The Impacts of AI Tools for Instruction on Learning: Evidence from the Dominican Republic

Last registered on March 05, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
The Impacts of AI Tools for Instruction on Learning: Evidence from the Dominican Republic
RCT ID
AEARCTR-0018003
Initial registration date
February 26, 2026

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
March 05, 2026, 6:40 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name
Raissa Fabregas
Affiliation
University of Texas - Austin
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Julian Cristia
PI Affiliation
Inter-American Development Bank
Contact Investigator
PI Name
Emma Naslund-Hadley
PI Affiliation
Inter-American Development Bank
Contact Investigator
PI Name
Pablo Celhay
PI Affiliation
Universidad de Chile
Contact Investigator
PI Name
Ariana Grossi
PI Affiliation
Inter-American Development Bank
Contact Investigator
PI Name
Maria Loreto Biehl
PI Affiliation
Inter-American Development Bank
Contact Investigator

Additional Trial Information

Status
On going
Start date
2025-12-01
End date
2027-01-31
Keywords
Education
Additional Keywords
AI, Instructional tools, math, pedagogy, secondary education, test scores
JEL code(s)
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
We evaluate the impact of an AI-based instruction platforms on student mathematics learning in public secondary schools in the Dominican Republic. In partnership with the Ministry of Education (MINERD), we randomly assign math teachers in 400 schools to receive access to one of three widely used AI platforms designed to support mathematics instruction, paired with structured training and ongoing implementation support. We measure impacts on student performance on end-of-year mathematics exams and use survey data to examine changes in teachers' instructional practices and time use.
External Link(s)

Registration Citation

Citation
Biehl, Maria Loreto et al. 2026. "The Impacts of AI Tools for Instruction on Learning: Evidence from the Dominican Republic." AEA RCT Registry. March 05. https://doi.org/10.1257/rct.18003-1.0
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
The intervention consists of three AI-based educational platforms on student mathematics learning, along with training and implementation support.
Intervention Start Date
2026-02-01
Intervention End Date
2026-06-06

Primary Outcomes

Primary Outcomes (end points)
End-of-year student test scores
Primary Outcomes (explanation)
The test score data wil be collected by the research team. Scores will be standardized to have mean zero and standard deviation one within the control group.

Secondary Outcomes

Secondary Outcomes (end points)
Teacher adoption and use of AI tools; teacher instructional practices; teacher time allocation; teacher well-being and job satisfaction, effect differences across platforms, heterogeneity by baseline teacher instruction+digital skills and student performance. 
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Schools are randomly assigned to one of three treatment arms, each receiving a different AI platform, or to a control group. Randomly selected math teachers in treatment schools receive platform access, training, and implementation support. Control schools continue with business as usual.
Experimental Design Details
Not available
Randomization Method
Randomization done in office by a computer
Randomization Unit
Randomization is at the school level (and randomly select math teachers within treated schools, if more than 1 math teacher in that school)
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
Total number of schools: 400

Sample size: planned number of observations
Total number of schools: 400. Total number of participating math teachers: 426. Total number of students: 12,780
Sample size (or number of clusters) by treatment arms
3,780 students in the control arm and 9,000 in one of the treatment arms (split equally among the 3 platforms). 126 teachers in the control schools, and 300 teachers in the treatment schools.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Our primary specification will pool all treated arms together. For this we expect to be able to detect MDEs of 0.13 SD for our primary outcome (student test scores) and MDEs of 0.3 SD for secondary teacher outcomes.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Etikos
IRB Approval Date
2025-11-25
IRB Approval Number
CEI-E-2025-31
Analysis Plan