Estimating Optimal Dynamic Treatment Regimes for Personalized Education: A Tutorial and Applications with Machine Learning
Abstract
Optimal dynamic treatment regimes (ODTRs) are increasingly used to generate sequences of personalized recommendations that maximize a final outcome of interest. Although ODTRs are widely used in biostatistics and precision medicine, their applications in education remain limited, especially for longitudinal decision-making. This study translates ODTR methods for single- and multi-stage designs into individualized high school math course recommendations, providing step-by-step guidelines and demonstrations with data from the High School Longitudinal Study of 2009 (HSLS:09). We use Targeted Maximum Likelihood Estimation with ensemble machine learning to estimate ODTRs that maximize students' math achievement and college enrollment, imposing feasibility constraints based on domain knowledge and propensity-score thresholds to ensure recommendations are practically implementable. Results show that unconstrained regimes yield higher estimated values but infeasible recommendations, while feasible regimes produce realistic pathways with stronger empirical and expert-informed support. Finally, we discuss practical considerations for designing ODTRs in education.
Related
- Estimating Effects of Time-Varying Special Education Regimes Using Targeted Maximum Likelihood
- Estimating Average Treatment Effects with Projected Instrumental Variables
- Imputation is all you need: double robustness, semiparametric efficiency, and automatic covariate balance for estimating the average treatment effect
- tteICE: An R Package for Estimating Treatment Effects on Time-to-Event Outcomes with Intercurrent Events in Two-Arm Trials
- Estimating Treatment Effects Through Intermediate-Event Pathways with Multi-State Data
BibTeX
@misc{pan2026estimating,
title = {{Estimating Optimal Dynamic Treatment Regimes for Personalized Education: A Tutorial and Applications with Machine Learning}},
author = {Chenguang Pan and Yuxuan Li and Thomas Harris and Youmi Suk},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.10843},
eprint = {2610.10843},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.10843},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Estimating Optimal Dynamic Treatment Regimes for Personalized Education: A Tutorial and Applications with Machine Learning
AU - Chenguang Pan
AU - Yuxuan Li
AU - Thomas Harris
AU - Youmi Suk
PY - 2026
DA - 2026/10/07
PB - arXiv
DO - 10.48550/arxiv.2610.10843
UR - https://arxiv.org/abs/2610.10843
AB - Optimal dynamic treatment regimes (ODTRs) are increasingly used to generate sequences of personalized recommendations that maximize a final outcome of interest. Although ODTRs are widely used in biostatistics and precision medicine, their applications in education remain limited, especially for longitudinal decision-making. This study translates ODTR methods for single- and multi-stage designs into individualized high school math course recommendations, providing step-by-step guidelines and demonstrations with data from the High School Longitudinal Study of 2009 (HSLS:09). We use Targeted Maximum Likelihood Estimation with ensemble machine learning to estimate ODTRs that maximize students' math achievement and college enrollment, imposing feasibility constraints based on domain knowledge and propensity-score thresholds to ensure recommendations are practically implementable. Results show that unconstrained regimes yield higher estimated values but infeasible recommendations, while feasible regimes produce realistic pathways with stronger empirical and expert-informed support. Finally, we discuss practical considerations for designing ODTRs in education.
ER -