Skip to content
DraconiticResearch
Search Sources API
186,541 studies31 sourcesUpdated 42 min ago
PreprintarXiv licencearXiv

Network Experiments with Edge Treatments and Node Outcomes

  • Artem Kuriksha
  • Kenneth Hung
Open at sourcePDFDOICite

Abstract

We present a methodology for analyzing node-level outcomes while experimenting with edge-level treatments in a population connected by an undirected graph. Under our design, nodes are randomly assigned to test or control, and each edge inherits the treatment of its endpoints, with conflicts resolved by randomization. We use each node's assigned status as an instrument for its treatment exposure. We show that the Wald estimator is consistent for the global average treatment effect (GATE), even when the edge weights used to construct the exposure are misspecified. We formalize the assumptions needed both in terms of the linearity of potential outcomes and the sparsity of the graph, and prove the asymptotic normality of the Wald estimator. The estimator is straightforward to implement, requiring no assignment simulations over the graph. Our Monte Carlo study demonstrates the strong performance of our approach in the context of a social platform.

Related

  1. Identifying Causal Effects in Experiments with Latent Treatments: An Instrumental Variable Approach
    Lukas Junker, Florian Pargent
    PsyArXivPreprintCC BY 4.0Full textStatisticsOSF
  2. AI-assisted design and analysis of experiments with unstructured treatments
    Eli Ben-Michael, Zach Branson
    arXivPreprintCC BY 4.0Full textStatisticsarXiv
  3. Factorial Multivariate Bayesian Causal Forests: heterogeneous main and interaction effects of multiple treatments on correlated outcomes
    Danilo A. Sarti
    arXivPreprintarXiv licenceFull textStatisticsarXiv
  4. Gramian-Informed Framework for Edge Flow Network Analysis Against Nodal Attacks
    Prasad Vilas Chanekar, Bala Kameshwar Poolla, Jorge Cortés
    arXivPreprintCC BY 4.0Full textOptimisation & controlarXiv
  5. A General Variable Neighborhood Search for Logistic Network Flow Design with Nonlinear Edge Costs and Transshipment Capacities
    Varvara Kovaleva, Alexander Ponomarenko
    arXivPreprintarXiv licenceFull textOptimisation & controlarXiv

Cite

BibTeXRIS
BibTeX
@misc{kuriksha2026network,
  title = {{Network Experiments with Edge Treatments and Node Outcomes}},
  author = {Artem Kuriksha and Kenneth Hung},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.07363},
  eprint = {2610.07363},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.07363},
}
RIS
TY  - UNPB
TI  - Network Experiments with Edge Treatments and Node Outcomes
AU  - Artem Kuriksha
AU  - Kenneth Hung
PY  - 2026
DA  - 2026/10/05
PB  - arXiv
DO  - 10.48550/arxiv.2610.07363
UR  - https://arxiv.org/abs/2610.07363
AB  - We present a methodology for analyzing node-level outcomes while experimenting with edge-level treatments in a population connected by an undirected graph. Under our design, nodes are randomly assigned to test or control, and each edge inherits the treatment of its endpoints, with conflicts resolved by randomization. We use each node's assigned status as an instrument for its treatment exposure. We show that the Wald estimator is consistent for the global average treatment effect (GATE), even when the edge weights used to construct the exposure are misspecified. We formalize the assumptions needed both in terms of the linearity of potential outcomes and the sparsity of the graph, and prove the asymptotic normality of the Wald estimator. The estimator is straightforward to implement, requiring no assignment simulations over the graph. Our Monte Carlo study demonstrates the strong performance of our approach in the context of a social platform.
ER  -