Are inconsistent decisions better? An experiment with pairwise comparisons

Abstract

This paper tries to contribute to the well-known and lengthy discussion on whether we should allow decision makers to be inconsistent, or, to put it in another way, if intransitive decisions better represent their preferences. We will focus specifically on the multiple-criteria decision making problem, since it is one of the most usual stances where this intransitivity appears. Results from an experiment carried out with graduate students shows that, when intransitivities are removed, the preferences of decision makers are not better represented, and therefore that the effort to do so may not be justified.

This paper tries to contribute to the well-known and lengthy discussion on whether we should allow decision makers to be inconsistent, or, to put it in another way, if intransitive decisions better represent their preferences. We will focus specifically on the multiple-criteria decision making problem, since it is one of the most usual stances where this intransitivity appears. Results from an experiment carried out with graduate students shows that, when intransitivities are removed, the preferences of decision makers are not better represented, and therefore that the effort to do so may not be justified.