Feedback manipulation and learning in games

This thesis uses theory and experiments to study how an efficiency loss can be minimized in two types of situations. While competing for a fixed prize people suffer an efficiency loss if they invest more than would be optimal. We show that these wasteful overinvestments disappear when people receive accurate feedback, can learn over time and know what others will choose. A different type of efficiency loss occurs when people adopt inefficient conventions from which no individual wants to deviate. We show that such inefficient conventions are more often overcome if players can easily reveal their choices and if a group contains more farsighted people. These findings highlight the importance of social media and could potentially be useful for group formation.