Can Machine Learning Help Predict and Improve Student Outcomes? Evidence from a Randomized Control Trial

Read full text Summarise with VIRAT
Authors: Jacob Meyer
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

We conduct and evaluate an educational intervention to identify 1-6th grade students at risk of poor academic performance. The intervention provides teachers with predicted end-of-year state-administered exam scores, where predictions are formed using machine learning. We use survey and administrative data to measure the effect of this information on teacher beliefs, teacher efforts, and student outcomes. Our partner is a charter school district in the United States. To test the efficacy of providing these predictions, we implement a randomized control trial at the school-grade level. Teachers in the treatment group receive access to a report showing predicted outcomes for their students. Teachers in the control group continue with status quo information.

Fields of Study & Subjects

Online Learning and Analytics School Choice and Performance Educational Assessment and Improvement

Publication details

Publication date
1 January 2026
Language
EN
Document type
Preprint