Embedded Files
My research emerges from my desire to improve the ability to detect small effect sizes in social-science settings. I approach this problem from both the design and analysis side.
Optimizing Multi-Level Adaptive Interventions
Optimizing Multi-Level Adaptive Interventions
This project proposes the Multi-Level SMART for optimizing multi-level adaptive interventions both with and without interference.
PWRD Aggregation
PWRD Aggregation
This project weights repeated measurements of the outcome to maximize power in settings where the intervention has a heterogeneous and delayed effect concentrated on a particular subset of individuals.
Model Selection through Simulation
Model Selection through Simulation
This project adapts the dry-run scheme proposed in Wyss et al. (2017) to randomized trials, enhancing the method through an additional subsampling step to deter overfitting.
For a vignette illustrating this process, see here
Improved Statistical Efficiency in SMARTs
Improved Statistical Efficiency in SMARTs
This project examines a suite of commonly used statistical techniques that can improve efficiency in SMART settings. It provides recommendations for which techniques should be applied across a variety of scenarios.
Page updated
Report abuse