Optimal allocation of sample size for randomization-based inference from 2 K factorial designs

Arun Ravichandran,Nicole E. Pashley,Brian Libgober,Tirthankar Dasgupta

DOI: https://doi.org/10.1515/jci-2023-0046

2024-01-01

Journal of Causal Inference

Abstract:Abstract Optimizing the allocation of units into treatment groups can help researchers improve the precision of causal estimators and decrease costs when running factorial experiments. However, existing optimal allocation results typically assume a super-population model and that the outcome data come from a known family of distributions. Instead, we focus on randomization-based causal inference for the finite-population setting, which does not require model specifications for the data or sampling assumptions. We propose exact theoretical solutions for optimal allocation in 2 K {2}^{K} factorial experiments under complete randomization with A-, D-, and E-optimality criteria. We then extend this work to factorial designs with block randomization. We also derive results for optimal allocations when using cost-based constraints. To connect our theory to practice, we provide convenient integer-constrained programming solutions using a greedy optimization approach to find integer optimal allocation solutions for both complete and block randomizations. The proposed methods are demonstrated using two real-life factorial experiments conducted by social scientists.

mathematics, interdisciplinary applications,social sciences, mathematical methods

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to optimize sample allocation in 2K - factor design to improve the precision of randomized inference and reduce costs. Specifically, the paper focuses on randomized causal inference in the finite population setting, a method that does not require model specification for data or sampling assumptions. The author presents the exact theoretical solution for the optimal allocation in 2K - factor experiments under complete randomization and explores it based on A -, D - and E - optimality criteria. In addition, the paper extends this work to factor designs with block randomization and derives the optimal allocation results under cost - constraint conditions. To combine theory with practice, the author provides an integer - constrained programming solution using the greedy optimization method to find the integer - optimal allocation solutions under complete randomization and block randomization. The application of the proposed method is demonstrated through two real - life factor experiments.

Optimal allocation of sample size for randomization-based inference from 2 K factorial designs

Optimal Designs for 2^k Factorial Experiments with Binary Response

On the Optimality of Randomization in Experimental Design: How to Randomize for Minimax Variance and Design-Based Inference

On Randomization-based and Regression-based Inferences for 2^K Factorial Designs

On Optimal Rerandomization Designs

Improving Covariate Balance in 2^K Factorial Designs via Rerandomization

Randomization-based Joint Central Limit Theorem and Efficient Covariate Adjustment in Randomized Block 2K Factorial Experiments

Causal Inference From 2(K) Factorial Designs By Using Potential Outcomes

Efficient Balanced Treatment Assignments for Experimentation

Optimal orthogonal designs for experiments with four-level and two-level factors

Sharpening randomization-based causal inference for $2^2$ factorial designs with binary outcomes

Randomization-based joint central limit theorem and efficient covariate adjustment in stratified $2^K$ factorial experiments

Rerandomization in $2^K$ Factorial Experiments

Nearly Random Designs with Greatly Improved Balance

Optimal Designs For Two Non-Interactive Treatment

Improving Covariate Balance in 2K Factorial Designs Via Rerandomization with an Application to a New York City Department of Education High School Study

Causal Inference from 2^k Factorial Designs Using the Potential Outcomes Model

Inference for Two-stage Experiments under Covariate-Adaptive Randomization

Optimal Stratification of Survey Experiments

Sample size planning for conditional counterfactual mean estimation with a K-armed randomized experiment

Design-based Causal Inference for Balanced Incomplete Block Designs