Explaining gender differences in confidence and overconfidence in math

This paper investigates empirically how and why men and women are different in their confidence levels. Using the data of the PISA test in math, confidence is decomposed into two dimensions: confidence in correct math knowledge and overconfidence in over-claiming false knowledge. The findings highlight that female students are not less confident than male students, but they are rather less overconfident. Furthermore, mathematical abilities have different effects on male and female students. While ability alone increases confidence and decreases overconfidence, the interaction effect of feminine gender and ability is negative. This means that the negative effect of ability on overconfidence is larger for female students than male ones, while the positive effect of ability on confidence is smaller for females. That being said, the negative gender gap in overconfidence against girls is greater for students in the higher quartiles of math scores than those in the lower quartiles. Also, the positive gender gap in confidence for girls is smaller for well-performing students than underperforming ones. The empirical evidence further reveals that such gender-asymmetric effects of ability can be explained by gender socialization that limits women's roles and undermines their achievements.