New JRSS-B Paper Quantifies How Strong Confounders Must Be to Overturn Research Conclusions

burny_tech · x · 2026-09-10

A paper by the author with Chad Hazlett is officially out in JRSS-B. It introduces new sensitivity analysis tools that precisely quantify how strong unobserved confounders would need to be to overturn a study's conclusions — turning 'how fragile is this finding to hidden bias' into a computable quantity.

Original post →

More from Research

Research channel →