Measuring Constructed Preferences 

John W. Payne1, James R. Bettman1, and David A. Schkade2
(1Duke University and the 2University of Texas)

Abstract

This paper discusses the implications of recent work on the psychology
of preferences for the measurement of values. In particular, the goal
of the paper is to begin a dialogue leading to a "building code" for a
world of constructed preferences. It is clear from recent research
that the preferences expressed for objects of any complexity and
novelty are often constructed - not merely revealed - in the
generation of a response to a judgment or choice task. Further, there
appear to be some common "faults" in the construction of preferences
(e.g., people often commit what might be called a fallacy of one and
focus on a single option, a single objective, or a single assumed
state of the world). The presence of such faults in constructing
preferences raises important questions for those interested in the
measurement of values, whether the purpose is for predicting future
decisions or for helping guide future decisions. For example, how
might information be provided to help ensure quality construction, and
how would we know a "well-constructed" value or preference if we saw
one? Building on the work of many researchers, this paper proposes
some guidelines for measuring constructed preferences (a "building
code") that may mitigate some common faults in decision
making. Further, we discuss how such a "building code" for
constructing preferences needs to be sensitive to the purpose of the
measurement exercise and to the nature of the person providing the
valuation.