Development of the overconfidence measurement instrument for the economic experiment

摘要

In this article results of the two experiments, aimed at the development of the instrument (test) that would enable construction of the comprehensive measure of individual overconfidence for the use in economic overconfidence experiments, are presented. Instrument was obtained in a two-stage procedure. In the first experimental phase, a pilot test, consisting of fifty general-knowledge questions of the unknown difficulty, was conducted to divide the items into three difficulty levels: hard, average-difficulty and easy questions. The second phase was aimed at verification of the replicability of results. Statistical tests supported the existence of the hard-easy effect, verified the success of categorization of questions into three levels of difficulty, and showed that gender was not associated with overconfidence in the developed instrument. The average group overconfidence measures obtained from both experimental phases did not differ from each other significantly. Instrument’s internal consistency was found to be good and acceptable for the use in social research. Compared to the tests used in the foregoing economic experiments, the obtained test is believed to result in the improvement of the overconfidence measurement quality.