in conjoint analysis may have the same origins. One explanation is that decision makers' responses only reflect the rank of attributes and not the strength of their preferences. A procedural source of biases is the normalization of attribute weights. One consequence of these two factors is that attribute weights change if attributes are divided in a value tree. We also discuss how these problems in attribute weighting could be avoided in practice.

Click here to access a pdf version of this paper.

Back to the Decision Analysis Working Paper Index