When Bayes goes bad: Weakly-regularized covariate adjustment leads to a biased estimate of prevalence

Swen Kuh, Lauren Kennedy, Qixuan Chen, Andrew Gelman · 2026

Read When Bayes goes bad: Weakly-regularized covariate adjustment leads to a biased estimate of prevalence by Swen Kuh, Lauren Kennedy, Qixuan Chen, Andrew Gelman (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-2603-29134.txt.

Description

Kuh and co-authors investigate a bias in Bayesian estimates of disease prevalence. They studied antibody prevalence in an Australian city in late 2020. As covariates were added one by one, estimates drifted downward. Simulations tested four possible causes of the drift. Weak priors turned out to drive extra partial pooling as the model grew. That pooling fed back through the test specificity estimate. The effect then lowered the final prevalence estimate. The authors give practical recommendations for such models.

Catalog record

CREATORSwen Kuh, Lauren Kennedy, Qixuan Chen, Andrew Gelman
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE80013
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.29134

Browse the full library · About the Tower of Babel