The Impacts of Tracking on Accumulated Learning Losses: Experimental Evidence from Brazilian Middle-schoolers Five Years After the Pandemic

How to cite

Costa, Thiago da; Lichand, Guilherme; Loeb, Susanna. 2025. "The Impacts of Tracking on Accumulated Learning Losses: Experimental Evidence from Brazilian Middle-schoolers Five Years After the Pandemic." AEA RCT Registry. AEARCTR-0016063. https://doi.org/10.1257/rct.16063-1.1

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

Thiago da Costa* , Guilherme Lichand , Susanna Loeb

* Primary investigator

Registry ID
AEARCTR-0016063
DOI
https://doi.org/10.1257/rct.16063-1.1
First registered
Last updated
Status
On going
Trial dates
2025-02-03 – 2025-12-31
Countries
Private
JEL codes
I21, I24, I28, C93
Registry record
http://www.socialscienceregistry.org/trials/16063

Abstract

This study evaluates a large-scale remedial education program implemented in public middle schools across São Paulo State, Brazil. The program identifies students who are behind in math and Portuguese and removes them from their regular classes a few times per week for small-group tutoring sessions. Tutors use materials tailored to each student's learning level, based on an initial diagnostic test. The goal is to help students catch up by focusing on basic skills from earlier grades, using a simple, low-tech approach. Unlike most tutoring programs, this one takes place during school hours, which makes it easier to scale and avoids the challenges of after-school logistics. The study will test whether this approach improves standardized test scores, and basic literacy and math skills. Results from this study will help policymakers understand whether it is possible to close learning gaps with cost-effective and scalable strategies that fit the reality of public education systems in developing countries.

Intervention

The intervention is a remedial education program designed to support middle school students in public schools who are significantly behind in math and Portuguese. Selected students are removed from their regular classes three to four times per week for 45-minute tutoring sessions in small groups of up to 15 students. These sessions are led by trained tutors using printed materials tailored to each student's current learning level, determined by a baseline diagnostic test. The materials cover foundational content from grades 1 to 5 and allow students to progress at their own pace. The program is implemented during regular school hours to avoid the logistical challenges of after-school programs and to maximize participation. The intervention is being rolled out across hundreds of schools in São Paulo State and aims to evaluate whether this low-cost, scalable approach can effectively reduce learning gaps.

Experimental design

his study uses a randomized controlled trial (RCT) to evaluate the impact of a remedial education program in São Paulo State public middle schools. Schools are randomly assigned to either a treatment group, which implements the individualized tutoring program, or a control group, which continues with regular instruction. Within treated schools, students are selected for tutoring based on low scores in standardized assessments. The program is implemented during regular class hours to ensure high participation. The evaluation will compare outcomes such as attendance, academic performance, and foundational skills between students in treatment and control schools to assess the effectiveness of the intervention.

Design details

Randomization method
Randomization done in an office by a computer using R.
Randomization unit
Schools
Clusters
999
Observations
Approx. 89000 students
Arms
399 schools control, 5000 schools treated.
Primary outcomes
The key outcome variables of interest in this experiment are: Standardized test scores in Portuguese and Mathematics, measured using the SARESP and Prova Paulista assessments. Foundational skills in literacy and numeracy, measured through diagnostic tools (ROAR for reading and SMARTE for math fluency). Learning levels across grade content, measured through a computer-based adaptive test (CAT) administered to all students in partnership with Parceiros da Educação and Catvante. Student attendance in Portuguese and Mathematics classes, used as a proxy for student effort. Grade progression and dropout rates, based on administrative data.

Cited by

No citation data available for this registration yet.

Keywords