Abstract:
The ability to model changes in preferences is crucially important for sound decision making and
effective communication. Much has been written about strategies for changing beliefs and preferences.
Typically such strategies have been driven by theoretical considerations, intuitive notions of rationality,
and an appeal to the principle of Minimal Change. In this paper we describe an experiment in which
people were asked to rank information, then given some new information, and asked to re-rank the
information. We analyse the results and provide comparisons with some well known computational
strategies. Some of the results are surprising, for example, a large percentage of human strategies can be
classified as either Conditionalization, Adjustment, or a combination.
