A simple questionnaire can change everything: Are strategy choices in coordination games stable?

This paper presents results from an experiment designed to study the effect of self reporting risk preferences on strategy choices made in a subsequently played 2x2 coordination game. The main finding is that the act of answering a questionnaire about one's own risk preferences significantly alters strategic behavior. Within a best response correspondence framework, this result can be explained by a change in either risk preferences or beliefs. We find that self reporting risk preferences induces an increase in subjects' risk aversion while keeping their beliefs unchanged. Our findings raise some questions about the stability of strategy choices in coordination games.