Isaac Ahimbisibwe (), Adam Altmejd (), Georgy Artemov (), Andrés Barrios-Fernández (), Aspasia Bizopoulou (), Martti Kaila (), Jin-Tan Liu (), Rigissa Megalokonomou (), José Montalbán (), Christopher Neilson (),
et al (3 additional author(s))
Sebastián Otero (), Jintao Sun () and Xiaoyang Ye ()
Additional contact information
Isaac Ahimbisibwe: Baylor University
Adam Altmejd: IFAU, Uppsala
Georgy Artemov: University of Melbourne
Andrés Barrios-Fernández: Universidad de los Andes
Aspasia Bizopoulou: VATT Institute for Economic Research
Martti Kaila: University of Glasgow
Jin-Tan Liu: National Taiwan University
Rigissa Megalokonomou: Monash University
José Montalbán: Swedish Institute for Social Research (SOFI), Stockholm University
Christopher Neilson: Yale University
Sebastián Otero: Columbia University
Jintao Sun: Renmin University of China
Xiaoyang Ye: Amazon
Abstract: Women make up only 35% of STEM graduates worldwide. We use ranked applications and assignments from ten centralized university admissions systems to distinguish gender differences in academic preparation (the pipeline) from differences in application decisions (the choice gap). Among students in the top 10% of each setting’s academic performance distribution, women are about 24 percentage points less likely than men to rank STEM first, averaging across settings. The pipeline gap varies widely and broadly tracks wage parity; the STEM choice gap remains large and negative everywhere. Linking applications to assignments shows that women’s underrepresentation largely precedes allocation. Pooled across settings, women are 51.2% of high achievers, 38.6% of those listing STEM anywhere, 34.0% of those ranking it first, and 34.7% of those assigned to STEM. Holding application rates fixed, a gender-balanced pipeline would still leave high-achieving STEM applicants between 57% and 75% male in every setting. A strong female academic pipeline can therefore coexist with substantial underrepresentation in STEM.
Keywords: gender inequality; STEM gender gap; centralized application platforms
Language: English
43 pages, October 28, 2026
Full text files
FULLTEXT01.pdfFull text
Questions (including download problems) about the papers in this series should be directed to Chloe Nibourel ()
Report other problems with accessing this service to Sune Karlsson ().
RePEc:hhs:sofile:2026_012This page generated on 2026-11-02 12:54:28.