Peer gender and STEM specialization

ABSTRACT

This paper shows that students are less likely to specialize in mathematics when exposed to a high share of male peers. I exploit a curricular reform that incentivized students to obtain a mathematics qualification post-16. I show that, for those students affected by the reform, the higher the share of same-gender classmates, the higher the likelihood of obtaining a mathematics qualification for boys, and the lower the likelihood for girls. I interpret this as suggestive evidence that one’s perceived ability in mathematics, a boy-dominated subject, decreases when the share of male classmates increases. This further affects STEM participation in higher education.

KEYWORDS:

• Female participation in STEM
• peer effects
• school curricular reform
• degree choice

JEL CLASSIFICATION:

• J16
• I28
• I21

Acknowledgments

I am grateful to Thomas Cornelissen for a useful discussion and to an anonymous referee for his/her comments that improved this paper. I acknowledge the support of the ESRC Postdoctoral Research Fellowship (award no. ES/S011900/1). The data is supplied by the Department for Education and the Higher Education Statistics Agency. The author will provide guidance on obtaining the data. The use of the data in this work does not imply the endorsement of any of the organizations cited above in relation to the interpretation or analysis of the data. All remaining errors are my own.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ Science, Technology, Mathematics and Engineering.

² Morando (2020) shows that there is no clear evidence of the reform affecting this margin. Indeed, this sample restriction does not affect the results since when I consider the whole sample the findings hold.

³ The latter is inferred by exploiting the way in which the educational system is structured. In the period here considered the system of teaching and assessing mathematics at junior high school was a 3-tier system. Students were divided into three different classes where the material studied and the exam papers differed in their complexity. Foundation is the lowest tier and grades here range from D to G. In the intermediate tier grades vary from B to E and in the higher from A* to C. The sample is composed of students with a mathematics grade A*/B and I define their classmates as those students with a mathematics grade A*/C (higher tier) in the same school and cohort.

⁴ Results available upon request.

⁵ The median of same-gender peers in the mathematics class is 0.486 for boys and 0.522 for girls; the minimum value is zero for both boys and girls.

⁶ The data does not have information on teachers so I cannot include teacher fixed effects. However, the fact that mathematics teachers at secondary school have a low turnover (as their national supply is very low, The Royal Society 2007) means that the school fixed effect partly accounts for teacher characteristics. When the analysis is repeated for small schools, where school fixed effect is very likely to be equivalent to a teacher fixed effect, results hold.

Additional information

Funding

This work was supported by the Economic and Social Research Council [ES/S011900/1].