Although women make up the majority of students in higher education in France, they remain significantly under-represented in the prestigious engineering schools, particularly the most selective ones. These schools recruit through competitive entrance examinations, which are intended to ensure objective selection based on academic performance. Preparing for these exams requires two to three years in preparatory classes for the grandes écoles (CPGE), which are renowned for their intensity. The low proportion of women admitted to the most selective schools could be explained by two types of factors: differences in performance (before or during the CPGE, or at the time of the entrance examination), or differences in preferences regarding which entrance examinations to sit or the choice of placement. Our analyses rule out both the role of preferences and that of performance prior to entering the CPGE as the main causes, as these factors actually work in women’s favour. The crux of the explanation lies in a gradual reversal of the performance gap between women and men during the preparatory years, particularly in the most selective and competitive streams, the ‘étoile’ classes. This decline is accompanied by a slight disadvantage for women on the day of the entrance examination, which also contributes – albeit to a lesser extent – to their under-representation in the most selective engineering schools.
How can we explain the under-representation of women at the most selective engineering schools?
Presentation
Key Results
- This report is based on a previously unpublished database, compiled by cross-referencing comprehensive administrative data from the Engineering School Admissions Service, administrative data from the Ministry of National Education and Youth, and data collected from 18 science-focused preparatory classes in mainland France, covering the period 2015–2023.
- Upon entering a science-focused CPGE, women, on average, achieve better academic results than their male counterparts. However, at the end of these two to three years of preparatory classes, they are less likely to gain entry to the most selective engineering grandes écoles. They account for only 20 per cent of the student body there, compared with 25 per cent in the science-focused CPGE classes.
- This under-representation can be partly explained by lower performance on the day of the entrance examination, but above all by a reversal of the gender performance gap: initially in women’s favour, this shifts in favour of men during the first year of CPGE and widens right up to the entrance examinations.
- Our analyses show that this gap widens particularly in the most competitive environments, notably in the so-called ‘étoile’ preparatory classes, which reinforce the trend of women dropping out at a relatively higher rate.
Method and Data
This study draws on four main sources of data:
Administrative data from the centralised admissions process for engineering grandes écoles: For the first time for research purposes, the authors analyse administrative data from the Service de concours écoles d’ingénieurs (SCEI) for the period 2015–2023. This data includes demographic information on applicants (age, gender, social background, geographical origin, scholarship status, disability status), information on their academic background (the sixth-form college where they attended their preparatory class, course stream, ‘star’ or ‘non-star’ class, number of attempts at the entrance examinations), as well as detailed data on the admissions process: examinations taken, results, ranking of schools in order of preference, and offers of admission at each stage. The authors focus on science-oriented CPGE programmes operating throughout the entire period and with at least 10 students enrolled per year.
Administrative data from the Ministry of National Education and Youth (2010–2020): Data from the SCEI were matched with those from the Directorate for Evaluation, Forecasting and Performance (MENJ-DEPP) using a unique encrypted identifier. This enables us to examine candidates’ previous academic performance, in particular their marks and average marks per subject in the baccalaureate and the National Brevet Diploma (DNB).
Data collected from 18 science-focused preparatory classes: Between February 2022 and April 2023, we collected data from 18 science-focused preparatory classes (CPGE) (MP/MP*, PC/PC*, PSI/PSI*, PT/PT*) that agreed to take part in our study. This data covers the demographic (gender, date of birth) and academic (stream, year group, marks per subject) characteristics of their students. This data provides us with detailed information on students’ marks during their preparatory year for approximately 17 per cent of our study sample. Figure 2 shows the geographical distribution of these CPGE institutions.
Analysis samples: The authors use several subsamples in our analyses: (i) the full sample of candidates (N = 165,450); (ii) a restricted sample, limited to CPGE programmes and streams offering both a ‘star’ class and a non-‘star’ class in the second year (N = 89,079); and (iii) a survey sample, based on data collected from the 18 CPGE institutions mentioned above, statistically matched to administrative data relating to their students (N = 16,789). The latter also includes the sample used for the discontinuity regression analysis of the effects of admission to a ‘star’ class (N = 6,630).
Partners
This paper was supported by CEPREMAP and the Chair in Education Policy and Social Mobility. Established in 2021 as part of a partnership between the Ardian Foundation, the Directorate for Evaluation, Forecasting and Performance of the Ministry of National Education and Youth (MENJ-DEPP) and PSE-Paris School of Economics, this chair aims to promote high-level research and the dissemination of knowledge.
The authors would like to thank the Engineering Schools Admissions Service for granting access to and providing their data.