Explaining Gender Differences in Confidence and Overconfidence in Math

Abstract

This paper investigates empirically how and why men and women are different in their confidence levels. In the analysis, confidence is disentangled into two dimensions: confidence in correct math knowledge and overconfidence in false knowledge. Using the data of the PISA test in math, the findings highlight that math abilities have different effects on boys and girls. Overall, math abilities increase confidence and decrease overconfidence. However, the positive effect on confidence is smaller for girls, and the negative effect on overconfidence is larger for them. This gender-asymmetric effect implies that well-performing girls are more constrained from gaining confident attitudes through their abilities, compared to well-performing boys. The empirical evidence further indicates that the gender-asymmetric effect of abilities can be explained by gender socialization that undermines women’s achievements and limit their opportunities.