Dependent Randomized Rounding for Budget Constrained Experimental Design

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yamin, Khurram, Kennedy, Edward, Wilder, Bryan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912430124171264
author Yamin, Khurram
Kennedy, Edward
Wilder, Bryan
author_facet Yamin, Khurram
Kennedy, Edward
Wilder, Bryan
contents Policymakers in resource-constrained settings require experimental designs that satisfy strict budget limits while ensuring precise estimation of treatment effects. We propose a framework that applies a dependent randomized rounding procedure to convert assignment probabilities into binary treatment decisions. Our proposed solution preserves the marginal treatment probabilities while inducing negative correlations among assignments, leading to improved estimator precision through variance reduction. We establish theoretical guarantees for the inverse propensity weighted and general linear estimators, and demonstrate through empirical studies that our approach yields efficient and accurate inference under fixed budget constraints.
format Preprint
id arxiv_https___arxiv_org_abs_2506_12677
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dependent Randomized Rounding for Budget Constrained Experimental Design
Yamin, Khurram
Kennedy, Edward
Wilder, Bryan
Machine Learning
Policymakers in resource-constrained settings require experimental designs that satisfy strict budget limits while ensuring precise estimation of treatment effects. We propose a framework that applies a dependent randomized rounding procedure to convert assignment probabilities into binary treatment decisions. Our proposed solution preserves the marginal treatment probabilities while inducing negative correlations among assignments, leading to improved estimator precision through variance reduction. We establish theoretical guarantees for the inverse propensity weighted and general linear estimators, and demonstrate through empirical studies that our approach yields efficient and accurate inference under fixed budget constraints.
title Dependent Randomized Rounding for Budget Constrained Experimental Design
topic Machine Learning
url https://arxiv.org/abs/2506.12677