An economic study of the preference for power.

Summary This work studies the preference for power, from its definition to its encounter in the laboratory or in the field, in an experimental way. We build this thesis around three questions: what is power and the preference for power? How to identify it, distinguishing it from other types of social preferences? What are the elements, conditions of experiences or subjects' situation, that affect it? We define power as a dispositional capacity, which can be applied or not. This allows us to define then the notion of preference for power, which is nothing else than a formalized version of the will to power that we meet since antiquity. We then define four types of preference: pure or instrumental preference for potential or effective power. In the second part we build a model of power that we test in the laboratory. We identify this preference for power in its four aspects, which is the central result of this work. We then seek to understand what impacts this preference, to observe that only gender has a significant effect. Finally, we measure this preference, to find that there may be a significant correlation between pure preference for potential and actual power. The third part aims at studying these power preference behaviors among decision makers in the synthetic world of Eve Online, which we observe to be a good substitute for the laboratory. In a counterintuitive way, our observations show that the decision-makers in the virtual world do not show more preference for power than the other subjects.