Increasing Parent and Teacher Investments to Improve Student Outcomes at Scale

How to cite

Bhargava, Palaash; Ye, Karen; List, John; Kattan, Faisal; Hwang, Jiwon. 2026. "Increasing Parent and Teacher Investments to Improve Student Outcomes at Scale." AEA RCT Registry. AEARCTR-0018365. https://doi.org/10.1257/rct.18365-1.0

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

Palaash Bhargava* , Karen Ye , John List , Faisal Kattan , Jiwon Hwang

* Primary investigator

Registry ID
AEARCTR-0018365
DOI
https://doi.org/10.1257/rct.18365-1.0
First registered
Last updated
Status
On going
Trial dates
2026-04-22 – 2026-12-31
Countries
Private
JEL codes
D24, D91, I20, I25, I28
Registry record
http://www.socialscienceregistry.org/trials/18365

Abstract

Learning outcomes are shaped by investments made by multiple actors around the child, most prominently parents and teachers. However, little is known about how to increase these investments at scale, or how they interact in the education production function. In this paper, we study whether low-cost behavioral interventions can shift parental and teacher investments, and whether these inputs act as complements or substitutes in shaping student outcomes. We implement a large-scale randomized evaluation in partnership with the Saudi Arabian Ministry of Education, embedded within its national communication platform. We cross-randomize goal-setting and reflection interventions targeting parents and teachers, allowing us to identify both independent and interaction effects. We first examine whether these interventions increase the quantity and quality of investments made by parents and teachers. We then study how these changes translate into student outcomes, using administrative data on grades and engagement. Finally, leveraging the overlapping design, we estimate whether parent and teacher inputs are complements in practice. Our setting allows us to evaluate behavioral interventions within government systems and provides evidence on both the scalability of such approaches and the education production function, with implications for education policy in both high- and low-income contexts.

Intervention

This study evaluates a scalable behavioral intervention implemented in partnership with the Saudi Arabian Ministry of Education. The intervention consists of two cross-randomized components embedded within the Ministry’s existing communication platform, Madrasati: one targeting parents and one targeting teachers. Over the course of six weeks, treated parents and teachers receive weekly structured prompts encouraging them to set goals related to children’s learning and to complete short reflection exercises on whether they met those goals. The intervention is designed to increase salience, reduce planning frictions and procrastination, and improve the quantity and quality of investments in children’s learning. It is fully asynchronous, low-cost, and minimally burdensome. Parents are prompted to undertake activities such as helping with homework, organizing study routines, reading with the child, and communicating with teachers. Teachers receive analogous weekly prompts plus a short primer describing evidence-based pedagogical practices, including comprehension checks, greater two-way interaction, support for struggling students, and communication with families. Control participants continue to receive standard Ministry communications through Madrasati but do not receive the intervention prompts.

Experimental design

The study is a cross-randomized field experiment implemented in 46 public schools in Saudi Arabia (23 primary and 23 intermediate schools). The sample includes students in grades 4–9, their parents or guardians, and teachers employed in participating schools. Parents are assigned to treatment or control through individual-level randomization of students with stratification by school-grade cells, median baseline engagement, and median baseline academic performance. Teachers are assigned through individual-level randomization stratified by school. The design allows identification of: (i) first-stage effects of the parent intervention on parental investments, (ii) first-stage effects of the teacher intervention on teacher investments, (iii) effects of parent treatment on student outcomes, (iv) effects of teacher treatment on student outcomes, and (v) interaction effects between parent and teacher interventions. The parent intervention operates at the student level, while the teacher intervention maps to student-subject outcomes, allowing analysis at a granular student-by-subject level.

Design details

Randomization method
Randomization is done in the office by the PI on a computer.
Randomization unit
Individual for both interventions.
Clusters
This is an overlapping randomized intervention. We work 46 schools (23 primary and 23 secondary) with 3 grades per school. Interventions are randomized at an individual level for both parents and teachers.
Observations
7328 parents, 687 teachers, 8182 students with 9 subject grades available per student.
Arms
Both parent and teacher level randomization done at an individual level. Therefore, for the parent intervention, we have a total of 3675 parents in the treatment arm and 3653 parents in the control arm. For the teacher intervention, we have a total of 349 teachers in the treatment arm and 338 teachers in the control arm.
Primary outcomes
Student academic achievement and first-stage changes in parent and teacher investments. Student achievement is measured using administrative end-of-semester subject-level grades. First-stage parent outcomes include indices of academic time investment, socio-emotional investment, communication with teachers, and diversity of activities. First-stage teacher outcomes include analogous indices capturing time investment, pedagogical practices, socio-emotional investment, communication with parents, and diversity of activities.
Secondary outcomes
Student engagement measures from administrative records, such as days logged into Madrasati, assignments completed, quizzes or learning tasks submitted, and interactions with digital learning materials. Belief measures for both parents and teachers: Likert-scale measures of self-efficacy and the perceived importance of educational investments, and hypothetical-scenario elicitation measures following Attanasio, Boneva, and Rauh (2022) to recover perceived returns to parental effort, school quality, and complementarities between the two. Weekly goal-setting reflections to measure compliance, uptake, activity choice, and descriptive mechanism variables, including completion rates and reported activity patterns over time.

Cited by

No citation data available for this registration yet.

Keywords