Rigorous research shows that boys and girls exhibit similar core numerical skills during infancy and early childhood, but the pervasive stereotype that boys are better than girls in the subject leads to behaviors that quickly transform that initial promise into later reality. As years of schooling progress, an achievement gap emerges and widens between the genders, and ultimately females end up consistently underrepresented in math-intensive college majors and jobs. Prior evidence suggests that the trouble starts early, and a new report in the journal Nature sharpens the picture.
A team of researchers from France and the United States, led by Pauline Martinot, a physician with a background in neurodevelopment and global health, makes use of four years of data from the French national testing program called EvalAide. It is a series of tests designed by the education ministry to give every French teacher in first and second grade a detailed picture of the needs, achievement levels, and progress of each child in their class, covering both language and mathematics. French law generally requires students to begin first grade in the September following their sixth birthday. Nationwide testing is conducted at the start of first grade, the middle of the first-grade year, and at the start of second grade. Results from each administration are given to teachers and parents, with the idea that teachers will explain to families what pedagogical means are being and will be used to boost outcomes for the next test, and families will have direct insight into what is and is not working along the way. Martinot and her team follow four cohorts of students entering first grade across the country from 2018 to 2021—more than 2,650,000 children.
The average mathematical performance of boys and girls was nearly identical at the start of first grade. By the middle of first grade, however, a small but already highly significant gap favoring boys was observed. By the start of second grade, that effect size had quadrupled, creating a giant gap between boys’ and girls’ test performance. This pattern repeated in each of the four years studied here—with no variation for Covid-impacted school years—and varied only slightly across family, class, and school type and across socioeconomic levels. Various other robustness checks were made, including analyzing only the parts of the EvalAide test that were identical across all test administrations (the majority of questions are designed to increase in difficulty with each test), controlling for variations in student age at test administration, and controlling for varying amounts of schooling received prior to first grade. The researchers even analyzed selected pairs of boys and girls who were closely matched on both math and language test results, as well as numerous demographic factors at the start of first grade. These students’ math outcomes still diverged in a similar pattern to the overall findings on both subsequent test administrations.
The researchers learned more about how the gap is forming: At the first test administration, although boys and girls had the same mean score, boys were over-represented at both ends of the distribution (lowest and highest deciles). Thus, similar averages for boys and girls. After one year of schooling, however, the picture changes dramatically, with the top 5 percent of test performers comprising more than twice as many boys as girls (2.3 boys for each girl). And this same pattern was observed in the vast majority of individual classrooms—tens of thousands of them across France—every year. Martinot and her team caution that these are not causal findings and likely don’t generalize beyond France, but they are satisfied that something is occurring within first grade classrooms across the country to manifest these alarming shifts in academic outcomes.
They conclude with speculation on the causes of this pervasive pattern, largely focusing on the behaviors of adults—teachers and parents/guardians—who adhere to gender-based stereotypes and transmit them to the young people in their care. It is also suggested that math starts to “get serious” as soon as it becomes an “academic subject” in its own right and that this new emphasis on math achievement might serve to amplify adult mindsets around it and to make students, especially girls, more receptive and responsive to those preconceptions, whether problematic or beneficial. As to solutions, the report encourages education for teachers and parents to erase gendered stereotypes, as well as any anxieties of their own on the subject, and suggests schools should provide access to strong role models of scientists, analysts, teachers, astronauts, and statisticians for girls to appreciate and emulate. Good practices for anyone interested in boosting the presence of women and girls in math and science.
SOURCE: Pauline Martinot et al., “Rapid emergence of a maths gender gap in first grade,” Nature (June 2025).
