– Important factors include stereotyping by family, teachers, and peers; cultural roles; role models; competition and risk aversion; interests; and beliefs not directly linked to gender.
– Arguments for early biological differences are not conclusive and do not affect ability at kindergarten entrance. The literature is converging on a consensus that there are only small gender differences in mathematics tests scores at early ages, but the gender gap widens by middle and high school.
– More boys perform in both tails of the mathematics distribution with significantly more in the right tail than girls. To the extent that those who study STEM are drawn from the right tail of the math ability distribution, this may give rise to some of the gender differences we observe in STEM.
– Gender differences in mathematics test scores are mutable and can be influenced by family, teachers, culture, stereotypes, and role models throughout the schooling process.
– Teachers and to a lesser extent family are important contributors to gendered stereotypes and can have a negative influence on girls’ mathematics performance. Role models and gender equality in a given culture can decrease the gender gap in mathematics performance.
– The competitive nature of test-taking may understate girls’ true mathematical abilities. Changing the “growth mindset” belief that success in mathematics requires effort can improve girls’ performance.