The Power of Feedback

Last registered on September 21, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
The Power of Feedback
RCT ID
AEARCTR-0019511
Initial registration date
September 08, 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
September 21, 2026, 8:02 AM EDT

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
Katherina Thomas
Affiliation
OECD
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Marc Gurgand
PI Affiliation
Paris School of Economics (PSE) - CNRS - ENS-PSL
Contact Investigator
PI Name
Ilze France
PI Affiliation
University of Latvia
Contact Investigator
PI Name
Tim Fütterer
PI Affiliation
Eberhard Karls Universität Tübingen
Contact Investigator
PI Name
Zuzana Baranovičová
PI Affiliation
Education Policy Institute, Ministry of Education, Research, Development and Youth of Slovak Republic
Contact Investigator

Additional Trial Information

Status
In development
Start date
2026-09-17
End date
2027-01-31
Keywords
Education
Additional Keywords
Artificial intelligence, AI, Formative assessment, Learning, EdTech, Mathematics
JEL code(s)
I21, O33
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
The Power of Feedback is an international cluster-randomised controlled trial evaluating the effectiveness of AI-enabled formative feedback for primary school math students in four countries (Latvia, Germany, France and Slovak Republic). Teachers of Grades 4 and 5 are randomly assigned to use the OECD’s Platform for Innovative Learning Assessments (PILA) with their students, either with an integrated AI-enabled tutor or without the AI-enabled tutor. Both groups use the same mathematics modules, digital activities and teacher monitoring dashboard, so that the availability of AI-enabled feedback is main difference between the treatment and control. Over approximately 5-8 weeks, students use PILA during mathematics lessons twice a week for around 30 minutes under teacher supervision. The AI-enabled tutor provides formative feedback and support while students work on mathematics exercises, supporting reflection and different solution strategies rather than simply providing answers. The objective is to estimate whether access to AI-enabled formative feedback improves students’ math skills, measured using assessments administered after the intervention. The study also collects student and teacher questionnaire data and platform-use data to measure engagement, attitudes towards AI, and implementation quality.
External Link(s)

Registration Citation

Citation
Baranovičová, Zuzana et al. 2026. "The Power of Feedback." AEA RCT Registry. September 21. https://doi.org/10.1257/rct.19511-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
The intervention provides Grade 4 and 5 students with access to an AI-enabled tutor integrated into the OECD’s Platform for Innovative Learning Assessments (PILA) during math lessons. The AI-enabled tutor provides short, constructive, age-appropriate formative feedback while students work on math exercises, including prompts to reflect on their reasoning, identify mistakes and consider alternative solution strategies. Students in both the treatment and control group use the same PILA math modules teachers the same monitoring dashboard. PILA is used during regular math lessons approximately twice a week for 30 minutes over 5-8 weeks, under teacher supervision.
Intervention Start Date
2026-10-05
Intervention End Date
2026-12-31

Primary Outcomes

Primary Outcomes (end points)
Mathematics achievement (measured by math test administered on platform)
Primary Outcomes (explanation)
Mathematics achievement (measured by math test administered on platform)

Secondary Outcomes

Secondary Outcomes (end points)
Math anxiety, motivation and efficacy, teacher adjustment (teaching practices), engagement, attitudes towards AI
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The study uses a cluster-randomised controlled design, with the teacher as the unit of randomisation stratified by country. Participating Grade 4 and 5 teachers are randomly assigned 50:50 to the intervention (AI-enabled tutor) or control condition (no -enabled tutor) at account creation for the digital platform by country.
Randomization method: Automated randomization up on sign-up of the teacher to the digital learning platform by country (50:50).
Experimental Design Details
Not available
Randomization Method
Automated randomization up on sign-up of the teacher to the digital learning platform by country (50:50).
Randomization Unit
Teacher
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
350 teachers (50-100 by country).
Sample size: planned number of observations
350 teachers x 20 students (if 1 class) = 7000 (target).
Sample size (or number of clusters) by treatment arms
50:50 assignment by country.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
0.12 SD
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Paris School of Economics
IRB Approval Date
2026-04-02
IRB Approval Number
N/A
IRB Name
University of Tubingen
IRB Approval Date
2026-06-05
IRB Approval Number
N/A
IRB Name
University of Latvia
IRB Approval Date
2026-06-10
IRB Approval Number
N/A
IRB Name
Commission for Artificial Intelligence in Education and Training, Slovak Republic
IRB Approval Date
2026-06-24
IRB Approval Number
N/A
Analysis Plan