Skip to content
DraconiticResearch
Search Sources API
174,684 studies31 sourcesUpdated 56 min ago
PreprintCC BY 4.0arXiv

Differentially Private Average Treatment Effect Estimation by Propensity Score Blocking

  • Duncan Stewardson
  • Grayson W. White
  • Adam Groce
Open at sourcePDFDOICite

Abstract

Average treatment effect (ATE) estimation in observational studies is a fundamental statistical tool used frequently in social science, medicine, and other fields. These fields often work with sensitive data where privacy protections are important, so a differentially private mechanism for ATE estimation is highly desirable. Here we present two propensity score-based algorithms for ATE estimation on observational data, one improving the inverse probability weighting (IPW) method used in prior work, and the other using blocking on the propensity score (BPS). Both show lower error and less bias than prior work, with the BPS-based algorithm frequently reducing error by 75% or more compared to prior work.

Related

  1. Differentially Private Synthetic Data via Generalized Moment Matching
    Rundong Ding, Yiyun He, Yizhe Zhu
    arXivPreprintarXiv licenceFull textProbabilityStatisticsarXiv
  2. Communication-Efficient Differentially Private Gradient Tracking for Distributed Optimization via Local Updates
    Mihitha Maithripala, Chenyang Qiu, Zongli Lin
    arXivPreprintarXiv licenceFull textOptimisation & controlarXiv
  3. From Privacy to Generalization: Linear Max-Information Bounds for Differentially Private Learning Algorithms
    Christoph H. Lampert, Max Cairney-Leeming, Hossein Zakerinia
    arXivPreprintCC BY 4.0Full textStatisticsarXiv
  4. Differentially Private Distributed Optimization via the Intrinsic Randomness of Zeroth-Order Estimators: Sampling Geometry, Convergence, and Privacy Guarantees
    Jiajun Wang, Yujie Tang
    arXivPreprintarXiv licenceFull textOptimisation & controlarXiv
  5. Towards Hyperparameter Transfer for Differentially Private Optimization
    Tianze Wang, Zhiqi Bu, Linjun Zhang
    arXivPreprintCC BY 4.0Full textStatisticsarXiv

Cite

BibTeXRIS
BibTeX
@misc{stewardson2026differentially,
  title = {{Differentially Private Average Treatment Effect Estimation by Propensity Score Blocking}},
  author = {Duncan Stewardson and Grayson W. White and Adam Groce},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.09536},
  eprint = {2609.09536},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.09536},
  note = {CC BY 4.0},
}
RIS
TY  - UNPB
TI  - Differentially Private Average Treatment Effect Estimation by Propensity Score Blocking
AU  - Duncan Stewardson
AU  - Grayson W. White
AU  - Adam Groce
PY  - 2026
DA  - 2026/09/08
PB  - arXiv
DO  - 10.48550/arxiv.2609.09536
UR  - https://arxiv.org/abs/2609.09536
AB  - Average treatment effect (ATE) estimation in observational studies is a fundamental statistical tool used frequently in social science, medicine, and other fields. These fields often work with sensitive data where privacy protections are important, so a differentially private mechanism for ATE estimation is highly desirable. Here we present two propensity score-based algorithms for ATE estimation on observational data, one improving the inverse probability weighting (IPW) method used in prior work, and the other using blocking on the propensity score (BPS). Both show lower error and less bias than prior work, with the BPS-based algorithm frequently reducing error by 75% or more compared to prior work.
ER  -