Acting Woman: Math Performance and Gender Identity

Abstract

In this paper I causally explore whether perceived math ability has an impact on perceived gender identity. I base my experiment on a framework proposed by Akerlof and Kranton (2000), and randomly assign individuals to receive a signal of relatively high or relatively low math ability. Because math is associated with male stereotypes, I hypothesize that individuals who perceive themselves as good at math identify as more masculine and believe they are perceived as more masculine by others. If true, this indicates women may suffer a gender identity cost for being good at math, which may be contributing to persistent gender gaps in STEM-fields. Including all respondents in my analysis I find no significant effect of treatment on perceived gender identity. When only including participants who consider math to be masculine or gender neutral there is a significant treatment effect indicating those who received a signal of high relative math performance identified as more masculine, and believed they were perceived as more masculine by others. This suggests women who perceive themselves as good at math may experience a gender identity cost, while men may experience a gender identity gain. However, the results are not robust to the inclusion of collected control variables and therefore needs to be confirmed in future research.