Mediation Analysis for Probabilities of Causation
Abstract
Probabilities of causation (PoC) offer valuable insights for informed decision-making. This paper introduces novel variants of PoC-controlled direct, natural direct, and natural indirect probability of necessity and sufficiency (PNS). These metrics quantify the necessity and sufficiency of a treatment for producing an outcome, accounting for different causal pathways. We develop identification theorems for these new PoC measures, allowing for their estimation from observational data. We demonstrate the practical application of our results through an analysis of a real-world psychology dataset.
Cite
Text
Kawakami and Tian. "Mediation Analysis for Probabilities of Causation." AAAI Conference on Artificial Intelligence, 2025. doi:10.1609/AAAI.V39I25.34886Markdown
[Kawakami and Tian. "Mediation Analysis for Probabilities of Causation." AAAI Conference on Artificial Intelligence, 2025.](https://mlanthology.org/aaai/2025/kawakami2025aaai-mediation/) doi:10.1609/AAAI.V39I25.34886BibTeX
@inproceedings{kawakami2025aaai-mediation,
title = {{Mediation Analysis for Probabilities of Causation}},
author = {Kawakami, Yuta and Tian, Jin},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2025},
pages = {26823-26832},
doi = {10.1609/AAAI.V39I25.34886},
url = {https://mlanthology.org/aaai/2025/kawakami2025aaai-mediation/}
}