Overconfidence or Something Else:

The question whether overconfidence is truly a psychological bias has attracted considerable attention in the overconfidence literature. Several studies have suggested that people only appear overconfident as a result of the experimental design, or some biases other than overconfidence. This paper researched the existence of genuine overconfidence. The term refers to a psychological bias that makes people believe they are better than average, overestimate their skills and knowledge and be excessively confident in their beliefs. However, several other biases could also produce the same effect and make people appear overconfident. This thesis investigated those biases. Then, it tried to measure overconfidence while minimizing their impact. Consequently, a new approach to measure overconfidence was needed. The new approach presented in this thesis is composed of two parts. In the first part, a typical general knowledge test is replaced by a cognitive game. The aim of the carefully designed cognitive game is to improve the often criticised standard method of measuring overconfidence and to mitigate the effects of other biases as much as possible. In the second part, incentivized choices are used to elicit subjects’ beliefs. Measuring beliefs with incentivized choices gives participants a financial incentive to state their true beliefs and incentivized choices are rarely used in overconfidence experiments. Results from the experiment show that even when given monetary incentives, participants were not more successful in placing their performance relative to others. Moreover, subjects were overconfident in all overconfidence measures. In addition to overplacing their performance they also overestimated it and provided overly narrow confidence intervals, which support several of the genuine overconfidence hypotheses. In spite of the findings, the limitations of this study prevent us from comfortably making the conclusion that overconfidence is indeed a psychological bias.