Beyond principal ignorability: Nonparametric sensitivity bounds for principal stratification
Xinyuan Chen Michael O. Harhay Fan Li · 2026
Read Beyond principal ignorability: Nonparametric sensitivity bounds for principal stratification by Xinyuan Chen Michael O. Harhay Fan Li (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-01669.txt.
Description
Chen, Harhay, and Li bound causal effects when a key premise may fail. Principal stratification handles intermediate variables in causal work. Point estimates of its effects need the principal ignorability premise. That premise cannot be tested in practice. Their frame is a sensitivity analysis for when it fails. A margin-free factor captures the size of the failure. It is built from relative risks tied to an unmeasured confounder.
Sharp bounds for each causal effect follow from that factor. The bounds sit inside worst-case bounds, with or without monotonicity. Cornfield-type rules say how strong confounding must be to nullify an effect. E-values state that same threshold in one number. The method extends to broader causal effects and pairwise estimands over a product space.
Catalog record
| CREATOR | Xinyuan Chen Michael O. Harhay Fan Li |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.01669