Reducing overconfidence – Causal evidence from intention-to-treat estimates

IZA Logo

IZA Seminar

Place: Schaumburg-Lippe-Str. 9, 53113 Bonn

Date: 15.04.2014, 12:15 - 13:30


Presentation by 

Hannah Schildberg-Hörisch (Heinrich Heine University Düsseldorf)


Overconfidence is a widespread and well-documented phenomenon. Overconfident individuals systematically overestimate their own abilities. Empirical evidence on the consequences of overconfidence highlights diverse negative implications such as underestimation of personal risks. The main contribution of this paper is to present an intervention that reduces the prevalence of overconfidence. We use data from a randomized social experiment and estimate intention-to-treat estimates to identify the causal effect of the intervention on overconfident behavior. More precisely, we compare children who have been randomly chosen to be eligible for a one year mentoring program to otherwise identical control group children. Finally, we document that a mentoring intervention bears the potential to close the developmental gap between children from families with low socio-economic status (SES) and children with high SES who are less likely to display overconfident behavior.

For more information, please contact