This HDOSE Strategic Reinvestment project, A Comparison of Machine Learning Methods to Detect the Causal Heterogeneous Treatment Effects for Multilevel Randomized Controlled Trials, supported methods work at the intersection of causal inference and machine learning.
Within the lab, this project focuses on using machine learning to study causal heterogeneity: not only whether interventions work, but also for whom and under what conditions.
Project Information
- Funding source: University of Florida HDOSE Strategic Reinvestment Fund
- Role: Principal Investigator
- Co-PI: Walter Leite
- Project period: May 1, 2022-April 30, 2023
- Award amount: $12,500
Project Focus
- Compare machine learning methods for detecting heterogeneous treatment effects.
- Address multilevel randomized controlled trial settings.
- Support methodological guidance for causal heterogeneity analyses in education.
Related Publications and Working Papers
- Li, W., Gao, X., Ren, S., & Dong, N. (2026). Heterogeneous treatment effects for impact evaluations. Manuscript under review.
- Konstantopoulos, S., Li, W., Miller, S., & van der Ploeg, A. (2019). Using quantile regression to estimate intervention effects beyond the mean. Educational and Psychological Measurement, 79(5), 883-910. doi:10.1177/0013164419837321
- Strickland, K. J., Hill, J., & Li, W. Estimating heterogeneous treatment effects of the gifted and talented program using Bayesian additive regression trees. Working paper.
Related Presentations
- Strickland, K. J., Hill, J., Lu, Y., & Li, W. (2026). Estimating heterogeneous effects of the gifted and talented program using Bayesian additive regression trees. Modern Modeling Methods Conference.
- Li, W., Leite, W., & Quan, J. (2024). Application of machine learning algorithms to detect treatment effect heterogeneity for three-level multisite experiments. AERA Annual Meeting.
- Li, W. (2024). Using Machine Learning Methods to Detect Heterogeneous Treatment Effects for Multilevel Studies in Education. UF Education Policy Brown Bag.