Introducing the b-value: combining unbiased and biased estimators from a sensitivity analysis perspective
Zhexiao Lin, Peter J. Bickel, and Peng Ding · 2026
Read Introducing the b-value: combining unbiased and biased estimators from a sensitivity analysis perspective by Zhexiao Lin, Peter J. Bickel, and Peng Ding (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-2602-16310.txt.
Description
Zhexiao Lin and coauthors introduce the b-value, a tool for combining unbiased and biased estimators from a sensitivity analysis perspective. In empirical research with multiple estimators for the same parameter, a central question is how to combine fair but less precise estimators with biased but more precise ones to improve results. The authors focus on a less studied inference question: how to conduct valid statistical inference with unknown bias.
They propose combining unbiased and biased values from a bias perspective, giving confidence ranges keyed by bias size. . The b-value is the key value of unknown maximum bias at which combining them does not yield a clear result. Applied to three standard combined cases, the authors find bias limits where results change. They advise giving the b-value based on the soft-threshold method, which is robust to unknown bias and achieves the lowest worst-case risk.
Catalog record
| CREATOR | Zhexiao Lin, Peter J. Bickel, and Peng Ding |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 184 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-16310