Encouraging STEM Specialization Among High School Girls

How to cite

Ye, Karen; List, John A.; Kattan, Faisal; Abdelbaki, Nour. 2026. "Encouraging STEM Specialization Among High School Girls." AEA RCT Registry. AEARCTR-0018840. https://doi.org/10.1257/rct.18840-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Karen Ye* , John A. List , Faisal Kattan , Nour Abdelbaki

* Primary investigator

Registry ID
AEARCTR-0018840
DOI
https://doi.org/10.1257/rct.18840-1.0
First registered
Last updated
Status
On going
Trial dates
2026-05-14 – 2028-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18840

Abstract

Women remain significantly underrepresented in science, technology, engineering, and mathematics (STEM) education and employment in the Middle East. This study investigates the barriers preventing female students in Grades 10 - 11 from selecting specialized STEM high school tracks and evaluates the effectiveness of two interventions designed to close this gap. Qualitative evidence from focus groups with students, parents, and teachers reveals that while 83% of girls report enjoying STEM subjects, positive attitudes fail to translate into enrollment decisions due to persistent social norms, family influence toward gender-conforming career paths, and limited awareness of STEM's educational and labor market returns for women. To address these barriers, we implement two interventions that are cross-randomized across 40 schools: (1) role model videos featuring successful women in STEM, shown to both girls and their parents to shift perceptions of social acceptability and career viability, combined with information letters sent directly to both parents and students, emphasizing earnings returns, available government support for women in STEM, and the regional needs for STEM majors; and (2) default enrollment, automatically enrolling girls within a school into specialized STEM paths with an opt-out option to test whether inertia can convert positive attitudes into action. This design enables causal identification of each intervention's effects and potential spillovers, while generating cost-conscious, policy-actionable evidence for governments on how to sustainably increase girls' enrollment and retention in STEM education pathways.

Intervention

This study employs two cross-randomized interventions targeting both students and their parents to encourage STEM specialization among high school students. The intervention components are: (1) role model videos featuring women in STEM careers, designed to shift perceptions and aspirations, combined with an information letter providing parents and students with information about STEM opportunities for women; and (2) default enrollment into a STEM track (opt-out design), leveraging behavioral insights around inertia and defaults to increase STEM take-up.

Experimental design

This study evaluates two cross-randomized interventions aimed at increasing STEM specialization among secondary school girls in Saudi Arabia. The first intervention, consisting of role model videos and information letters, uses a randomized saturation design that first assigns schools to one of four saturation arms (0%, 20%, 50%, 80%), determining the share of students within each school eligible to receive the intervention, and then randomizes at the student-level within-school. The second intervention, consisting of the default enrollment condition, is randomized at the school-level, where schools are assigned to either an opt-in or opt-out framing.

The information intervention consists of two components delivered to both students and their parents: role model videos featuring women in STEM careers and an information letter about STEM opportunities. Additionally, the second intervention is default enrollment into a STEM track (opt-out) compared to active choice enrollment (opt-in). The primary outcome is students' high school track specialization (STEM vs. non-STEM). Secondary outcomes include changes in attitudes toward women in STEM, spanning social and family norms, perceptions of gender equality and capability, and views on STEM careers.

Design details

Randomization method
At the school level, randomization occurs across two dimensions simultaneously: (1) treatment saturation as schools are assigned to one of four saturation arms (0%, 20%, 50%, 80%), determining the share of students within each school eligible to receive the information intervention; and (2) default enrollment condition, where schools are assigned to either an opt-in or opt-out framing, crossed with saturation to yield eight school-level treatment groups. Target allocation across saturation levels is 30%, 20%, 30%, and 20% of schools respectively, split equally between opt-in and opt-out schools within each saturation level. School randomization is stratified by school size (above vs. below median enrollment). Within schools, individual students are randomly assigned to treatment or control (for the information intervention) consistent with their school's saturation target, with treatment status harmonized among siblings. A rerandomization procedure was implemented computationally using Python. Following school assignment, students within each school were randomly assigned to treatment or control consistent with their school's saturation target. A post-processing step then unified treatment status among siblings sharing the same parent within a school, with ties broken randomly. Across 50,000 iterations, the procedure selected the assignment minimizing the maximum deviation between realized and target shares of schools and students across saturation arms, subject to a tolerance of 2 percentage points.
Randomization unit
- Information intervention: Student-level randomization (w/ saturation randomized at school-level) - Default enrollment: School-level randomization
Clusters
40 schools
Observations
4,575 students
Arms
- 6 schools (pure control, opt-in), 6 schools (pure control, opt-out), - 4 schools (20% saturation, opt-in), 4 schools (20% saturation, opt-out), - 6 schools (50% saturation, opt-in), 6 schools (50% saturation, opt-out), - 4 schools (80% saturation, opt-in), 4 schools (80% saturation, opt-out)
Primary outcomes
High school track specialization, specifically whether students choose a STEM or non-STEM specialization track
Secondary outcomes
o For Students: - College and career aspirations and expectations - Beliefs about own ability to succeed in specialized STEM tracks, as well as relative ability of girls vs. boys - Attitudes toward gender norms in STEM - Second-order beliefs regarding gender norms in STEM - Beliefs about STEM majors and careers - Expected returns to STEM - 11th and 12th grade grades (after choosing STEM/non-STEM specialization track, if possible) - College enrollment and major (if possible) o For Parents: - College and career aspirations and expectations for daughter - Beliefs about daughter's ability to succeed in specialized STEM tracks, as well as relative ability of girls vs. boys - Attitudes toward gender norms in STEM - Second-order beliefs regarding gender norms in STEM - Beliefs about STEM majors and careers - Expected returns to STEM

Cited by

No citation data available for this registration yet.

Keywords