Thomas Breda and Clotilde Napp; (PNAS July, 2019)
Although girls now have better academic careers than boys, obtaining better results at the brevet or baccalaureate, and pursuing higher education more often, they nonetheless remain significantly under-represented in fields related to mathematics. For example, they represent less than a third of the enrollment in engineering schools. However, it is often these fields that lead to the best paid professions, in the fastest growing sectors and the least subject to salary gaps between women and men.
However, the differences in level of mathematics between girls and boys are now very small in most countries and do not explain the large differences in education and career choices between the sexes. Social science researchers have therefore turned to other explanations such as differences in self-confidence, preference or discrimination.
The article sheds light on the debate on the causes of gender segregation between professions in a new light. We reconsider the role of academic performance and we show that taking into account not only results in mathematics but also those in letters (or reading) makes it possible to account for a large part of the differences in orientation between girls and boys.
Small differences in performance
To do this, we use individual data from the international PISA survey, Program for International Assessment, conducted in 2012. The study of performance in mathematics and literature on a sample of 300,000 15-year-old students in 64 countries (the 35 countries of the Organization for Economic Cooperation and Development OECD and 29 partner countries) reveals that boys remain slightly better than girls in mathematics, while girls are much better in letters. The gap in letters in favor of girls is three times greater than the gap in mathematics in favor of boys. These differences give girls a real comparative advantage in literary disciplines compared to boys: two thirds of them are better in letters than in mathematics, while this is the case for only a third of boys.
The PISA 2012 survey also includes questions to measure intentions to pursue studies and careers related to mathematics. We find that boys have more intention of studying mathematics, with differences observed that vary greatly from one country to another. We see that small differences in performance in mathematics can only explain around 10% of these differences in intentions.
On the other hand, the results are radically different when we try to explain orientation and career intentions by the comparative advantage for maths rather than by the level in maths alone. Taking comparative advantage into account makes it possible to account for 75% of the differences between the sexes in terms of intentions to study mathematics across all 64 countries studied.
This means that if we focus on students with the same gap between grades obtained in mathematics and grades obtained in French, the difference between girls and boys in terms of intention to study mathematics decreases by 75%. Girls are mostly better at letters than at mathematics and when it comes to making choices, they would identify themselves as more literary than scientific and their comparative advantage in letters would lead them to abandon more mathematical courses.
The PISA 2012 survey also makes it possible to measure students' declared interest in mathematics as well as their self-confidence in this subject. These variables do much less to explain the orientation gaps between girls and boys than comparative advantage.
Furthermore, we observe the same phenomenon as when we seek to explain career choices: differences in level of mathematics explain very little of the differences in declared interest or self-confidence in mathematics between girls and boys while Differences in comparative advantage make it possible to account for this almost completely. This shows that self-confidence or interest in a field is established by comparing one's performance in different fields: at the same level in mathematics, the better level of girls in letters harms the way they perceive themselves in mathematics, and in particular to their self-confidence in this discipline.
This important role played by gender differences in academic performance and comparative advantage at age 15 raises questions about their origin. We suggest that they are probably largely of cultural origin, determined by previous socialization processes, in the family environment and at school.
For example, we observe that performance gaps between the sexes are higher in countries in which stereotypes associating mathematics with boys are stronger. We also note that the organization of education systems can reduce these gaps.
To promote more equal representation of girls and boys in mathematics subjects, limiting differences in comparative advantage, for example by trying to improve the level of boys in French, should be effective. Another option would be to improve information for students when making choices, to encourage them to rely less on comparative advantage and more on career prospects. These interventions would of course be carried out in addition to those intended to limit gender stereotypes and their impact, from a very young age, on the academic careers of girls and boys.