Research
Job Market Paper
"Gender Gap in High-Stakes Exams: What Role for Exam Preparation?" with Léa Dousset
Elite higher education institutions often rely on competitive entrance examinations, which are presumed to measure students' abilities objectively. However, emerging evidence suggests a gender bias, with women frequently underperforming in high-stakes and competitive environments. This paper corroborates these findings within a highly selected group of elite STEM programs undergraduates in France. We extend our investigation beyond the effect of the high-stakes exam itself to consider the role of exam preparation in this gender performance gap. Our analysis reveals that heightened competitiveness in exam preparation exacerbates the gender performance gap. We also delve into a potential mechanism behind this widening gender disparity: the differential responsiveness to high-ability peers. These insights enhance our understanding of the gender gap in access to selective STEM higher education, which is particularly important in explaining the persistent gender pay gap at the top of the income distribution.
Notice relative au traitement des données collectées dans le cadre de ce projet
Proportion of female students, by decile of selectivity of STEM graduate schools
Working papers
"Unequal access to higher education based on parental income: evidence from France," with Sébastien Grobon, WID.world Working Paper, N° 2022/01, January 2022 (revise and resubmit, Journal of Human Capital).
In this paper, we provide new stylized facts on inequalities in access to higher education by parental income in France. On average, an increase of 10 percentile ranks in the parental income distribution is associated with a 5.8 percentage point (p.p.) increase in the proportion of children accessing higher education — 4.2 p.p. in the bottom half of the income distribution and 9.3 p.p. in the top half. This overall level of inequality is surprisingly close to that observed in the United States. We then document how these inequalities in access to higher education by parental income combine with inequalities related to parental occupation or degree. Finally, we assess the redistributivity of public spending on higher education and more generally of all public spending on young adults and their parents, and present a new accounting method to take into account the tax contribution of parents in our redistributivity analysis.
Access to higher education in France and the United States, by parents' pre-tax income
“The Concentration of Investment in Education in the US (1970-2018),” WID.world Working Paper, N° 2020/04, March 2020.
This study aims to analyse the concentration of investment in education in the US from 1970 to 2017. I study both the distribution of spending for K-12 and Higher Education and then present different scenarios to combine both inequalities. Even if the distribution of education spending is less unequal than the one of income or even wages, these spending are still very unequally distributed and, as for income and wages, inequalities have significantly increased over the past four decades, due to spending in higher education. Indeed, the top 10% of students for whom the most is spent used to have 28% of the overall amount of instructional expenditure in 1970 and now have more than 36%. Inequalities in educational investments are coming from two sources: unequal length of studies and unequal spending per grade, the latter being the main driver of the concentration observed. As a matter of fact, if everyone were to have the same educational attainment, the level of inequalities would almost be the same. The only way to reduce significantly the concentration in educational spending would be to equalize spending within each grade across districts and universities.
Work in Progress
"Price of Admission: The Impact of Application Fees on STEM Graduate Schools Applicants"
This study examines the impact of application fees on the application behavior and admission outcomes of STEM graduate school applicants in France. Using a Regression Discontinuity Design and data from the centralized admission process to STEM graduate schools from 2015 to 2020, the findings reveal a significant reduction in the number of applications of candidates who have to pay application fees, leading to adverse admission outcomes, especially for male students, those from lower socio-economic backgrounds, or with lower academic achievements. Additionally, the study shows that the design of application fees, whether decentralized or centralized, significantly impacts application patterns. The eviction effect of application fees is more pronounced in schools requiring a specific fee for each application, compared to those with a common application fee structure.
Publications in French
“Grandes écoles : des politiques d’« ouverture sociale » en échec”, with Pauline Charousset, Julien Grenet and Georgia Thebault, Éducation et Formations, no. 103, pp. 156—174, Avril 2022.
Grâce à des données riches et inédites, cet article analyse les inégalités sociales d’accès aux grandes écoles et leur évolution depuis le milieu des années 2000. Malgré les dispositifs d’« ouverture » qui ont été mis en place par certaines grandes écoles pour diversifier le profil de leurs étudiants, ces institutions d’élite sont restées presque entièrement fermées aux élèves issus de milieux sociaux défavorisés et leur base de recrutement n’a guère évolué au cours des quinze dernières années. Cette permanence des inégalités d’accès aux grandes écoles ne s’explique qu’en partie par les écarts de performance scolaire entre les élèves issus de milieux très favorisés et les élèves issus de milieux défavorisés. Elle trouve sa source, en amont, dans l’absence de diversification du recrutement des classes préparatoires et des écoles post-bac au cours de la période. Ce constat d’échec invite à repenser les leviers qui pourraient être mobilisés pour élargir le recrutement des filières sélectives.
“Les grandes écoles : De fortes inégalités d’accès entre élèves franciliens et non franciliens”, with Pauline Charousset, Julien Grenet and Georgia Thebault, Éducation et Formations, no. 102, pp. 367—393, June 2021.
À partir de données riches et inédites, cet article analyse les disparités géographiques d’accès aux grandes écoles et leur évolution depuis la fin des années 2000. Malgré les dispositifs d’« ouverture sociale » qui ont été mis en place par certaines grandes écoles pour diversifier le profil de leurs étudiants, la base géographique de recrutement de ces formations d’élite reste étroite et n’a guère évolué au cours de la période. Alors que les Parisiens ne représentent que 3 % de la population, ils constituaient 8 % des effectifs des grandes écoles en 2016-2017 et près de 25 % des effectifs des écoles les plus prestigieuses. À l’inverse, les non-Franciliens, qui représentent 82 % de la population, ne constituaient que 70 % des effectifs des écoles et 49 % des effectifs des écoles les plus prestigieuses. Les inégalités géographiques d’accès aux grandes écoles ne s’expliquent qu’en partie par les écarts moyens de performance scolaire et de composition sociale entre les départements d’Île-de-France et les autres départements. Leur persistance trouve notamment sa source, en amont, dans l’absence de diversification du recrutement des classes préparatoires et des écoles post-bac au cours de la période étudiée. La très forte concentration spatiale de ces filières sur le territoire francilien, particulièrement pour les plus sélectives d’entre elles, tend à avantager les élèves originaires d’Île-de-France dans l’accès aux grandes écoles.